PDA

View Full Version : Attn Starrett No. 199 Owners!



Fasttrack
01-28-2010, 11:07 AM
There was a discussion here a few weeks ago about the Starrett no. 199 master precision level.

(http://bbs.homeshopmachinist.net/showthread.php?t=39090)


YOD suggested that the levels are built with a degree of concavity on purpose and that they should NOT be scraped flat.

Well I contacted Starrett yesterday and got the following response:



Yes they are made concave. If you scrap it flat, the heat of your hand could cause it to rock as the metal expanded. If the scrape marks look good, I would suggest leaving it alone.

Mcgyver
01-28-2010, 11:15 AM
interesting, did not know that, I guess it makes sense. whenever i've used mine, for scraping or for setting up a machine, its a comparative tool; in words being dead on level doesn't matter, is the relative position of the bubble when readings from two places are compared.

if you want to scrape a concave surface, I have a surface plate you could use :D

lazlo
01-28-2010, 12:30 PM
I didn't know that either Tom! (One of the Starrett 199 owners) :)

oldtiffie
01-28-2010, 11:25 PM
There was a discussion here a few weeks ago about the Starrett no. 199 master precision level.

(http://bbs.homeshopmachinist.net/showthread.php?t=39090)


YOD suggested that the levels are built with a degree of concavity on purpose and that they should NOT be scraped flat.

Well I contacted Starrett yesterday and got the following response:


Originally Posted by Paul Nault (Starrett Tech Support)
Yes they are made concave. If you scrap it flat, the heat of your hand could cause it to rock as the metal expanded. If the scrape marks look good, I would suggest leaving it alone.

I missed seeing this post earlier.

That's fine - I can see the logic.

What I can't see is just how bowed it is and from that how far it needs to "bow" under hand heat until it is dead flat - and what the operating temperature range is.

Next of course is if the level needs to be "re-done", just how "bowed" do you scrape it - and how do you actually and accurately verify it?

Send it back to Starrett? At what cost?

J Tiers
01-29-2010, 12:10 AM
That is fine if you always have the whole thing on a surface, but sometimes you use it on parallels as checking twist on a lathe. Then concavity is a potential problem.

I strongly doubt that the concavity is much... about as much as you would see by a lightening of the "marking" towards the center.

personally, I'd make it flat, and pick it up with a cloth in your hand...... more versatile. better an error when I might expect it (when heated by hand) than perfect in bad conditions and bad in perfect conditions.

oldtiffie
01-29-2010, 02:39 AM
Well JT,

you may have opened a can of worms here - again.

If the level is only supported at the ends and is in affect a "bridge" between its ends and if it is able to twist, then it seems that there is a good case to have adjustable screws as a tri-pod with one screw on one end and the two other screws the other end. A tripod once settled will not rock.

With that unknown "hump" it makes it a bit suspect as regards using the base of the level as a straight edge too.

Which, given that a "Camel-back" bridge straight edge is machined and/or scraped flat and given too that the basic structure is similar, why is the same caution that is applied to the level not applied to the Camel-backed straight edge that people use as a reference for scraping other flat surfaces?

And lets not forget the possibility (probability) of uneven lighting and uneven heating on the surface plate too.

So, while others cogitate and obfuscate about that lot, I will get back to my machine-top faces and my float-glass "flats".

oldtiffie
01-29-2010, 03:40 AM
Just as a matter of interest, I chased up the details of the Starrett 199 and similar level details at the Starrett web site.

Very interesting.

Their level castings are "seasoned", the base is scraped and there is no mention of any "curve" or "hump" either.

The accuracy is about 1:24,000 (inch) or 1:25,000 (metric).

The metric 1:25,000 is 0.0004:1 which is 0.0004"/" or 4 "tenths" of an inch per inch which re-stated is 4.8 tenths per foot which re-stated again is 1:25,000 is the Atan of 0.00292 degrees x 60 = 0.1375 minutes x 60 = 8.25 seconds which is pretty close to an 8" "Phase 11" or "Vertex" rotary table which is indexed (via the vernier) to 10 seconds.

So.

Let's not get sucked in by all the hype.

It IS a VERY good level - it just is not all that some who "beat" or "hype" it up before having a look at - or understanding - what the figures really mean.

I have no good reason to doubt its performance - or those made in other areas - such as Asia and Europe - either.

The link to the *.PDF file on the Starrett web site is:
http://www.starrett.com/download/338_p439_444.pdf

Forrest Addy
01-29-2010, 05:31 AM
I think it boils down to scrape it how you like it.

I've gone for generations thinking Starrett levels were scraped flat. Starrett sales info and specs make a big deal out of their scraping job but they say nothing about a calculated concave sole intended to counter the heat absorbed from the user's hand. Given absense of information to the contrary, I've taken for granted the level's sole was flat. I've even used it as a scraping reference a few times when I needed something that size.

I attributed the concavity I found to small changes in the level's frame casting over time. So now I hear from a trusted source supported with a quote from Starrett that the Starrett #199 level soles are deliberately scraped concave.

I feel Starrett kept imporant information from me. I don't care if something is deliberately one way or another. I just want to know precisely what has been monkeyed with and why so when I use the tool to its maximum potential I can make the appropiate compensations.

I don't know how many Starrett levels I've re-scraped flat thinking the frames warped over time. I even remarked on it in recent posts. Shame on me. Now I feel like an idiot and think of Starrett as patronizing the knowledgeable users of its products. It's as though a skilled user couldn't be cautioned about heat input from casual handling of a precision tool.

I present myself as an arbiter of such things and to be ignorant of Starrett's scraping practice on their level soles after 50+ years in the trade is embarassing to say the least. I apologise to my justifiably shrinking cadre.

All that said, the concavity is small (I estimate about 0.0005" over the length) and not enough to count unless you bridge the level inconsistantly across 1-2-3 blocks or something similar.

Still the Starrett #199 level is a superb tool and should be used and trusted to the limits of its calibration - or better if reversal technique is used. I have several and their bases will remain flat as I scraped them. I will continue to handle them with cotton gloves when the degree of accuracy warrants; the heat of my hands will not unequally warm the frame and so distort the sole. Concavity/convexity is not an issue.

If stuff is scraped with a deliberate bias as in the production Starrett#199 level it should be placarded on the tool, noted in its calibration cert, and on the storage box label.

Your Old Dog
01-29-2010, 07:31 AM
That is fine if you always have the whole thing on a surface, but sometimes you use it on parallels as checking twist on a lathe. Then concavity is a potential problem.............

That's likely why they come in so many sizes, that would be my guess.

If we are leveling a surface plate the level being concave is a moot point. If we are measuring machinery that has been scrapped dead flat it's still a moot point. But, it's not a moot point if we are measuring anything with a slightly convex surface. A flat level would simply teeter back and forth like a rocking horse on a slightly convex surface. How accurate could that be?

Don't they make dedicated straight edges for that use?

Mcgyver
01-29-2010, 08:24 AM
I don't know how many Starrett levels I've re-scraped flat thinking the frames warped over time. I even remarked on it in recent posts. Shame on me. Now I feel like an idiot and think of Starrett as patronizing the knowledgeable users of its products. It's as though a skilled user couldn't be cautioned about heat input from casual handling of a precision tool.

I present myself as an arbiter of such things and to be ignorant of Starrett's scraping practice on their level soles after 50+ years in the trade is embarassing to say the least. I apologise to my justifiably shrinking cadre.


.

Forrest, you're as good a writer as you are metal worker, that was funny...I don't think you have to fret any cadre shrinkage.

when i read Tom's account, i thought of applications where I've used it, (none being as a spotting referece) and i can't see how the work would suffer for the concavity....Provided the level was in close to the same fore & aft position when checking a lathe, i don't think such small concavity would affect the reading......then again i wonder about all those bowed things you scraped :D

J Tiers
01-29-2010, 09:04 AM
You cannot "level" a non-flat surface, but you CAN level between any two POINTS or LINES, and frankly, I don't want some "decision maker" at Starrett deciding that my level must be "pre-distorted" to compensate for something which may not occur if a person with some sense uses it. See below for disclaimer.

NOT ONLY THAT

It sounds like total bull$hit to me, even if some idiot at Starrett said it. And I use the term "idiot" advisedly.

If you "pick something up", you are probably holding it BY THE TOP PART.

So the TOP will be "heated" more than the bottom, which will EXPAND THE TOP, and BOW THE ENDS DOWN. Until it equalizes, of course.

In other words, that will make the concave scraping worse, i.e. the level will be MORE CONCAVE.

if you DO "equalize it" at a HIGHER temperature than teh machine, then when you set it down, the bottom will CONTRACT, BOWING THE CENTER UP and MAKING THE CONCAVITY WORSE.

Apparently the claim is that they are trying to compensate for people who pick it up, and set the contact surface of the level on their bare hand as they carry it around.

or for people who set a cold level on a hot machine surface.

ONLY THOSE actions (or similar) could heat the BOTTOM, expanding it and causing the ENDS TO BOW UP. ONLY THOSE could be compensated by concave scraping.

I ain't buying it. Someone is not quite telling the truth here. They probably ARE concave, BUT NOT FOR THAT REASON. Eitehr the concavity is totally ineffective, as suggested by some who say it doesn't matter, or it can matter.

I tend to think it is fairly inconsequential as far as measurement, and it brings up the point of why Starrett saw fit to bother.

My conclusion is that IT HAS NOTHING WHATEVER TO DO WITH HEATING, THAT IS ALL SMOKE AND MIRRORS. The TRUE REASON is that if you scrape it concave, very slightly, you will be sure it will sit flat. END OF STORY.

I have done that with vise bases, and other things which I wanted to have sit flat and not rock. It works, provided you KNOW about it, and there is no good reason why that is a problem for the device in use. Starrett most likely scrapes them that way for that reason, if in fact they DO put in a concavity. At a half tenth concavity, you may not even notice a slight lesser marking in the center. Might be a bit more than that, but probably not much.

Nothing like the "banana" I had to send back to Grizzly.




If we are leveling a surface plate the level being concave is a moot point. If we are measuring machinery that has been scrapped dead flat it's still a moot point. But, it's not a moot point if we are measuring anything with a slightly convex surface. A flat level would simply teeter back and forth like a rocking horse on a slightly convex surface. How accurate could that be?



Nonsense...........

How "level" CAN a non-flat surface EVER be? IT CAN'T BE LEVEL, of course......

So there is no point in discussing leveling a cylinder in the round direction..... and every convex part is essentially a portion of a large cylinder, albeit possibly not a perfectly round one.

Fasttrack
01-29-2010, 10:20 AM
Forrest, no worries about your "shrinking cadre"! In regards to the level, I posit the following:

Starrett scrapes the sole slightly concave in order to keep the level from "rocking" on a flat surface. It is probably done as a "safety factor". If slight uneven heating/sunlight etc causes the sole to move, it will either become more concave (acceptable if you are measuring at the ends of the level) or it will become flat. I seem to recall reading that the levels should only be used at the ends - i.e. it should be supported at either end and no where else.

So it is a combination of what Starrett says and what JT says. I suspect it is done mostly to "cover their own a**es" but it does compensate for expansion due to uneven heating.


Anyhow, I didn't post this to criticize anyone or their practices. Just passing on an interesting tidbit of information I found ;)

gwilson
01-29-2010, 11:00 AM
My mint 199 is also hollow. I never liked it,especially when I have to set the level on 2 parallels across my lathe's bed. What I've always done is just set the 199 with exactly the same amount of overhang at each end of the lathe,so that it doesn't give a false reading due to the concavity.

I also have always worn gloves when handling it. I agree with the above post that says that heating the TOP of the level would only make it MORE concave. Makes perfect sense.

I haven't tried to measure the concavity of my 199,but it sure seems like it is more hollow than they say it is.

The whole thing doesn't make any sense to me at all. What idiot would carry their 199 in their bare hands while resting the base of the level in those warm hands?

I suppose I should carry the level around for a while,and then test it with blue in the surface plate?

I just find it irritating that the level is concave at all. If I used it more often,I think I would do like Forrest and scrape the damned thing flat. I'd like it better that way!!

Jpfalt
01-29-2010, 12:18 PM
I also am a Starrett 199 owner and find a couple of inconsistencies. I don't handle the level by the sole, but rather from the frame. If the frame heats, the sole gets more concave, not less.

Second, Starrett didn't say how much it is concave. It may be only a tenth, but it will still stop rocking of the tool.

I heard much the same discussion about large bandsaw wheels. Some said the wheels should be crowned across the tire, some say the wheels should be flat. However, both agree that the wheels must not be hollow.

I would say that flat or slightly hollow on the level should both work as long as you know what the condition is and accommodate the geometry in your work method.

Your Old Dog
01-29-2010, 12:54 PM
With the exception of my surface plate and my lathe bed I don't think anything I have tried to level has been flat as pi$$ in a pan. Therefore, the ever so slight arch serves me instead of working against me. I don't use it as a straight edge like I don't don't use a phillips screwdriver for a straight bladed screwdriver. I wouldn't expect my digital caliper spine to be straight enough to take the place of a machinist straight edge made for the purpose.

Forrest, what do you propose to do now? Now that you know the level is bowed do you plan on doing a recall on a life time of work? Your going to be busy grasshopper :D

Forrest Addy
01-29-2010, 01:15 PM
Old Dog, I scraped the soles my of levels flat so the problem is moot.

On the subject of heat: when I was an apprentice there was an old millwright came through the shop once in a while. He had a cart with bicycle wheels he carried all hs stuff in and among it were a couple of precision levels in nice wood boxes. He was a famous grouch easly irritated by most anything and scrutiny by a young guy was cause for hm to spin around and yell: "What the hell are you gawkin' at! Mind your own F'n busness!!" and variations. Us apprentices were terrified of him.

He died not long after I started my econd year. His hunting buddy put up his personal tools for sale in the shop. Milton Brown bought his levels which were not Starrett #199 but of the same precision, Taft Pierce I think. The old grouch had padded the frames with fleece hair side in and baseball stitched it neatly around the frames. We were puzzled by this refinement but now I know they were shrewdly insulated for ease of handling.

oldtiffie
01-29-2010, 03:10 PM
Oh dear, oh dear, oh dearee dearee me.

The wailing and gnashing of teeth and only part-muffles by all the sack-cloth here is dreadful, dreadful - do you hear me? - dreadful!!!!

It seems that not only has that paragon of virtue and omnipotence in all things Engineering - aka Starrett - not only not told you of its practices but has with-held the reasons for them.

It must be like seeing Old Nick in the tabernacle when you've donned your white gloves, done your cleansing and incantations, offered sacrifice and burned incense etc. before opening the Tabernacle marked "Starrett".

It must be just as bad as - or it is actually - being cuckolded?

Perhaps its more a chastity belt than a tabernacle that's needed for that old tart Starrett.

Me?

I'd drown it in the baptismal font.

Old Nick, Starrett and Tiger Woods all seem to have similar score cards.

oldtiffie
01-29-2010, 08:15 PM
Well,

I'd always wondered why High Priests and Vestal Virgins (in the Starrett Temple??) wore beards and veils respectively - but the self-satisfied smirk and the twinkle in the eyes gave it away!!!

Naughty naughty!!

Perhaps the Temple Starrett is a bordello (for all comers) or a Harem (for the "insiders" and the "selected few" (more likely).


With the exception of my surface plate and my lathe bed I don't think anything I have tried to level has been flat as pi$$ in a pan. Therefore, the ever so slight arch serves me instead of working against me.

YOD's "pi$$ in a pan" (font) is looking more likely by the minute.

But while I am on a roll, I thought I'd re-visit the Starrett site here in OZ.

http://www.starrett.com.au/

http://www.starrett.com/pages/3_precision_tools_and_gages.cfm

http://www.tru-stone.com/

http://www.tru-stone.com/pages/smp.asp

http://www.tru-stone.com/pages/smp.asp#plate

Ta-da (drum roll and trumpets) - and here we are:
http://www.tru-stone.com/pages/forms_lit.asp

with the "inch" specification here:
http://www.tru-stone.com/pdf/FedSpecGGG-P-463c.pdf

and the amended (metric) specification here:
http://www.tru-stone.com/pdf/FedSpecGGG-P-463c%20amendment%201.pdf

I've known for quite a while that the "specification" for "flatness" that just about every one here seems to use (the inch version) - at page 6, clause 3.4.4.1 - is a metric standard in "inch" clothes.

ie Total flatness tolerance in micro-inches = (40 + D^2/25) is actually an "inch-ised" version of the metric standard as 40 ~ number of inches ("D") in a metre (39.37) and 1/25 is ~ 0.03937 ~ 0.04 the approximate inch equivalent of 1mm.

Note too that the "rounding" figure for grade AA plates is 25 which just happens to co-incide with 25.4 micro-inches ~ 1 micrometer - aka 1 micron.

So the "inch" standard is a metric one in disguise.

Now onto the updated version of the "Specification" which - surprise, surprise is all metric.

http://www.tru-stone.com/pdf/FedSpecGGG-P-463c%20amendment%201.pdf

Pages 3, 4 and 5 are the pages of interest here.

All dimensions are metric - plate sizes (mm) mass (Kg) and tolerances micro-metres (um) - aka microns.

The plate flatness tolerances (um) are at Table 111 (page 4 of 9) and are in um.

The flatness formula - at page 5 of 9 - is:
(1 + 1.6 D^2) x 10^-6 micro-metres - aka um - aka micron

Here is Starrett's ISO 9001 Certification (Note: ISO is the international metric standard):
http://www.tru-stone.com/pdf/iso9001%20cert.pdf

Next, NIST has been metric for quite a while.

The Starrett web site does not show the sizes of the surface plates it makes, but I'd guess they make metric and "inch" plates although I suspect that the "inch" plates are actually made as metric plates.

So back to the Starrett levels.

Have a look here:
http://www.starrett.com/download/338_p439_444.pdf

Yep - metric and inch.

I don't blame Starrett at all for all the hull-a-balloo and gnashing of teeth and wringing of wrists here.

Too many believe too many myths for too long and regard them as "Gospel" - when, as it turns out, all that was needed was to ask Starrett and they'd have told you.

Their web-site said all that it needed to be said.

I am amazed - no not true - I was once but not now - that nobody with all the "expertise" and "knowledge" here had every thought to check their Starrett level bases to see how flat they were "ex-factory".

Or perhaps many most? all? only bought used ones and never bothered to get or read the original manual that came withe the tool/level when new.

To that extent its getting to be a re-run of the "co-axial indicator" thread.

I don't think that Starrett has reamed anyone but some have done a good job of it for themselves - on themselves.

A lot of stuff here should start with "Once upon a time ............. " and hopefully end with "............................. and lived happily ever after".

A lot less gullibility might help too.

Your Old Dog
01-29-2010, 08:31 PM
Forrest I don't doubt your story for a second but I question his motive. We had a woman where I worked who used to take quick little choppy steps and never left her desk without an arm load of video tapes..... I referred to them as props! She always looked busy but I'm not so sure she was.

I think your millwright might have realized the spin that would circulate the shop about his precision levels. Some guys believe job security is more often secured by BS then skill. And then there's the church mouse type who no one thinks much of until one day when he don't show up for work.

Here's a touching story I just received in an email:

Here's a truly heartwarming story about the bond formed between a little 5-year-old girl and some construction workers . This will make you believe that we all can make a difference when we give a child the gift of our time.

A young family moved into a house, next to a vacant lot. One day a construction crew started building a house on the empty lot.

The young family's 5-year-old daughter naturally took an interest in all the activity going on next door and spent much of each day observing the workers..

Eventually, the construction crew, all of them "gems-in-the-rough," more or less, adopted her as a kind of project mascot. They chatted with her, let her sit with them while they had coffee and lunch breaks, and gave her little jobs to do here and there to make her feel important.

At the end of the first week, they even presented her with a pay envelope containing ten dollars. The little girl took this home to her mother, who suggested that she take her ten dollars to the bank to start a savings account.

When the girl and her mom got to the bank, the teller was equally impressed and asked the little girl how she had come by her very own pay check at such a young age. The little girl proudly replied, "I worked last week with a real construction crew building a new house next door to us."

"Oh my goodness!" said the teller, "and will you be working on the house again this week

The little girl replied, "I will, if those a$$holes at Home Depot ever deliver the fxxxxn' sheet rock.."

Kind of brings a tear to the eye - doesn't it?

fasto
01-29-2010, 11:13 PM
I've got a Polish-made VIS level, 0.0005"/8", so similar in accuracy to the 199. Of note, the base is relieved about 0.050" for about 4" in the middle, so that it bears only on the 2" of base on each end. The base also has an involute groove, so there are 4 points of contact when used on a flat surface. The 4 points of contact are scraped and blued nicely when I checked against my surface plate.

I bought this new, so the relieved sections are supposed to be that way.

oldtiffie
01-29-2010, 11:39 PM
As far as I can recall the Hilger & Watts (UK) precision levels and inclinometers were similar.

michaelantipode
01-29-2010, 11:46 PM
Thanks for the info Fasttrack, and J Tiers, I can't find much wrong with your logic. I brought this up because I was just curious about the same condition of my level and that of one belonging to a friend. I suppose I will scrape mine one of these days even though the factory scraping looks pretty good and, unless I hear some better line of argument, in spite of the factory account. My problem is not when I use the full length of the level, but when I need to sit it on smaller areas and can't EXACTLY? center it to account for the hump. I am no rocket scientist nor an old hand machinist, but as my grandmother used to say, "A job worth doing, is worth doing right." I know that one can never actually get things perfectly true, but accepting that the hump is to account for something which isn't straight already just doesn't make much sense to me; why bother scraping it at all?
Michael

oldtiffie
01-30-2010, 02:19 AM
If you read this carefully:
http://www.starrett.com/download/338_p439_444.pdf
you will see that each graduation on the vial is 0.0005"/foot ie ~ 0.00004"/inch.

That condition is true if the level is within or does not exceed or pass beyond the first vial graduation either side of centre ("zero").

So the level would be correct if it were +/- 1 graduation ie the tolerance for accuracy are +/- 0.0005 and the limits are 0.001"/foot apart.

Further, if Starrett were to set the level on 2 rollers,the top of which were very accurately leveled with each other at a set distance apart - say level 1/2 span or at the 1/3 or 1/4 span and ran the curved base over those rollers and the level bubble did not exceed or stayed with the first graduation either side of the zero mark, it would still be in tolerance and would pass QC/QA.

Of course the easy and sensible thing to do would be to calm down and ask/e-mail Starrett during business hours.

It is all too easy to misunderstand or misinterpret all of this, but asking and a cool head will work wonders.

Reading or inferring things into a script that are not there is a lot less than sensible or helpful.

My guess is that most here either inherited their level or were given or bought it from someone other than Starrett.

In that case Starrett owes you nothing and is not obliged to you in any way.

I'd expect that the warranty that went to the original buyer would be a commitment to that original buyer only for a specified period under specified conditions and did and does not extend to his successors in any way.

Perhaps those with a problem with their level should take the matter up and pursue it with the person you got or bought it from.

I'd be interested to see how many here who own one of those Starrett levels (still?) has the original certificate and Owner/s Hand-book/Manual.

I'd be even more interested to see/read a copy of it to see what is in it.

As I see it, Starrett is 100% "clean".

michaelantipode
01-30-2010, 04:46 AM
Tiffie,
I think you are getting your knickers in a twist over nothing.
I have no argument with Starrett. I bought my level used and I accept that the master scraper at Starrett didn't put it in my hands. I never said otherwise. I just wanted to know what sense, if any, there was to a hump in the level. I have not been too fussed by the hump, but then I have only had the level a short while and only started to use it to true up a machine.
The numbers and +/- are not my issue. Honestly, no matter how inconsequential they may be, it doesn't explain why they are there. Regarding warming of the metal, it seems to me that J Tiers makes a pretty good argument to me that the concavity is actually more likely to be accentuated by warmth.
What I do with my level, whether or not I scrape it, has little to do with Starrett or anyone else. After all, it is my level. I initially posted for no other reason than to learn from others, to see if anyone else had noticed this condition. Along the way, I thought that maybe if it were a general condition, maybe there was a reason. Maybe I should have gone to Starrett first, but since I didn't obtain my level from Starret, I thought it a bit rich to ask them first; I wanted to see if my friend and I were the only ones with this situation.
Hey, I didn't go to Starrett first. So arrest me....

oldtiffie
01-30-2010, 06:18 AM
Thanks michaelantipode.

My replies were not just intended for you - at all - although on reflection, it could have seemed that they were.

It was more general answers to general questions in both general and specific cases.

So far as I can see, Starrett has been very co-operative and informative to the - so far as I know - only case where some one has actually asked them and to which Starrett have given an early, direct and unequivocal reply.

I can't see why you could not or should not ask them questions that are for your specific requirements and to expect a similar reply as well.

Some here seem quite upset that Starrett bowed the sole plate without them knowing when all they needed to do was ask.

My guess is that everyting has a tolerance and that a "bow-down" will "rock" where-as a "bow up" will not "rock" that Starrett used a tolerence of "zero down" (ie "flat") to perhaps 0.00025" maximum "up" to make a concave sole plate - which seems pretty sensible and reasonable to me (see later).

So long as the level tolerances are not exceeded Starrett can do as it likes.

I did not see anywhere in the Starrett literature that the sole plate was bowed - or flat - or not. Nor did Starrett say that the sole plate could or should be or not be used as a straight edge.

If people are so concerned about the "bow" they can soon measure it in any number of ways with stuff that many have in their shop/s.

If it were me that was so worried, and I wanted it removed, I'd do it - scraping or (more likely) surface grind it and perhaps scrape it for "looks".

I'd certainly ask Starrett if they could provide me with a manual for it - as I'd be quite prepared to pay for it.

The Starrett - and similar - levels are very accurate so far as a level is concerned but many extol its virtues without realising that its tolerance limits are +/- 0.0005"/foot - ie 0.001" between "HI" and "LO" over a 12" "swing" when they'd possibly want a lot better than that when "true-ing up" a vise or tramming a mill spindle.

As an aside, if the "bow" were "upward" and uniform, and 0.00025" at the mid-point of the 12" level the end away away from the supported end would "drop" 0.0005" at the unsupported end - and still be in tolerance.

I doubt that a maximum of 0.00025" (2 1/2 "tenths") - if that is the case - is going to have too many serious adverse effects.

Frankly, if that level were mine and it "blue-d up" pretty well at the ends on my surface plate and running a good (0.0005"/0.01mm) indicator over it showed no more than a 0.00025" "dip/hollow" I'd be satisfied.

I did get a bit pi$$ed-off with some "old hands" that either didn't (want to?) believe what Starrett had said and resented that it was not flat as they "always knew" it to be.

I "ran" the figures to see and to explain what they did - and didn't - mean.

I think that it is highly likely that your level is very accurate, reliable and use-able.

I can assure you that using a "better" or a "more accurate" level on is own is a total PITA. I have two that pretty well never come out of their boxes.

I prefer to start using a common cheap "bulls-eye" level, followed by an Electricians "torpedo" level followed by a very good precision-machined extruded "box"-section Carpenter/Builders level which has an accuracy of 1:2,000 (0.5mm/metre ~ 0.006"/foot ~ 0.0005"/inch) before I even think of using a "Starrett quality" level.

I rarely need my ("Chinese") "precision" levels.

There is an art in setting up and using a good level.

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance12.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance5.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance6.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance7.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance8.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Precision_grinding/Wheelbalance2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Machinist_Square2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Machinist_Square1.jpg

J Tiers
01-30-2010, 10:15 AM
As an aside, if the "bow" were "upward" and uniform, and 0.00025" at the mid-point of the 12" level the end away away from the supported end would "drop" 0.0005" at the unsupported end - and still be in tolerance.

I doubt that a maximum of 0.00025" (2 1/2 "tenths") - if that is the case - is going to have too many serious adverse effects.



Of course, that dimension is only an assumption, which may or may not be justified. it could be any dimension, actually, so long as it is not disclosed.

I am not terribly worried about it, but I DO think it should be 'disclosed".

While "one graduation" is a convenient unit, it is fairly common to interpolate between graduations, particularly when they are widely spaced. If the true concavity is unknown, produced by merely "scraping heavy" in the middle, or if it is actually a tenth, the numbers may or may not, come out different in a significant amount.

Even if the true concavity is unknown / not controlled, it's existence and amount ("approximately 0.00005" or "less than 0.001", etc, etc) should be disclosed in order to explain what the user will see as far as consistency.

A 12" level with 8" of concave surface in the middle, and a maximum of 0.0001 concavity would have a potential error in level of up to 0.0003 per foot if set across pins. If the concavity is 0.00025, the maximum per foot error is 0.00075. Not a lot, but worth knowing about in a level with a "sensitivity" of 5 tenths per foot, as it corresponds to 1.5 divisions.

Paul Alciatore
01-30-2010, 12:14 PM
This leaves a lot of questions in my mind. First, I really wonder if it is a big to-do about nothing. The sensitivity is given as 0.0005" per foot. This is really a pure number so it is neither English nor metric. It is a ratio: 0.0005/12 = 4.166 x 10^-5. There are no units since the inches in the numerator and denomator cancel. It is a pure number. It can also be 4.166 x 10^-5 mm/mm or 4.166 x 10^-5 miles/mile or 4.166 x 10^-5 light years/light year or4.166 x 10^-5 corn kernals/corn kernals. It is a pure ratio, a pure number: no units except as added for convience. Tenths of an inch per foot is just a way of visualizing it's meaning in use.

OK, 0.0005" per foot. So just how concave are their levels? 0.0001"? 0.0005"? 0.001"? 0.01mm? Have any of you folks who have noticed this feature actually measured it before you scraped it out? Starrett seems to have company in the "not expressing the facts" department. I can guarantee you that, if I purchased a Starrett level and found it to be 0.0001" concave, I would never, never, NEVER try to correct it. Why not? Well, think. If it has a fairly smooth curve to it that is 0.0001" high in the center, then it would be only a fraction of that at any reasonable point where you would rest it on ways that are less than 15 inches (the length of the Starrett level) apart. BUT, it is not the height of the curve at the point where you sit the level that counts, it is the DIFFERENCE in the heights of the two points that it sits on. So, the real question becomes, 1. how symetric is the curve about the center of the level and 2. how much difference is there in the placement of the level from one measured point to the next. We are talking Starrett here and a precision level that many think is the best in the business so I would suspect that the curve they scrape would be fairly symetric. Certainly the variations at a given distance from center will be uniform to 20 or even 10%, perhaps even better. So if the total height of the curve is only 0.0001" then at a point between the center and the ends, it may only be 0.00007" or less. And we are talking about perhaps a 20% variation on that number or perhaps 0.000014 or 14 millionths. There are two such points so we double it for the worst case error (one is high and the other is low) or about 0.00003" or 30 millionths.

The other half of the error is from the placement of the level by the person making the measurement. How much difference would there actually be here? I don't know about you, but I would certainly place the level within about a quarter inch of the same position each time. But lets double that to 1/2 inch. Even the worst case machinist would not exceed that, would you? So 1/2 inch off on each side would move us up and down on the curve. We have a curve that slopes 0.0001" in 7.5" so a simple proportion gives us a change of 13 millionths. Again, we double that because one side is moving up and the other down so the worst case error here is 26 millionths. so the worst case error is 30 + 26 = 56 millionths.

So with a 0.0001" concavity, our sensitivity is 500 millionths and the worst case error from the curve is 56 millionths. This is about 1/10 the sensitivity. Unless you have a separate, very heavy duty foundation for the machine, you are going to get this much variation just by walking around the room. There is certainly no reason to go altering the level at this level of error.

Even at double this level of concavity, it is doubtful that you would ever see the difference. If it is 0.001" concave, perhaps it would actually matter. But even there, consistency on the part of the person using the level could EASILY cancel out any error. Simply placing the level on the ways in the same relative position each time would completely take care of the problem. COMPLETELY!

So, what is the actual amount of concavity that they scrape into the levels?

It could be easily measured on a "factory" scraped level. This should be done on either on a well supported surface plate or on a well mounted machine table with the gibs locked down. Sit the level on a pair of parallels with one parallel at the very edge of the level and the other at the center position. To maintain balance a very small weight should be placed on the end, above the parallel. Place an identical weight above the center parallel for symetry. Also put just enough foam under the unsupported end to take up about 90 - 95% of it's weight so the level does not bow. Observe the bubble. Without moving the parallels, move the level to a centered position on the parallels and again place the small weights above each parallel. Observe the difference in the position of the bubble. The bubble's difference divided by two should be the approximate height of the curve scraped into the level.

So, somebody please measure it. I'd do it, but I don't have a 199. Better yet, lets get measurements from 5 or 10 different 199s so we can get an average and an indication of the range.

Your Old Dog
01-30-2010, 04:26 PM
Paul, that's about how I saw it but could not state it as eloquently as you. A good healthy fart in the shop could raise temperatures enough to create as much error as that bow in the base. If which I still maintain the slight bow works to your favor in allowing one to place things in relative level that are convex enough to keep the ends of the level from resting as they should.

Now, if I hit Forrest's estate sale 50 years from now I'll likley have to rescrape a bit of bow in my cheaply found levels :D

krutch
01-30-2010, 04:42 PM
I missed seeing this post earlier.

That's fine - I can see the logic.

What I can't see is just how bowed it is and from that how far it needs to "bow" under hand heat until it is dead flat - and what the operating temperature range is.

Next of course is if the level needs to be "re-done", just how "bowed" do you scrape it - and how do you actually and accurately verify it?

Send it back to Starrett? At what cost?


I sent a Sterrett level, the one with the tubular cover, in about three years ago. The vial had leaked. Cost me shipping to Sterrett. Returned like new, no charge. Can't say about scraping.
Krutch
I looked it up in my records, and it was a model #98-6 Starrett level. Returned in a new box, cleaned and like new. NO CHARGE! It was the same level, not a replacement. I sent it in for a millwright friend of mine. I'd say that is good service. I am not the original buyer, nor my buddy and I was not asked when I called for a repair order #.
If your level is not 'flat', then go ahead and scrape it. It's yours. If you sell it, then you should disclose what you have done. Then you can claim to be better than Starrett! How often have you not charged, in such a case? BTW, I looked at the Starrett site and it states not to use a level longer than the work. Makes sense if the level has a 'bow'
Krutch 1 Feb. 2k10

oldtiffie
01-30-2010, 05:10 PM
The end-point here - as Paul has said - is that a level - even with a hypothetical flat base is not very accurate anyway as the bubble is both "slow" and "sticky".

A reasonably strong light can "pull" a bubble even when there are no others factors at work. Its a common problem on the 10 or 20 (or "finer") arc seconds alidade and transit leveling bubbles on a Surveyors theodolite and has to be checked and corrected for all the time. It sometimes took up to half an hour for our theodolite to come to and adapt to ambient conditions. You can see this if you shine a bright light on the level vial ("bubble") from opposite ends of the level.

My guess is that Starrett makes its levels and tested them in accordance with an approved testing regime that has a US/NIST/DIN/ISO standard and issues a certificate that the level is in accordance with that standard. Starrett has ISO9000 accreditation.

It should not be beyond the wit of man or anyone who is interested enough here to "go looking" as you would then have a credible reference as a base-line instead or a badly wounded ego or unnecessary or unjustified "worry".

Next, I would not be the least bit surprised or disappointed if Starrett made those level flat enough by machine and just scraped ("mottled") them for "looks".

You can interpolate between vial marks all you like and take your chances but Starrett has only warranted that the level is correct to and at the first index mark.

People either don't need or don't want this sort of stuff that they seem to be demanding of Starrett "as of right" when they don't require it of lathes and mills - or even of surface grinders.

There is far too much quite unjustified awe and mystique attached to Starrett and not a little pi$$ing in ears and pockets too here about it.

The Starrett and similar levels are - in a lot of ways - just an "up-market" version of a Carpenters or Builders level that spends half its life in the sun/weather and in tool-boxes in the back of trucks.

oldtiffie
01-30-2010, 07:15 PM
OK.

Let's just get our heads out of the clouds and our noses out of our own and each others collective ar$es and bring it all down to a realistic shop-level scenario.

Lets say the level is 12" long and that the surface plate (assumed accurate) minimum width is say 4>6" longer than the level at the least dimension so that there is plenty of room for the level anywhere on the surface plate.

Looking at (from the Starrett web site):
http://www.tru-stone.com/pdf/FedSpecGGG-P-463c.pdf

pages 6 and 7 of 26 pages and clauses 3.3.3 and 3.3.4 and Table 111 we see that for 18" x 24" surface plates the surfaces of the plates are to lie on or between parallel surfaces that are spaced at:
Grade AA: 75 micro-inches (0.000075") = 0.75 "tenths"
Grade A: 150 micro-inches (0.000150") = 1.50 "tenths"
Grade B: 300 micro-inches (0.000300") = 3.00 "tenths"

It is all OK if the plate surface is "bowed" up or down or is in "waves" or is flat etc. as all meet the accuracy criteria.

I'd expect that most here will have Grade "B" surface plates so I will stick with that.

The worst case scenario is when the surface plate has a 0.0003" "bow" "down" and the base of the level is also "bowed up" - say 0.0002" and both are additive. In that case there will be a 0.0005" ("half a thou") gap somewhere between the surface plate and the level base and both are (just) within tolerance individually.

The best case scenario is when the surface plate has a 0.0003" "bow up" which matches the 0.0003" "bow up" of the base of the level. Both are bowed 0.0003" but there will be no gap between them.

Both the level and the plate can have "waves" within the tolerance zone or limits with each wave at a different pitch. If all waves on both the plate and the level were of equal pitch and magnitude, they would fit each other precisely whether they fitted "into" or "onto" each other.

There are other scenarios as well but that lot will suffice for the purposes of the discussion.

So, scraping to a surface plate may not be as precise or as accurate as some might have you believe - but on other hand it maybe - or it maybe anywhere in between.

Ideally, if you were scraping to a Grade B level, you should be using a Grade AA or A surface plate to scrape to.

So far as I can see, the "white glove" treatment is probably good but I can't imagine them staying too white - or clean either - when scraping.

Drops of sweat or oil or smears of grease or rubbing your hands or fingers over a surface that is being scraped doesn't help the scraping process - or your patience either!!

Starrett and its proceedures and accuracy are looking better by the minute. I can't see that they are at fault in any way despite the wailing and howls of rage by some here.

So, let's all come down from those higher ethereal planes that some seem to be in and come back to planet earth and reality in the shop.

J Tiers
01-30-2010, 08:56 PM
My guess is that Starrett makes its levels and tested them in accordance with an approved testing regime that has a US/NIST/DIN/ISO standard and issues a certificate that the level is in accordance with that standard. Starrett has ISO9000 accreditation.

You can interpolate between vial marks all you like and take your chances but Starrett has only warranted that the level is correct to and at the first index mark.

People either don't need or don't want this sort of stuff that they seem to be demanding of Starrett "as of right" when they don't require it of lathes and mills - or even of surface grinders.




Ol Tiffster....... It IS NOT about interpolating. And it isn't about some sort of 'right" to great accuracy......(you left out the bit about "for free", which I assume you had intended to add)

If you have 0.00025 concavity, you can have measuring across parallels or pins, as much as 1.5 marked division error from level when level is indicated, if the level is 12" long and has an 8" length of perfectly regular concavity.

yes, you have to put it off-center for this to be fully expressed. Which is necessary in some situations.

I personally don't care, I don't own one and don't want to.

But with a device sold as the best precision level they sell, they should disclose the concavity to prevent confusion.

The foreign level mentioned somewhere above was very reasonable.... they machined out the center, so there was no question. Quite honest, perfectly functional.

If teh Starrett unit is NOT SCRAPED in teh middle, that's fine. But I am perfectly sure it IS scraped, which serves to imply that it is flat. Basic honesty about an otherwise precision device demands that it be disclosed, EVEN IF A PRECISE AMOUNT IS NOT GIVEN.

it would be perfectly acceptable to me if there were marks on the sides to indicate the area within which flatness was not guaranteed.

It isn't so good to withhold information about a precision device. But in Starretts case, they have done it for many years. Look at an old catalog and as far as I can determine, they don't even give approximate sensitivity for any of the levels except the top model.

oldtiffie
01-30-2010, 10:49 PM
Thanks JT.

The whole premise seems to rest on the accuracy of the surface plate as without that any "within tenths" is pretty meaningless.

My plate is a metric Class Grade 00 (Inch Grade AA). It is 4 x as accurate as a "Shop" Grade B plate.

I bought it new because I not only wanted the certificate and accuracy but I really wanted the "map":
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplategradediag1.jpg

As a matter of interest 1 micro-metre is ~ 40 micro-inches.

Clearly, the best position on that plate is along the diagonal from top left to bottom right as there is only a difference (middle) of 0.12um x 40 = ~ 0.48
micro-inches = 0.00000048" = 0.48 millionths of an inch.

The probably worst case is along the back edge where the 0, 1.22 and 0.49 um measurements are.

The magnitude of the "bump" at the 1.22 position on a line between 0 and 0.49um is:
1.22 - (4.9 - 0)/2 = 1.22 - 4.9/2 = 1.22 - 2.45 = = - 1.23um ~ 62 micro-inches ~ 0.000062" = 0.62 "tenths".

So, unless someone has a "map" of their plate to make best use of it, I can't see how they can be definitive as regards anything of a relative or absolute nature as regard comparing one "flat" (the base of the level) with a reference (the surface plate).

I daresay that the Starrett level has - or may have - a "number" that is or may be "NIST-traceable". But who is to say that the "as tested" and the current state of the level are the same - or not. Certainly not without a pretty accurate surface plate - with a "map".

Given that the base of the level may be anywhere between "dead flat" and "humped" "upward" by several "tenths" at least, I can't see why it is as important as a similar certificate would be for a Starrett micrometer etc.

No one seems too concerned about a certificate and test result for a "1 tenth" micrometer - so where is the need or justification for one for the level?

I don't think that Starrett is obliged to keep or send a copy of the test nor do I think that anyone other than the original owner has any right to demand one - if the original owner has such a right either.

PS.

Way way down here in darkest OZ, that bit of rock works real good as a base for stuff on the 20 ton press and as an anvil - with my big, big BFH. I use the levels for setting the bird-baths level as most times the water levels in a hose are more than good enough for setting up stuff in the shop.

J Tiers
01-31-2010, 10:06 AM
Thanks JT.

The whole premise seems to rest on the accuracy of the surface plate as without that any "within tenths" is pretty meaningless.


What has a surface plate got to do with it? I never mentioned surface plates, we're talking about a level here. I'm sure your analysis is very nice, although I have not read it, but unfortunately it has no relevance to this discussion.


I already said it isn't likely to be a big deal, but it could be a problem.

A machine with certain sorts of surfaces that are intended to be leveled parallel might end up "out" by more than one division on a 12" level if the concavity is 0.00025, and the middle 8" of the level is concave. Whether that is important or not just depends on what you are doing,. doesn't it?

YOUR point seems to be that "other things are probably out by more than that so it's irrelevant".

MY point is that they should disclose the concavity, instead of hiding it. Merely saying verbally to an inquiry that they "make it concave" is not disclosure.

After all, here is a "precision" device, the best the company makes. It has a certain "sensitivity, and certain tolerances, unknowns, etc.

Starrett HIDES, or fails to mention, this feature of the device, which affects its performance potentially by an amount over one graduation on it's "output display".

A DMM is given an accuracy statement, which for a (formerly, at least) decent company like Fluke is fully disclosed as to both basic accuracy and uncertainty. And the amount is similar, one or so "divisions" on the output display.

Why should Starrett not at least disclose the size of the area of concavity, and make some statement of its amount?

Why do they not even reference the standard you mentioned?

It seems so SIMPLE to just do that. Since they do not, what else are they hiding?


BTW, the 199 is 15" long... several people had given that as 12", I believe.

http://www.starrett.com/download/338_p439_444.pdf

Forrest Addy
01-31-2010, 12:55 PM
Tiff, you're arguing all around the issue. JR articulated my beef perfectly: Starrett has until now failed to disclose deliberate concavity in a precision tool; a refinement possibly having an effect on the use of the level. Period.

There's nothing wrong with the level itself except that the sole has been supplied with the concavity. If assymetrically supported the concavity will definitely have an effect of the finished leveling job. The level registers 0.0005" in 10" per graduation. Using reversal technique you can reliably improve the sensitivity of the level to 1/10 the graduated accuracy (I use a hair or a thread to mark the end of the bubble) maybe better. There have been many times I've done just that and days later someone else has verifed the leveling job using the same technique with a different level.

So 0.0005"/10" is not the working limit of senstivity. Nope that's the graduation interval. With good technique and care you can work repeatably to 0.000050" in 10". That's 1 in 500,000 slope or 1 arc second roughly. You certainly don't need this level of refnement around the home shop but there have been many a floor mill runway resurrected with skillful leveling.

Another point is, yes, I've scraped several level soles flat but when I've had two or more, I scraped them serially on the same surface plate. Then, just for grins, I've blued up one and tried it aganst the other. Simultaneous generation of surfaces anyone? The result was a little touch up and the level soles were if anything a bit flatter than the surface plate.

Mcgyver
01-31-2010, 01:29 PM
You certainly don't need this level of refinement around the home shop but there have been many a floor mill runway resurrected with skillful leveling.


I think we should refine that colloquialism about whats required in the home shop. maybe make it a "around a machine shop", or "small job shop" or "general machine shop". It's the job at hand that dictates what's require, not the location. I know what you meant, but he assumption that because its done at a home shop that its less exacting is ever so slightly insulting; I can't always produce the same quality as a high end tool room or jig grinding shop but i can a that of the run of mill job shop.

lazlo
01-31-2010, 02:08 PM
I think we should refine that colloquialism about whats required in the home shop. maybe make it a "around a machine shop", or "small job shop" or "general machine shop". It's the job at hand that dictates what's require, not the location.

That's hopeless. Home-shop precision ranges from "Ah ain't making parts for NASA" to measuring microns :)


I know what you meant, but he assumption that because its done at a home shop that its less exacting is ever so slightly insulting

Agreed. Each person has their own criteria for precision. For some folks, they're just trying to get the lawnmower working. Other folks want their model engine to run. Still other folks enjoy precision as a challenge, including rebuilding and/or re-scraping old machine tools back into spec.

Some here seem to have the opinion that if it's good enough for them, it should be good enough for you. Sit and spin. :rolleyes:

Mcgyver
01-31-2010, 02:59 PM
That's hopeless. Home-shop precision ranges from "Ah ain't making parts for NASA" to measuring microns :)


you're right, its hopeless, but thought the point should be made. I don't think Forrest intended in any way to malign, but the point should be made; this default assumption that the home shop is only capable of rough work is just plain wrong. take a look at what guys like Bill Huxhold or Rich Carlstedt turn out. Bill for example works in a temp controlled environment and running machines like a Hardinge has more quality in his basement than many job shops

Microns? i work to microns all the time, I just can't be bothered keeping track of them all

Forrest Addy
01-31-2010, 04:58 PM
No slur intended. I was reacting to the "what's good enough for me is good enough for the rest of the world" crowd - whose rant is a slur on the sanity of those having requirements not only different but reasonable and well justified.

I often sneer at hot-rodders, shade tree mechanics, ignorant-ass distorters of technical information, nuts, screwballs, single issue zealots of every description, knee jerk reactors to current events, etc. I ruthlessly scold my family members when they repeat propaganda, rumor, etc without fact checking before blurting.

No, I don't have many friends and the ones I do have are nt only patient but have hides like rhinoceri.

Getting back to the topic of home shops, there are many among us who have them and many of them produce the occasional marvel. The rest of us mess around once in a while completing a project, fixing and overhauling etc. My current task is prepping project pieces and making scrapers for my coming scraping class.

Getting back to the OP, my views have been stated. Starrett failed to inform buyers of their #199 level of the non-conventional technique employed in scraping its sole.

Paul Alciatore
01-31-2010, 06:35 PM
Well, unless I missed something here, no one took me up on the challenge to actually measure the bow in a "factory" Starrett level.

J Tiers
01-31-2010, 08:24 PM
If I had one, I'd do it.

But I don't have one, and I'm not going to buy one out of curiosity, so it will need to be someone else.

I'd measure my 98, but that wouldn't help, and besides, I scraped that flat already (aside from the "V", which I could do without quite nicely).

Greg Q
01-31-2010, 08:24 PM
A couple of things:

Is the reply from some guy at Starrett the oficial party line, or just some convenient bull****? Perhaps the real reason for the slight concavity is a wear allowance, or the slightest deviation from the flat condition towards the concave to ensure no convexity (an insurance practice effectively)

My 199 is concave to be sure, but it seems to be on the order of one scraping cycle...maybe 0.0001-0.0002" deep.

I will scrape my 199 flat before I use it again. It will be more useful that way, and I am confident of doing so to 50 millionths or the limit of my lab grade granite plate.

Having an affordable access to the local gravity vector provides me with a handy and reliable reference standard from which much information can be derived. While I also have a French made, involute grooved base model, I am more pleased to own a 199.

As an aside, I wonder how the Sherr-Tumico level base compares?

The Starrett is an old and venerable tool, pre-dating ISO9000 and the like. It dates from an era when cosmetic scraping would have been scorned. For those who have never seen one, the scraped surface is a wonderful example of how scraping should look.

Fasttrack
01-31-2010, 10:16 PM
I often sneer at hot-rodders, shade tree mechanics, ignorant-ass distorters of technical information, nuts, screwballs, single issue zealots of every description, knee jerk reactors to current events, etc. I ruthlessly scold my family members when they repeat propaganda, rumor, etc without fact checking before blurting.




LOL - good on you, Forrest. I know I've occassionally perpetrated the sin of "repeating propaganda, rumor, etc without fact checking" but I try to keep that to a minimum. It really pisses me off when other people do it :D





Is the reply from some guy at Starrett the oficial party line, or just some Perhaps the real reason for the slight concavity is a wear allowance, or the slightest deviation from the flat condition towards the concave to ensure no convexity (an insurance practice effectively)


That is how I read the reply. I think the heat argument is mostly BS for reasons already discussed. I do believe, however, that the concavity is intentional and is there as an "iunsurance" policy against convexity.

duckman
02-01-2010, 12:10 AM
I used to work at a company that scraped the Starrett level bodies, we did it when Starrett got behind , the guy that scraped them would never scrape any thing but flat and as close to perfect as any body could . Scraping 200 spots per inch was nothing for this fanatic. Unfortunetly that co. is now closed due to mismanagement.

oldtiffie
02-01-2010, 12:23 AM
OK.

Here is the OP - a very good one in my opinion - at:
http://bbs.homeshopmachinist.net/showpost.php?p=514631&postcount=1



There was a discussion here a few weeks ago about the Starrett no. 199 master precision level.

(http://bbs.homeshopmachinist.net/showthread.php?t=39090)


YOD suggested that the levels are built with a degree of concavity on purpose and that they should NOT be scraped flat.

Well I contacted Starrett yesterday and got the following response:


Originally Posted by Paul Nault (Starrett Tech Support)
Yes they are made concave. If you scrap it flat, the heat of your hand could cause it to rock as the metal expanded. If the scrape marks look good, I would suggest leaving it alone.

Note the response time - which seems to be of the order of 24 hours.

krutch at:
http://bbs.homeshopmachinist.net/showpost.php?p=515489&postcount=29
seems to have got a very good result from Starrett just by asking too:



Originally Posted by oldtiffie
I missed seeing this post earlier.

That's fine - I can see the logic.

What I can't see is just how bowed it is and from that how far it needs to "bow" under hand heat until it is dead flat - and what the operating temperature range is.

Next of course is if the level needs to be "re-done", just how "bowed" do you scrape it - and how do you actually and accurately verify it?

Send it back to Starrett? At what cost?


I sent a Sterrett level, the one with the tubular cover, in about three years ago. The vial had leaked. Cost me shipping to Sterrett. Returned like new, no charge. Can't say about scraping.
Krutch

So Starrett gave us all some very sound advice that was pretty well along the lines of "if it looks OK it should be OK - and leave it alone and don't fiddle (or fu*k around) with it".

That makes an awful lot of sense to me.

And if I had a Starrett 199 level?

I'd do just as Starrett advised in that email as it made eminent sense to me.

If some here who "know better" or "demand" to "be informed" and all but bash on the Starrett door or on "Management's?" desk (sure - nothing but the top for you eh??) as your "rights" have been trodden on and the integrity of your "knowledge", "image" and "reputation" has been impugned etc. etc.

Perhaps you "lost your innocence" - as it were?

Or is it that the once infallible and omnipotent "Starrett" that you put on a pedestal as an icon and worshiped at the feet of is a bit (a lot?) less credible because they don't show you that awe and "respect" that some seem to feel that they have an "entitlement" to and are "owed" and hence can "command" and "demand" "as of right"?

Or as it seems here that as Starrett = God and some here can not only demand that God accede to their wishes that God is not only subservient to them but is accountable and sub-ordinate to them.

Wow!!!!

Such machismo, such testosterone, such authority.

and

Such Bull-$hit.

There is a whole lot more talking about the point and little if anything of anyone actually doing anything about it.

Why doesn't some-one? - anyone? (no one??) of you who cares so much or is concerned so much actually email Starrett and ask them. The precedent has been set and Starrett were prompt, precise and polite about it.

Post the result here and we will all be better informed even if none the wiser or better off.

How is or can it be that so many of such competence and skill and investigative skills and pursuers of Holy Grails, who confirm everything and assume nothing have never got around to actually noticing, let alone measuring at least one and preferably several Starrett 199 levels to actually determine the magnitude and frequency of the curved bases?

I've asked several times for some one to post a copy of the original hand-book/manual that came with a new Starrett 199 level to post a copy of it here. It is conspicuous by its absence up to date. Perhaps no one has one. Perhaps it doesn't say that the sole is flat or bowed or to what extent and to what limits or tolerance - perhaps it does? But we won't know until we see that hand-book/manual.

To that extent its similar to my asking for one to be posted on the "Blake" co-axial indicator thread too, as despite my asking there as well, the only one that was posted was the one that came with my new "Chinese" much-derided "fake Blake".

It seems reasonable to me for me to assume that very few - if any have that 199 hand-book/manual or have seen or read one either as no body has sen fit to say what is in it or posted a copy of it as asked.

Has anybody actually asked Starrett about the extent of the "bow" and/or for a copy of the manual?

If not - why not?

And does it really matter?

There has to be a tolerance and limits on everything - no matter how exact or how big or small or important or unimportant the "thing" is - Starrett 199 levels included.

Starrett only said that the base was "curved". It did NOT say by how much. It did NOT say that there was only one curve either - there could have been several or many.

Some here - for what ever reason - have or seemed to have deduced or decided that there is only one curve that is a geometrical and mathematically regular (ie part of a circle, parabola, conic section etc.) form that goes from end to end of the level sole plate and is in effect an arc subtended (spanned end to end) by the perfect/ideal/straight sole plate that acts as a chord to that arc etc.

Starrett certainly did not say that.

If that perfect arc and that perfect chord were the tolerance and limits, it would not matter how many regular or irregular "shapes" such as "waves" or "lumps" or "bumps" there were as long as they stay within Starrett's limits - all are OK.

I suggest that some here could do well to stop reading things that do not exist into or out of what is written so as to have it be - or seem to be - as they want it to read or to be.

It does not help if you/"they" then compound it by actually saying that what the text/manual said is what they think - and perhaps believe - the text did say when it may not have said that at all.

I have no objection at all to anyone working at any level that he chooses to suit himself for his own reasons - at all.

Some like the job to be "closer" or "look nicer" etc. etc. and to have "better" "more accurate " tools and machines etc. etc. as is their inalienable right here.

Some just like the job to be "good enough".

If both meet their respective objectives then one is as good as the other.

But to criticise others and infer that they or their tools or machines and/or what they make of do with them because it is "not (as?) good" as their own is "a bit rich".

I am continually amazed here at just how well some who are new to machining and/or who have "not as good" tools and machines and limited resources actually do so very well.

(Continued next page/post - page 2 of 2)

oldtiffie
02-01-2010, 12:24 AM
(Page 2 of 2)

Now back to Forrest's lot:



.................................................. ......

Getting back to the OP, my views have been stated. Starrett failed to inform buyers of their #199 level of the non-conventional technique employed in scraping its sole.

Starrett didn't "fail" at all as it didn't have to - and didn't - for what-ever reason.

And just how and when are you going to actually "do something" to see that that "failure" by Starrett is rectified and that they are made to be and are
to be seen to be accountable to you?



Originally Posted by oldtiffie
Thanks JT.

The whole premise seems to rest on the accuracy of the surface plate as without that any "within tenths" is pretty meaningless.


What has a surface plate got to do with it? I never mentioned surface plates, we're talking about a level here. I'm sure your analysis is very nice, although I have not read it, but unfortunately it has no relevance to this discussion.

................................................
............................................


If you have a 199 level or what-ever and need to test its "flatness" a surface plate and its assumed accuracy is a pretty good start.

I thought my analysis was "pretty nice" too.

I'm damned if I can see how you can both say that you have not read it yet pronounce it to "have no relevance ........... ".

Real "head-scratcher" that one.

I was wondering who it was that lifted their skirt and flounced off with a "sniff" and (her?) brollie (parasol?). I thought it was Mrs. Grundy.
http://en.wikipedia.org/wiki/Mrs_Grundy

It seems that I was mistaken.


Tiff, you're arguing all around the issue. JR articulated my beef perfectly: Starrett has until now failed to disclose deliberate concavity in a precision tool; a refinement possibly having an effect on the use of the level. Period.

There's nothing wrong with the level itself except that the sole has been supplied with the concavity. If assymetrically supported the concavity will definitely have an effect of the finished leveling job. The level registers 0.0005" in 10" per graduation. Using reversal technique you can reliably improve the sensitivity of the level to 1/10 the graduated accuracy (I use a hair or a thread to mark the end of the bubble) maybe better. There have been many times I've done just that and days later someone else has verifed the leveling job using the same technique with a different level.

So 0.0005"/10" is not the working limit of senstivity. Nope that's the graduation interval. With good technique and care you can work repeatably to 0.000050" in 10". That's 1 in 500,000 slope or 1 arc second roughly. You certainly don't need this level of refnement around the home shop but there have been many a floor mill runway resurrected with skillful leveling.

Another point is, yes, I've scraped several level soles flat but when I've had two or more, I scraped them serially on the same surface plate. Then, just for grins, I've blued up one and tried it aganst the other. Simultaneous generation of surfaces anyone? The result was a little touch up and the level soles were if anything a bit flatter than the surface plate.

Forrest.

I note your concerns.

If you have three 199 levels and use them in the classic "3 plate" scraping method to achieve optimum mutual accuracy - good for you.

My concern would be that as all are equal and were all manufactured in the same environment that one should be stored in that environment as the "reference flat" (the Grade AA?) for the other two, one of which would be used for "inspection" (the Grade A?) for work carried out by the "shop" level (Grade B).

Now if the B needed to be re-scraped then all would need to be re-scraped.

Given that all are scraped to be identical there is no separate reference to determine just how "flat" the "flatness" of the three 199 bases is/are.

Many here don't have a Starrett 199 level or its equivalent let alone three of them and can spare two of them and just use one for "B" work.

Further, now that Starrett has "let the cat out of the bag" - so to speak, many will be quite adequately assured (by Starrett) that leaving their 199 level "as is" will be adequate for their needs and will leave - and use - it "as is".

I have yet to see any good reason for anyone to say definitively that the Starrett 199 level "as shipped" (or "looking good enough") is not in fact "good enough".

I am not arguing the right of the owner to "re-work" it and to "make it better" as and if he chooses - as it is his tool and his right to do as he likes with it.

Greg Q
02-01-2010, 04:29 AM
Just for kicks I looked up my Starrett catalogue (1976 ed), 15th "Machinery's",
Moore Special Tool's "Holes Contours and Surfaces", "The Fundamentals of Dimensional Metrology*" (Ted Busch) and "The Handbook Of Dimensional Measurment" (FT Farago), the local and Henry Ford school texts and "David Berkowitz on Manners"

Starrett sez this:

-Ground and graduated main vial of 10-second accuracy, one division= 0.0005"/foot

-Special Alloy Iron employed to Obtain Freedom from Thermal Effects

-Castings thoroughly seasoned, machined and scraped.

-Insulation from handling by means of a top plate of non-conductive material.

Moore is mute on the subject, as is Machinery's and the trade school books.

The two metrology texts have chapters devoted to the tool, and nowhere is there a mention of any deliberate concavity of the tools' soles. Indeed, there is ample mention of using parallels to gauge a distance less*** than the 15" body of the tool-a practice likely to end in frustration if the tool isn't dead flat.

***this of course introduces the dreaded Blake indicator lever arm sensitivity conundrum effect.:p

*"Fundamentals of dimensional metrology", Library of Congress cat # 64-12593 is the best single metal shop text book of the many that I own.

Greg

oldtiffie
02-01-2010, 05:24 AM
Greg.

It gets better and betterer - or sumfink.

So far all that has been considered/discussed is a level on a flat surface.

If the surface was concave the ends of the level would be in contact with the arc and perhaps no one would be the wiser.

If the surface was convex only one part of the base of the level would be in contact with the arc and so when does the inevitable rocking become evident - and where is the single common contact point on the arc and the flat base of the level?

For those small amounts that the level vial may not "see" or "pick up", perhaps the salvation may well be a very small "hollow" arc on the "flat" base of the level. A "rise" of say 1>3 "tenths" between the arc and the chord may be of practical benefit after all.

If the level were riding on two "bumps" that were level and the rest of the job being checked just "dropped away" from the base of the level, the level would or may not "pick it up".

I am not at all certain that an otherwise "correct" surface and/or level might not "move" and register as a "tilt" as the level vial is referenced to gravity.

There are others.

I have toyed with the idea of emailing Starrett and asking them to advise on matters of concern here by some who seem to think that Starrett "owes" them and is or maybe in infringement of the(ir) "right/s" and must accede to the(ir) "demands".

I might ask Starrett for a copy of their manual/hand-book for the 199 level too as a lot of others seem to "know" what is in it but are - at best - "reluctant" to share it with the rest of us.

darryl
02-01-2010, 05:43 AM
And here all this time I've been using one of those levels as a straightedge to cut my pancakes straight. No wonder they tasted funny :)

Greg Q
02-01-2010, 05:43 AM
Yeah, but you are talking about variations in surfaces from flat, which is the province of spotting tools to quantify and map. The level as I understand it is simply used to qualify the plane (often imaginary) between two points and compare that against the independent and always calibrated gravity vector.

My own use of the precision level is as often as not on parallels spaced at 8"-9"and as such need the certainty of a dead flat sole.

The care and calibration of shop references is an ongoing responsibility in my book-we calibrate our measuring tools, and having to touch up flat surfaces made from iron is just part of the deal. My B & S camel back and other masters all need a periodic fine tuning. Adding the levels to the "to-do" list is really no big deal.

Greg

oldtiffie
02-01-2010, 06:15 AM
Agreed Greg.

The level is but a bridge between two points. Is a matter of how flat the base of the level is and where it bears on those two points.

We are now getting into "metrology" level (sorry) stuff where everything is controlled and checked and counter-checked for optimum accuracy and consistency.

But how that is transferred to the shop and how effective and consistent it is there is quite another matter.

If some feel it necessary - for what-ever reasons - to set up or check stuff in the shop to the n^th degree - they have my best wishes.

But if some of them infer that because they do or can do it that everybody should or must do likwise is "laying it on a bit thick" - and bloody patronising and insulting - as it is everybody's right to seek or take advice and use it or not as they wish. They are accountable or answerable to nobody (else).

To each his own.

I like to see excellence in work, aptitude and attitude in the shop.

The level of excellence is limited by a whole raft of variables - and there-in lies a challenge.

Knowing and having the appropriate level of excellence for a particular shop activity and applying and achieving it is a skill not easily learned either.

A range of levels is well within this category as well. My "lower order" ones get quite a work-out and the "higher up" levels get increasingly less (???) use.

In my mind, a skilled glass-blower is something that I could watch for hours. There are lots of others.

J Tiers
02-01-2010, 09:29 AM
I thought my analysis was "pretty nice" too.

I'm damned if I can see how you can both say that you have not read it yet pronounce it to "have no relevance ........... ".

Real "head-scratcher" that one.


And obviously the answer here is that you are ANSWERING THE WRONG QUESTION.

The whole premises of your replies, as usual, is apparently summarized as:

1) you have no right to any particular accuracy from Starrett

2) The difference that is made by this issue is of no relevance because you will <probably> never need that accuracy and wouldn't know it if you saw it.

3) they didn't SAY it IS flat, so it doesn't have to be.

This is merely "sea lawyer" logic chopping..... whether "technically and legally" true or not.

And of course you missed the point.

Check the link I gave to the device. It implies all over the place that this device is accurate to XXX, and describes the underside as a "reference surface".

Now, I don't know how many facets or curves YOU think a "reference surface" has. I consider that it should be flat enough to not adversely affect the operation of the device, at a minimum. Adversely meaning to degrade the accuracy.

http://www.starrett.com/download/338_p439_444.pdf

it is obvious to one 'skilled in the art", that such a level may be used in all sorts of setups, over pins or parallels, on flat surfaces, etc.

And it is obvious that a deliberate concave surface of unstated concavity has the potential to degrade that accuracy when use in the various possible AND NOT DISALLOWED ways.

You your ownself suggested 2.5 tenths, I believe, and I have shown that such a concavity can affect teh level by over one division. yet one division is the sensitivity.

What we have here is a device that is purported to be the best accurate level...... But what is really happening is that Starrett GIVES NO SPECS beyond the basic accuracy statement.

Apparently they do some unmentioned compensations, adjustments, etc, and tell teh user, "nice user, don't worry your head about these things, we've taken care of them and are helping you".......... This is not helpful.

All it takes is a full disclosure of the accuracy, of the level accuracy, and of the "reference surface".......

If the old HP, or Fluke, etc made that tool, there would be a page of specs under various conditions, giving full limits and etc.

Starrett chooses to simply say in effect, "we are the best, and you should just not worry your head about it".

Rather than STARRETT being "petted" for not telling you you CAN do that, they should be "slapped" for NOT GIVING THE USAGE RESTRICTIONS THAT APPARENTLY EXIST.

it's not a big deal, until you really NEED the accuracy they claim......

oldtiffie
02-01-2010, 03:20 PM
JT.

All that is required of you to resolve the issue with Starrett is for you to email Starrett.

If, for what-ever reason you can't bring yourself to "demand and expect" - and get - an answer to your liking from someone (Starrett?) that you don't (now) "like", then I am pretty sure you won't have to lower your dignity (and your sights) if you just simply "ASK" them. Just don't let them trample on your rights and "dis-respect" you. Don't let 'em ignore you or push you around as you've got "rights".

I daresay Starrett will not expect you to humble yourself on bended knee at all as I'd expect Starrett to be quite civil, polite and accommodating in giving you a correct answer.

Same applies to any other "concerned" (out-raged?) Starrett owners.

I don't have - and don't need - a Starrett level of any type (199 included) - but I'd like to have one. If it looked OK, I'd leave it "as is" - as Starrett suggested - and get on with my life. If it didn't "look OK" I'd check it on my "AA" Chinese surface plate and take what ever action - if any - was needed on the Starrett level or else I'd email Starrett (OZ) and see about Starrett having any necessary restorative work done - at my expense. I'd ask for a copy of the hand-book manual too - and pay for it if necessary.

If I were to consider "doing it up" myself - if it needed "touching up" - I'd do it - but not until I'd "touched up" the blonde down the road. I will need a pretty big half-round or 3-point (definitely not a flat) scraper for that - for a big "bearing" job - there. Perhaps I'll need a "Hilti" (jack-hammer) - a big one - first though.

Then back to the 199.

Lottsa water and use a classic "figure of eight" motion. First of all - with the wood-troweled finish on the shop floor slab - and for that bit of "class" I'd finish off on the steel-troweled part of the floor slab. A few nice "swirls" with the angle-grinder would be a "nice touch" too.

I might even hand-lap it - on a brick.

Gotta keep up "standards" here under the shade tree in our (my?) part of Hillbilly land here in OZ.

One must "keep up appearances" - mustn't one?

michaelantipode
02-01-2010, 10:31 PM
I sent a note to Starrett tech support asking about the level.

This is the reply:

Michael,



Yes, these are made intentionally with a .0002" concavity in the base.
All machinist levels are made with this. Other levels have more than
the .0002". The reason is if there were none, when a level is picked
up, the heat of your hand can cause the metal to expand (usually in the
center) and cause the level to rock.



Hope this settles this for you.

Best regards,

Paul Nault
Technical Support Department
Starrett
978-249-3551 x178
pnault@starrett.com

Still not sure I understand the physics of this, but I guess it depends on how and where one grabs the thing.

oldtiffie
02-01-2010, 10:45 PM
Thanks mightily Michael.

You are the only one who took up the challenge and had the "guts and nuts" - and eminent good sense - to ask Starrett and to post the reply.

Now extending - but keeping to - Starrett's logic, it would seem that those that scrape/d the sole/base of their (199) levels may well be doing themselves a dis-service.

Note - again - that Starrett did not say what shape the curve is or even whether it is symmetrical.

A highly satisfying and gratifying out-come.

Again - many thanks.

Greg Q
02-01-2010, 11:25 PM
The reply from Starrett came from the same person as the OP's reply. There is no new information there, so sorry.

After reading that, I did the following: I went and cradled my level in my hand for four minutes, then checked the markings on the surface plate with canode red spotting dye. The resulting markings were the same as prior to the intentional mishandling.

I measured the temperature with a calibrated 3M Heat Tracer...the readings from left-middle-right on the tool botom were 26C/34C/27C

Next, I assaulted the tool with a heat gun, gently ramping up the temp in the middle to 50C. Again, there was no change in the resulting markings. with indicated temps 29/51/30

My level is hollow by about 0.0002" +/- 0.0001. You would have to pretty hot stuff to have body heat to distort the casting to the degree under discussion.

That's it, I'm outta here.

Greg

J Tiers
02-01-2010, 11:44 PM
Thanks mightily Michael.

You are the only one who took up the challenge and had the "guts and nuts" - and eminent good sense - to ask Starrett and to post the reply.

Now extending - but keeping to - Starrett's logic, it would seem that those that scrape/d the sole/base of their (199) levels may well be doing themselves a dis-service.

Note - again - that Starrett did not say what shape the curve is or even whether it is symmetrical.

A highly satisfying and gratifying out-come.

Again - many thanks.

Except that it is at least in part the merest moonshine......... We disposed of this "answer" a long time ago in this same thread, and the answer is obvious BS as to the reason, although it may be accurate as to the amount.

I won't bother with anything further on the insulting and swipey "guts" comments..... which with the "indignant rights" comments add nothing to the discussion except some political input......

They obviously scrape it concave to make it sit flat. The "heat of hand" stuff is just BS and moonshine, presumably intended to provide a logical reason. They could just as well simply state the concavity and state that it is to ensure it sits flat. I certainly won't think any less of them on account of it.

And it means that the level will ONLY be accurate to one division if both ends of the 15" (381mm) length are on the surface being "tried". In many cases that could be difficult. With a concavity, the outer part of the level will "dip" by a certain amount depending on how much of the bottom is concave, and how much of the bottom is on the surface.

Even if the level is 'centered", as is pointed out, the concavity is not any shape, and so the outcome if the "ends" are not both on the surface, is unknown and undefined.

Well, we know the same information that we had several pages ago.... not much more, other than a "number" of 2 tenths that represents the approximate concavity maximum...... This communication has added no significant new facts.

As for the "heat of hand" and the idea that scraping flat is a bad thing.......... it's nonsense.

The 199 has a heat-insulating "handle" to avoid heat effects...... the blurb states "insulation from handling by means of top plate of nonconductive material."

However, the top is NOT the problem......... The heat of the hand, if the level is used by anyone having the slightest amount of sense, will be input on the TOP, where it is held, which will, as we went through before, bow the level DOWN MORE, ACCENTUATING the concavity.

Only a remarkably stupid individual who carries the level around resting the middle of it on the palm of his hand, can get any possible benefit of the concavity on a heat basis. And from Greg's tests, that may be pretty minimal or nonexistent.

I suspect Greg's test came out that way because the metal conducted the heat and ended up the same temp overall anyway, so there was no "differential" expansion. Which is about what you'd expect.

If it is a *widely known* feature, as they seem to imply ("all manufacturers do this") then why is it not a *plainly stated* feature?

But if Forrest didn't know it, after presumably using the things for 40 years, it can't be that widely known.

It seems so simple to just be clear and plain about it so that everyone will know and understand that.

or better, just cut away the middle section, so there is no doubt what part is flat and what relieved, like the european level mentioned earlier.

Well, we were given a rough figure as to the amount, which is pretty much what we had assumed, and isn't precise enough to do anything much with.............and are left to guess how much of the base is really flat..... nothing between the ends, I guess......

oldtiffie
02-02-2010, 12:01 AM
It seems that what-ever the effect of heating the level (seemingly none to perhaps 0.0002"?) it seems that while the level base may move it will not pass "flat" (and so become convex) but will or may vary between "flat" and concave to a maximum of 0.0002".

If taken over the length of the level base it is at or better than a lot of surface plates.

I'd need to check what Grade but I will leave that for now.

That alone would make it pretty hard to scrape it any flatter on a normal "shop" (Class "B") surface plate.

I am satisfied with the outcome as regards the 199 level.

I'd still like to see a copy of the hand-book though.

I would not be the least bit surprised if Starrett were "reading" this thread. I hope they are and if they are that they post me the images of the 199 hand-book and approval for me to post them here and I will post them.

(For: Starrett:

if you are "reading" this please email me at mickgatsatlinkdotcomdotau

Thanks)

Greg Q
02-02-2010, 12:12 AM
What hand book? There is an instruction card in the box lid for the benefit of people who failed fifth grade, little elaboration in the catalogue, and little worthwhile in the web link.

With regard to scraping a level on a surface plate: Tiffie, you should be able to get to less than 0.0001". After a little bit of instruction at the last scraping class we were able to get to half of that or better: 50 millionths. Moore Tool speaks of routine scraping to 20 millionths, albeit in skilled hands.

I was able to confirm the temp plot as I described-there was a definite temp delta across the tool base and at least 22 C difference over 7 inches.

WRT to posting pics: This site won't let me. When it does, I will. Not interested in a huge web exercise to prove an obvious point.

J Tiers
02-02-2010, 12:26 AM
Greg..... to get a difference, the TOP of the level needs to be at a different temp than the bottom.

Differences across the length will have minimal effect, I'd suspect.

Greg Q
02-02-2010, 12:32 AM
But wouldn't expansion at the top make the bow more pronounced? By applying differential heat in the middle of the base I was trying to get the bow to unbend. I thought that my technique was the worse case scenario.

Greg

Paul Alciatore
02-02-2010, 12:56 AM
Well, we finally got an actual number, 0.0002 +/- 0.0001". So, interpolating from my previous analysis, we are talking about 40% of the stated sensitivity as the maximum error from this curve that is scraped into the level.

OK, I can see how some may want better than that. It is still quite small, but not the 20% in my example calculation.

What I would really like now is an explanation of how handling a level will make it convex on the bottom. I mean, if you grab the top, then the top will expand and the bottom will become concave. You would have to work hard to grab it by the bottom. I mean, first grab the top and then use the other hand to grab the bottom? Who would do that? I mean, really, WHO?

So if you are reading this Starrett, please explain. Please!

oldtiffie
02-02-2010, 01:41 AM
What hand book? There is an instruction card in the box lid for the benefit of people who failed fifth grade, little elaboration in the catalogue, and little worthwhile in the web link.

With regard to scraping a level on a surface plate: Tiffie, you should be able to get to less than 0.0001". After a little bit of instruction at the last scraping class we were able to get to half of that or better: 50 millionths. Moore Tool speaks of routine scraping to 20 millionths, albeit in skilled hands.

I was able to confirm the temp plot as I described-there was a definite temp delta across the tool base and at least 22 C difference over 7 inches.

WRT to posting pics: This site won't let me. When it does, I will. Not interested in a huge web exercise to prove an obvious point.

Thanks Greg.

That closes the loop - no hand-book - that's another item off the list.

All your points are noted and agreed as regards flatness, thermal effects etc.

If I had to scrape it - I would - although I'd be inclined to put it over the surface grinder first.

As you say, getting to 0.0001" (100 millionths ~ 2.5 um) is quite possible and perhaps half of that with the right tools and environment.

From memory - I need to check - a Grade B plate has a tolerance of 3um ~ 120 millionths so it will need to be a pretty good and well-mapped Grade B plate to do the job. My guess is that we are looking at Grade AA (laboratory grade) to be sure.

OK - I've checked. See page 6 of 26 Table 11:
http://www.tru-stone.com/pdf/FedSpecGGG-P-463c.pdf

For a table diagonal of up to 30" the allowable tolerances for repeat measurements are:
Grade AA: 35 micro-inches (~ 0.9um)

Grade A: 60 micro-inches (~ 1.5um)

Grade B: 110 micro-inches (~ 2.75um)

See Table 111 - Total Flatness Tolerance in micro-inches (page 7 of 26).

For an 18" x 24" plate the flatness tolerances are:

Grade AA: 75 micro-inches (~ 1.9um)

Grade A: 150 micro-inches (~ 3.8um)

Grade B: 300 micro-inches (~ 7.5um)

So the only 18" x 24" plate that will do the job for 100 micro-inches (0.0001") is a Grade AA (the "top of the range). It may squeeze 75u" at a pinch but getting to 50 will require some very high class work.

Starrett have pretty well said that there is a maximum rise/gap above "true flat" of 0.0002" between the end points and zero at the end points.

I can't see me needing that - or better - although in my circumstances it will be more "nice" than "need" to have so would suit my purposes admirably - on the rare occasions that it got out of the box. I am not interested in "bragging rights" although others may be.

As that 199 is within a maximum of 0.0002" of "flat" it should make (and in fact is) a very passable straight edge.

I do get a bit "disappointed" when people needlessly and sometimes irresponsibly toss "a tenth"around like a drunken sailor (I have considerable "hands on" experience there) or as if they were pulled out of thin air.

Posting pics/images here is achieved by use of an external server (I use PhotoBucket) and linking to the image file for that image on PB.

So far as I know, none of us have the right or access to attach pics.

There is a very good tutorial as a "sticky" at the top of the first page.

J Tiers
02-02-2010, 09:14 AM
Well, we finally got an actual number, 0.0002 +/- 0.0001". So, interpolating from my previous analysis, we are talking about 40% of the stated sensitivity as the maximum error from this curve that is scraped into the level.

OK, I can see how some may want better than that. It is still quite small, but not the 20% in my example calculation.



I think it could be more.....They don't give any limits for teh ends of the concavity, so we may assume 2.5" of flat.... and max concavity at the 0.0003 per the here-stated limits (where did they come from?).

if so, and you have to use the level over 1/2 its length only because the flat machine surface is only 7.5" and has a vertical surface behind it. The level will touch the flat at 2.5" in, and the center of the concavity will be on the edge of the surface.

That is then 0.0003" in 7.5-2.5", which is 12 *(0.0003* (7.5-5))= 0.00072 per foot or 1.4x the unit of sensitivity per foot, i.e. an error of 1.4 marked divisions.


That isn't too bad, but it would be good if Starrett actually gave limits, and suggested accuracies for real-world conditions which may not be ideal. 1/2 the length is about the maximum in that direction, and with a 15" level would almost HAVE to be considered. I would not expect any accuracy going with less than 1/2 the length........ even centered on pins or parallels. 15" is pretty darn long, though, most folks would be at half that or so leveling a lathe.


of course, that is for using the level as an absolute measuring tool. If you use it as a COMPARISON tool, the errors drop out, and you can probably get to 0.0001 per foot resolution, despite their 'fixing" a problem that you may not have.

And, since if you use the level, or any other 'error per unit length" device, at a fraction of its length, you can expect only a proportional accuracy, it might not be totally surprising. But if you assume that the level is several times better than the divisions, then at 1/2 the length you should still be within the stated limits. But you may not be.

They DO suggest in the catalog that the level should not be longer than the work, but interestingly, they only give that further down in the #98 section. I have no idea what they give in the instructions (such as they may be).

IIRC there was a very clear explanation of these things given by a different manufacturer, possibly Lufkin, who made a number of level types in the same or better accuracy. I have the catalogs somewhere, I might look it up.

chrisfournier
02-02-2010, 10:32 AM
I am one of the people who was foolish enough to buy a used pristine 199 for $79.00 instead of a new one for $700.00. Do I feel stupid now...

IIRC the top portion of the 199 is a seperate component from the base casting and I believe that it is made of plastic. It is also screwed to the base casting further isolating it from the base as far as imparting any dimensional input or stresses.

Would it not be prudent to handle the level by this portion when using it on your machinery. You don't need fingernails like Diana Ross to grip it in this fashion.

I realise that I don't have the Starrett instruction booklet, at a cost savings of $630.00 but is this technique not reasonable.

Oh yeah, I downloaded the instructions for free.

Honestly the indignation about this level's "potential tenths" on a home machinist site where much of the equipment is offshore import grade installed on whatever floor surface is available, meaning no engineered monolithic slabs is a bit much.

michaelantipode
02-02-2010, 08:24 PM
When I started the first discussion about these 199's it was because a friend and I had noticed a concavity in our levels. It was somewhat relevant to me because I have been truing up a lathe. The carriage is not 15 inches wide and so the level didn't sit on its ends; I was using the concave area. Having noticed that the base was concave, I wondered if it was "out" and if that mean other things might be suspect with it. Whether or not I try to get my home shop as good as I can is not the issue. WTF, I have been in people's houses that I wouldn't keep a roach in, but that doesn't mean I have to live like that.
Anyway, I sent another email to Paul at Starrett. I don't know what the US Fed Specs are on levels so don't know how to make sense of that. I also don't know how to support the whole of the level if there is a .0002 concavity, and if the surface it is sitting on is "better" than that. Here is the reply:


Michael,

We do not list all specs on our tools in the catalog. There is a US
federal spec for concavity of all levels that we must meet. You have
the hand heat backwards. The heat of your hands will cause the metal to
expand (in all directions) thus causing the base to expand downward and
cause it to rock. All levels need to be fully supported on the
workpiece to read accurately. This includes down to a carpenters level.
It would be helpful to note the concavity but I will say I have never
noticed a mention of this from any level manufacturer and every level
from every manufacturer has to have a concave base.

The top plate will help on the thermal expansion but many, if not most,
will pick the level up with their hands also on the cast body.

Best regards,

Paul Nault
Technical Support Department
Starrett
978-249-3551 x178
pnault@starrett.com

-----Original Message-----
From: garwood@netspace.net.au [mailto:garwood@netspace.net.au]
Sent: Tuesday, February 02, 2010 7:14 AM
To: Paul Nault
Subject: Re: FW: Submission from E-Mail Response Form

Paul,
Thank you for the reply and information.
I have some questions about this information:
I don't understand one is meant to hold the level and thereby heat it.
If one
holds it in the middle at the top (as do I) I would think that the
center would
heat, expand, and thus accentuate the concavity.
I have looked through the info on the Starrett web site regarding the
199 and
cannot find any mention of the .0002" concavity. I know of several
people who
have scraped their levels flat. Shouldn't Starrett make mention of this
and
explain it a bit better? There is mention of the nonconductive top plate
and of
the special alloy used to obtain freedom from thermal effects, but I see
no
mention of the concavity. This concavity makes a difference if one is
not using
the whole length of the level and that is why I am curious.
Thank you for any further information about this issue.
Regards,
Michael


> Michael,
>
>
>
> Yes, these are made intentionally with a .0002" concavity in the base.
> All machinist levels are made with this. Other levels have more than
> the .0002". The reason is if there were none, when a level is picked
> up, the heat of your hand can cause the metal to expand (usually in
the
> center) and cause the level to rock.
>
>
>
> Hope this settles this for you.
>
> Best regards,
>
> Paul Nault
> Technical Support Department
> Starrett
> 978-249-3551 x178
> pnault@starrett.com
>

oldtiffie
02-02-2010, 08:55 PM
Thanks Michael.

You do go to a lot of trouble to be sure you are well informed and to pass that on to the rest of us.

Well, well.

So the furthe it/this goes the more Starret is vindicated. It seems that it is as white and as innocent as the driven snow.

It seems that some here have "lost their innocense (and their virginity and credibility)".

I am very pleased about that.

So there is a US Standard on that. Fancy that. Fancy some of them here who chase and quote/cite such things both not knowing and not making it their business to find out and to be "informed".

The more so as it is "Starrett" and "good old American iron" (and "know-how").

So, accepting Starrett's advice at face value - ie correct - then any 199 - or "other" level that has been scraped either flat to to a lesser "hump" than the standard defines is "wrong".

It gets better as all who read your post are "informed" and if they sell their Starrett 199 etc. without advising the buyer that are in fact guilty of all that Starrett has been wrongly accused of previously here.

It doesn't get better than this.

Based on the "errors" - even in a "Graded and Certified" surface plate - it might be better if you were to precision-grind (to a flatness and parallel tolerance limit - not exceeding 0.0001/0.00005"- a range of parallel strips of varying lengths that you could span symmetrically under and by your Starrett 199 level.

Your 199 vial ("bubble") will need to be checked periodically and certainly before, during and after any job that required the level of accuracy that you seem to infer that you need (it for).

Greg Q
02-03-2010, 02:09 AM
Not so fast sport.

Here is a copy of the single page of the US Federal Spec A-A-50685:

1.
iWWIZ.
COMMERCIAL ITEM DESCRIPTION LEVEL, PRECISION MASTER, SINGLE FACE
The General Services Administration has authorized the use of this commercial item description as a replacement for Type IX, class 1, single face, precision master level in superseded specification GGG-L-211C.
1.1 This commercial item description covers a single face, precision master level used to calibrate other levels.
2. REQUIREMENTS.
>
2.1 ~ . . The item shall be made of steel or close grain cast iron of a design to preclude warping or deflection due to temperature change. It shall also consist of a single faced stock, one adjustable level vial, one f~ed cross test vial, a vial holder, a cover and a case. The vial shall have a cover or top plate of heavy nonferrous metal or plastic to shield vials from heat of operator’s hand. It shall be 15 inches long and have vial accuracy of 0.0005 inches per foot of single faced stock. Concavity and convexity of single faced stock shall not exceed 0.0002 inches per foot. The vial shall be a glass tube with a vial sensitivity of 10 seconds, maximum. The vial shall have no less than three auxiliary lines, permanently marked, spaced l/10 inch apart on both sides of both bubble defining lines. The angular value of these lines in either degrees or inches per foot shall be permanently marked on the level.
2.2 The level, including all parts and accessories, shall be free from burrs, rust, external sharp or rough edges, imperfections, and any other defects which may impair serviceability, durability, appearance, or operational performance.
3. QUALITY ASSURANCE.
3.1 ~ The contractor shall certify that the product offered meets the salient characteristics of this description and that the product conforms to the producer’s own drawings, specifications, standards
and quality assurance practices, and is the same product offered for sale in the commercial marketplace. The government reserves the right to require proof of such conformance prior to first delivery and thereatl.er as maybe otherwise provided for under the provisions of the contract.
4. PREPARATION FOR DELIVERY. 4.1 l?reservat ion. Packaeinz. Packinz and Markin~& Preservation, packaging, packing, labeling,
and marking shall be as specified in the contract or order.
MILITARY INTEREST:
Custodians: Air Force -99
AMSC N/A DISTRIBUTION STATEMENT
Preparing activity Air Force -84 Project No: 5220-F217
A: Approved
for public
release; distribution is unlimited.
A-A-50685 &JULYd989
IFSC 52201


Emphasis is mine. Note.

US Fed Specs are not covered by copyright and are free to reproduce.

Forrest Addy
02-03-2010, 02:58 AM
OK. Cool. If Starrett makes their levels to a standard then they should cite the standard. I could find no standard cited for the Starrett #199 master precision level in my box of Starrett catalogs nor on their website. In the same catalogs their surface plates and other precision tools are documented to the nostrils with relevent standards. Starrett sells precision tools to professional users so citing relevent standards is plain good industrial ethics as well as good marketing.

However the spec Greg cited states a bilateral tolerance for precision levels whereas the Starrett rep says they hold a unilateral toleranceso the level sole is concave. If the 11 levels I scraped and calbrated (and I think checked and adjusted a dozen others over the years) most all were concave more than 0.0002" in a foot exceeding Greg's Federal Spec A-A-50685 cited tolerance.

That's the core of my complaint with Starrett. Tiff, you can chortle over who overplayed their hand but I think anyone using a Starrett #199 level to the accuracy I've been called on to attain would have good reason to complain about more than two tenths per ft deliberate concavity.

For the record, I know excactly how much a tenth of a thousandth is. I've chased them on and off for nearly my whole career. Sometimes they don't mean much; just data. Sometimes an accumulation of tenths leads you to confusion, frustration, or disaster if you fail to account for every one.

And by the way, I was just out in my shop running one tests. I stoned off my planer table and placed one of my #199 levels on 1-2-3 blocks set a foot apart on the 3" end. Then I set a LDVT in the middle of the sole. The LDVT was connected to a gage amp to register vertical movement for some heat tests. In every case placing my hand on top of the level insulator and lapping my fingers around to contact the cast iron warmed the upper part of the frame. In every case the level sole became MORE concave as a result of thermal expansion of the upper limb. Handling a level from the top insulator is the way I was taught to handle a preision level: always by the inulator. I was also taught to never under any circumstances cradle the sole.

If one were in the habit of cradling the sole, the level frame would indeed go convex cancelling Starrett's scraped in bias as the sole expanded in advance of the back. How long do you have to hold a Starrett #199 level by the sole for it expand enough to straighten out? About 40 seconds. Don't believe me? Try your level on your surface plate. Push an end sideways. If it is in thermal equilibrium and the sole is a bit concave, the level should pivot from the opposite end. Next step: hold the level in the palm of your hand by the center of the sole for one minute. Place the level on the surface plate and push one end sideways. Pivots in the center now doesn't it? The level went convex like the runner of a rocking chair.

The shape the level's sole assumes is dependent on how it's handled, a point Mr Nault of Starrett faled to address these alternatives in handling technques. If one were involved in exploiting the level's maximum accuracy, he would likely take care to keep the level (and the whole of the item beng leveled) in thermal equilibrium. For that reason alone I scrape a precision level flat as I can get it.

Point is when things get fussy, small influences have huge consequences. That's why I hammer these hair-splitting topics. There are no mysteries in the machinist's trade. Only people who fail to pay attention to details.

chrisfournier, that black plastic thingy may serve the functon of a bezel covering the vial but it's an insulator the way I use it. I can bare hand the black plastic all day without transmitting heat to the level frame. It's even better if you glue 3/16" thick leather on the level's cheeks between the insulator and sole. Congrats on your #79 level, by the way. Good is good. Good and cheap is even better.

Is this horse dead yet? I feel like I've been beating it long enough.

luthor
02-03-2010, 05:04 AM
Is this horse dead yet? I feel like I've been beating it long enough.

It did take a while but I think you put it out of it's misery now.

Greg Q
02-03-2010, 05:59 AM
Wait, we were just getting wound up;) Or at least someone was.

We still haven't covered the DIN 877 spec yet, or the fact that the US spec supersedes an older one that I couldn't find. Or that the GAO accepted a commercial item description as the specification. Talk about log rolling: Here we have a specification for a tool that is itself taken from the manufacturer's description. (Read the first paragraph). Circular logic is at work*

I understand now why all the US built master precision levels are 15" long-the spec requires it. The Europeans, not being so constrained, offer a variety of lengths (and hence deeply relieved soles-you match the tool length to the work without having to use parallels. Sells more tools that way, too:cool:

Finally, what of the grooved base levels? Please don't anyone claim that the involute or vee grooves are similarly concave.

*see:circular logic

luthor
02-03-2010, 06:15 AM
Has anyone got a copy of the applicable British Standard?

oldtiffie
02-03-2010, 06:24 AM
Forest.

First of all, if you or one of the wounded complainants had emailed Starrett at all, preferably earlier, then you and the other wounded complainants would have been better informed - earlier.

Greg - from down here in OZ didn't seem to have much trouble winkling out the relevant spec - and posting it here. Maybe it just gravitated to the bottom of the trash bin - just where you might expect me in OZ to be.

Here is an excerpt from the spec - the emphasis is mine:


2.1 ~ . . The item shall be made of steel or close grain cast iron of a design to preclude warping or deflection due to temperature change. It shall also consist of a single faced stock, one adjustable level vial, one f~ed cross test vial, a vial holder, a cover and a case. The vial shall have a cover or top plate of heavy nonferrous metal or plastic to shield vials from heat of operator’s hand. It shall be 15 inches long and have vial accuracy of 0.0005 inches per foot of single faced stock. Concavity and convexity of single faced stock shall not exceed 0.0002 inches per foot. The vial shall be a glass tube with a vial sensitivity of 10 seconds, maximum. The vial shall have no less than three auxiliary lines, permanently marked, spaced l/10 inch apart on both sides of both bubble defining lines. The angular value of these lines in either degrees or inches per foot shall be permanently marked on the level.

The bubble is to be protected from the heat of the hand.

As it should be - as any heat source - hands very much included - will seriously "pull" the bubble in the vial.

Anyone who has ever used an engineer's level or a dumpy level or a theodolite will be very well aware of that phenomenon - and its effects and potential consequences - and guards against it. It is also why my tri-pods were wood and not aluminium. All of my tapes and staffs were calibrated at 20*C - and if required I had to adjust for heat effects. I had to keep my theodolite tribrach, transit and alidade levels correct to at least 20 and sometimes 10 seconds. So I am well aware of the effects of heat an temperature. That includes using a level in the shop if the degree of accuracy required justified it and localised lighting or sun-light may have been a problem.


Concavity and convexity of single faced stock shall not exceed 0.0002 inches per foot.

I note that the level face may be anywhere between a concavity or a convexity of 0.0002" which places the limits 0.0004" apart.

Starrett is quite correct in opting for the concavity as the level will not "rock" as it would or may if the sole-plate was convex.

So Starrett have exceeded the requirements of the spec. and have applied a lot of good sense and best shop practice.

I think that the OP, Greg and Starrett are deserving of a bit more thanks and appreciation of and for their efforts for all of us as those qualities are very conspicuous by their absence or scarcity.

They all certainly do have my appreciation and gratitude for a good job well done.

If I had a 199 level, I'd very pleased to get that information - period.

J Tiers
02-03-2010, 09:02 AM
hey there, oh tifflepiffle.......

There is NO REQUIREMENT for anything in the US to be made to ANY "Federal spec" without a law or regulation to require it.

In fact, the laws and regulations are NOT "Federal specs", they are laws, like pollution limits, etc.

So self-righteously pointing to the "Federal spec" and saying , in effect, "if you fools would have just read it, you'd know..... but you were too damn dumb to do that", is silly and wrong-headed.

Starrett simply kept back the information, and that is the SOLE ISSUE HERE.

As I have written several time so far....... it would have been SO SIMPLE for Starrett to have given complete information, referenced the outside spec they use, etc.....

Starrett CHOSE not to do that, resting apparently on their smug assumption that they are the best, are unchallenged, and that everyone knows and accepts that.

I strongly suspect that these assumptions are NOT true, and that Starrett is extremely vulnerable. This sort of thing does not help their reputation, in my estimation. Not among "those who know".....

I would prefer that they stayed in business, but they probably will not, not unless that smugness is wiped off their face.

ulav8r
02-03-2010, 03:47 PM
Forrest,

Could you do one more test? Set the level(with concave base, not one you have scraped flat) on a surface plate with each end supported on a riser of some kind and record the bubble location. Then move one riser to the center of the level and note the change in bubble location. Based on Starrets' admission that the levels are supplied with non flat bases, we know that there will be a difference. But do we know if it will repeat if rotated end for end if supported by one end and somewhere other than the center or other end? IOW, is the concave the same from each end? Is it a constant radius or, more likely, some random nonflat surface?

Another test would to be check the level of a precision flat that was 2/3 or so of the length of the level, alternating ends of the level. I believe that you wold probably find a difference. I do not believe that Starret would bother to scrape the recess to a constant radius.

In case no one realizes it, I side with J Tiers and Forrest on this. The reason I ask Forrest if he could do this instead of doint it myself is that I do not have a precision level of any kind, or the experience and skill to do it myself.

Forrest Addy
02-03-2010, 05:04 PM
Ulav8SR

I tried what you suggested some time ago wondering what effect the concave sole would have. In the worst case the error was only about 1/2 a graduation. No biggie you might think unless you needed to work to the limits of the level's senitivity.

oldtiffie
02-03-2010, 05:23 PM
hey there, oh tifflepiffle.......

There is NO REQUIREMENT for anything in the US to be made to ANY "Federal spec" without a law or regulation to require it.

In fact, the laws and regulations are NOT "Federal specs", they are laws, like pollution limits, etc.

So self-righteously pointing to the "Federal spec" and saying , in effect, "if you fools would have just read it, you'd know..... but you were too damn dumb to do that", is silly and wrong-headed.

Starrett simply kept back the information, and that is the SOLE ISSUE HERE.

As I have written several time so far....... it would have been SO SIMPLE for Starrett to have given complete information, referenced the outside spec they use, etc.....

Starrett CHOSE not to do that, resting apparently on their smug assumption that they are the best, are unchallenged, and that everyone knows and accepts that.

I strongly suspect that these assumptions are NOT true, and that Starrett is extremely vulnerable. This sort of thing does not help their reputation, in my estimation. Not among "those who know".....

I would prefer that they stayed in business, but they probably will not, not unless that smugness is wiped off their face.

Thanks JT.

I can and do appreciate your position and response and I really do admire you for it.

My point here is that all that was needed - certainly now and possibly earlier - was to simply ask Starrett if there is a code, Law,specification etc. that they used as a basis or bench-mark for manufacturing the 199 level.

The OP not only did it but had no problem asking Starret.

Greg had no real problem that I am aware of in finding the spec just by knowing there is one and not even knowing its name or title.

So that wasn't too difficult either.

I certainly would expect people to at least consider those options and sources or advice and reference material/codes etc.

I don't think that I at least should expect Starrett to "spoon-feed" me.

Those who purport to or who seem to give the impression that they are privy to (custodian of?) all the "knowledge" should perhaps be a little more cautious and better informed.

Enunciating stuff like Colonel Blimp, Moses or the Pope without adequate support is both "not the way to go" and just asking for trouble - and in this case, perhaps, "getting it".
http://en.wikipedia.org/wiki/Colonel_Blimp

The "because I/he/they/my mentor and "Fred down the road" said/say so" as reference is hardly a good basis to support a strong assertion.

If I had been proved or seemed to be wrong in this matter and if I had a Starrett 199 level, I would be both very pleased and walking backward on my (now) previously-held beliefs at 100 mph (well 98 1/2 anyway). I'd also make sure that all and sundry were aware or my change of position, the reasons for it, and issue any necessary apologies.

I think the 199 is an excellent tool and I'd be glad to have one. If I really needed it and could not get a good one on the "market" I'd either buy a new 199 or one from Mitutoyo or similar. If I had to scrape it - I'd do it on my Grade AA surface plate which is not only accurate enough to find out if the 199 was "up to spec" but to scrape it to as well.

I would be surprised if those who are critical or doubtful of or concerned about the "0.0002" over 12" " could put their hands on their hearts and assure first themselves and then anyone else that any other straight edge or parallel strip is as accurate as that level.

My guess is that many will be very pleasantly surprised to find out just how good that 199 actually is.

If my assessment is correct that Starrett 199 is flatter than many of the surface plates are that may be used to check it are.

I really can't see me having real job-related need for a level better than that.

I would be more concerned about checking and if necessary re-setting the bubble/vial to the accuracy needed to get the optimum performance from the base of the level. That is a bit of an art in itself.

I must say that it seems odd and not a little ironic that it is I - the so-called "anti old 'meric'n iron" protagonist being in the position of actually defending an "American Icon" (two actually) - Starrett as a manufacturer and its 199 precision level when there are so many self-proclaimed "defenders of the "American way and icons" " arraigned against it.

Perhaps they are more concerned with defending their own "honour".

I can assure them that I will NEVER EVER remove, take or aspire to "ruin" THAT!!! So, to that extent, it is quite "safe" although it may have been lost by and not taken from them - long ago.

I have this vision of an outraged parent/guardian/"father" casting his "disgraced and ruined" daughter/child and her "bundle" out into the cold cold snow and not seeing that it is he that is in fact "carrying/stuck with the "baby" ".

Odd that.

Fasttrack
02-03-2010, 05:34 PM
Ulav8SR

I tried what you suggested some time ago wondering what effect the concave sole would have. In the worst case the error was only about 1/2 a graduation. No biggie you might think unless you needed to work to the limits of the level's senitivity.


Hey Forrest! Check your pm ... ;)
(pm unrelated to this thread)

Greg Q
02-03-2010, 07:11 PM
Tiffie, I share the belief that a tool should be openly documented to the extent that any assumptions about its performance may be faulty.

As Forrest pointed out, the rest of their "foundation" tools carry a statement of accuracy and often cite the pertinent reference. It can be surprising to discover that there are hidden gotchas in many tools.

Just for kicks I looked up Scherr-Tumico's blurb for their master precision level. It claims to be precision ground and hand lapped. I don't know enough about lapping-is it possible to lap a center section 0.0002 hollow?

Earlier I mentioned that the spec explains the 15" length. I have since discovered 12" and 18" models of some levels. But it is my understanding that a manufacturer's product must meet the published spec if it has any hopes of selling to branches of the US federal government. If that is wrong can somebody correct me please.

On measuring a distance shorter than 15": If a parallel is place in the center of the level and another at one end it follows that the sensitivity is doubled. (A 0.0002 hollow will be the same as a 0.0004 difference over the full length). Hence, the error caused by the hollow can be even more vexing than half a division.

Tool maintenance: I'd guess that almost everyone on this board maintains their own lawnmowers and most guys have had a crack at cars and chainsaws too. Sharpening screwdrivers and drills and plane blades etc is common. Why not tool maintenance in the precision end of the shop? Since you can make a basic scraping set-up* for less than $100 + granite surface it is possible to maintain all of the iron surfaces to whatever degree of refinement desired/required reasonably cheaply.

Even the cheap surface plates should have an area flat to 0.00005. Using thin spotting ink, followed by an alcohol haze spotting can yield some crazy flat surfaces. So flat they just about wring to the granite. (Of course this exercise has drawbacks in a world full of 10 micron dust, but you get the idea)

There may be another surprise on this topic if Mike chimes in;)

*diamond grinding wheel on ebay + cast iron lapping wheel & diamond paste + carbide tipped scraper shop made + spotting medium

oldtiffie
02-03-2010, 07:20 PM
A very interesting, informative, thought-provoking and focused post Greg.

I have no real issues with it at all.

I have an appointment in Frankston shortly and will respond in the detail that your post deserves/requires - later in the day.

michaelantipode
02-03-2010, 07:56 PM
I have sent another missive to Paul at Starrett. It and his reply are below.

Michael,

US govt. spec GGG-L-211D
I have an older copy in front of me (version 211c). The difference
between c & d were minor and not related to the master levels according
to our calibration lab. Table XXI column B lists the tolerance spec for
12, 15 & 18" master levels (Type IX) as .0002.

Best regards,

Paul Nault
Technical Support Department
Starrett
978-249-3551 x178
pnault@starrett.com

-----Original Message-----
From: garwood@netspace.net.au [mailto:garwood@netspace.net.au]
Sent: Wednesday, February 03, 2010 3:32 AM
To: Paul Nault
Subject: Starrett 199

Paul,
Thank you for your reply.
Still some questions about all of this. I looked at the Federal Specs.and
came up
with the info below.
I am sorry. I see no mandated .0002 concavity. In fact, what I see is a
+/-
.0002 tolerance limit.
Can you illuminate this a bit? Are these the correct specs.?
Thank you,
and Best Regards,
Michael

Here is a copy of the single page of the US Federal Spec A-A-50685:

1. COMMERCIAL ITEM DESCRIPTION LEVEL, PRECISION MASTER, SINGLE FACE
The General Services Administration has authorized the use of this
commercial
item description as a replacement for Type IX, class 1, single face,
precision
master level in superseded specification GGG-L-211C.
1.1 This commercial item description covers a single face, precision
master
level used to calibrate other levels.
2. REQUIREMENTS.
2.1 The item shall be made of steel or close grain cast iron of a
design to
preclude warping or deflection due to temperature change. It shall also
consist
of a single faced stock, one adjustable level vial, one f~ed cross test
vial, a
vial holder, a cover and a case. The vial shall have a cover or top
plate of
heavy nonferrous metal or plastic to shield vials from heat of
operator's hand.
It shall be 15 inches long and have vial accuracy of 0.0005 inches per
foot of
single faced stock. Concavity and convexity of single faced stock shall
not
exceed 0.0002 inches per foot. The vial shall be a glass tube with a
vial
sensitivity of 10 seconds, maximum. The vial shall have no less than
three
auxiliary lines, permanently marked, spaced l/10 inch apart on both
sides of
both bubble defining lines. The angular value of these lines in either
degrees
or inches per foot shall be permanently marked on the level.
2.2 The level, including all parts and accessories, shall be free from
burrs,
rust, external sharp or rough edges, imperfections, and any other
defects which
may impair serviceability, durability, appearance, or operational
performance.
3. QUALITY ASSURANCE.
3.1 ~ The contractor shall certify that the product offered meets the
salient
characteristics of this description and that the product conforms to the
producer's own drawings, specifications, standards
and quality assurance practices, and is the same product offered for
sale in the
commercial marketplace. The government reserves the right to require
proof of
such conformance prior to first delivery and thereatl.er as maybe
otherwise
provided for under the provisions of the contract.

Greg Q
02-03-2010, 08:28 PM
Thanks Michael, well timed:D

OK folks, here's the other thing: Starrett cites Fed Spec GGG-L-211D.

Follow the link below. It is a list of cancelled specifications. The one Starrett mentions was cancelled in 1998. The GAO now uses A-A-50685 which has been posted here. I'm sure my ebay 199 was made before 1998. I see that the old spec also mentions 0.0002 according to Paul of Starrett's reply, but he doesn't say if that's a unilateral tolerance.

http://aero-defense.ihs.com/document/abstract/UDOIIAAAAAAAAAAA

also: http://apps.fas.gsa.gov/pub/fedspecs/sort5g.cfm

Along the way I discovered this cool site for specs:

http://www.everyspec.com/FED+SPECS/

Greg

J Tiers
02-03-2010, 09:37 PM
On measuring a distance shorter than 15": If a parallel is place in the center of the level and another at one end it follows that the sensitivity is doubled. (A 0.0002 hollow will be the same as a 0.0004 difference over the full length). Hence, the error caused by the hollow can be even more vexing than half a division.

I pointed this out, but Tiffie seems to have ignored that, And the fact that given the priorly stated tolerance of 0.0002 hollow +- 0.0001, the level could be almost 1.5 divisions "out" if used over half length, starting from one end (presumably because of the way the surface is bounded).




I certainly would expect people to at least consider those options and sources or advice and reference material/codes etc.

I don't think that I at least should expect Starrett to "spoon-feed" me.



How unutterably odd...... Spoonfeeding now? Advertising literature and product specs are "spoonfeeding"...? A peculiar looking-glass world THAT came out of, I'd say.

Bah..........

I mentioned that the key to this is not the EXISTENCE of a spec, but the fact that Sterrett (in this case) actually USES the spec.......

I have lots of federal specs on ref here, and IEEE, ASME, even UL and EU standards. Standards are one thing, following them, or needing to, is quite another. But it is a bit important that the vendor should have PAID ATTENTION TO THEM, eh? Otherwise reading them up is not nearly as useful.

There isn't anything compelling them to follow the spec. They easily might not. There are manufacturers who don't follow UL standards also...... and those are a tad more important, being safety standards enforced by the federal authorities in the form of OSHA for workplaces.

So before you write more voluminous tomes of smug superiority, you might perhaps consider that........ We might read the spec and say..... " I wonder if Starrett follows THIS? They certainly must not, because if they DID they surely would reference it in their literature." Quite a reasonable conclusion, and nothing to prove it one way or t'other unless you are sufficiently interested to ask, which I am not.

I was more interested in the fact that the specs Starrett gave were quite sketchy, the only one of any substance was the sensitivity.

And we'd already seen the Starrett answer anyway.

YOU think (for some strange reason) that it is "spoonfeeding" if Starrett were to actually say that they do in fact follow a particular Federal spec. Now, most normal people would suppose that it was marketing, adding credibility and weight to the claims.

Eh, I suppose the actual spec of 0.0005 per inch is spoonfed as well, since we could determine it ourselves with some shims and bar stock....... Why even claim the thing is a level? Surely you can see that by yourself, now.

I must admire your rugged individuality, Tiffie, your healthy self-reliant attitude of "let the customer figure it all out, don't tell the bastiches a thing... they need it they have to buy it, we don't need to tell them anything". Just the program to keep the customer base energized and fiercely loyal to the brand..........

There goes Tiffie.........Proudly marching with Starrett................................ and Toyota........................

Ta...............I've got some box tools to do up.

oldtiffie
02-04-2010, 12:24 AM
I have sent another missive to Paul at Starrett. It and his reply are below.

Michael,

US govt. spec GGG-L-211D
I have an older copy in front of me (version 211c). The difference
between c & d were minor and not related to the master levels according
to our calibration lab. Table XXI column B lists the tolerance spec for
12, 15 & 18" master levels (Type IX) as .0002.

Best regards,

Paul Nault
Technical Support Department
Starrett
978-249-3551 x178
pnault@starrett.com

-----Original Message-----
From: garwood@netspace.net.au [mailto:garwood@netspace.net.au]
Sent: Wednesday, February 03, 2010 3:32 AM
To: Paul Nault
Subject: Starrett 199

Paul,
Thank you for your reply.
Still some questions about all of this. I looked at the Federal Specs.and
came up
with the info below.
I am sorry. I see no mandated .0002 concavity. In fact, what I see is a
+/-
.0002 tolerance limit.
Can you illuminate this a bit? Are these the correct specs.?
Thank you,
and Best Regards,
Michael

Here is a copy of the single page of the US Federal Spec A-A-50685:

............................

..............................




Nice going Michael.

Starrett have all but said that it has all necessary accreditations and the opinion of its certified Calibration Laboratory is that all necessary requirements have been met. I would doubt that they would over-turn or retract that opinion and I further doubt that anyone with sufficient "clout" to make them do it would even bother to try.

So the essence of it is that: "that's your lot".

oldtiffie
02-04-2010, 04:25 AM
Thanks Greg.

I will respond on a para-graph/item by para-graph/item basis. So please excuse me for seemingly dissecting your post.


Tiffie, I share the belief that a tool should be openly documented to the extent that any assumptions about its performance may be faulty.

Having a belief and having it realised are not the same.

Starrett may or may not do it else-where and while perhaps a good idea and applied for that tool it may not be sufficient precedent to oblige them to do so elsewhere.

Its a Starrett "call".

If I really "needing" as apposed to just "wanting" that information, it I'd have emailed or phoned Starrett OZ. I have no reason to believe that I would not get the same service and courtesy as was obtained in this thread - and just as easily.

I amazed that so many here who wanted a "good" level did not make those inquiries before they bought those levels. My guess is that most were bought second-hand/used on an "as is" basis (no warranty) from other than Starrett. On that basis alone I don't see that Starrett has any obligation to owners other than the original purchaser at all. That they do it is a very big "plus" for Starrett.

I don't see that Starrett needs or is obliged to rescue people from there own actions and short-comings.

Its all very well for some to sit on there ar$es and decide what Starrett must and should and will do for them - and expect it "as of right".

It will do some here no harm to get them off their ar$es and the Starrett "tit" and take a bit of positive action and initiative for themselves.

As Forrest pointed out, the rest of their "foundation" tools carry a statement of accuracy and often cite the pertinent reference. It can be surprising to discover that there are hidden gotchas in many tools.

Its just as applicable to Forrest as it is to anyone else - perhaps the more so given his extensive use and history with Starrett tools in general and this one (199) in particular. Perhaps any who were looking to Forrest for advice and example in this instance didn't get it. That's a matter for Forrest to address in his own way in his own time.

Refer the matter to Forrest.
Just for kicks I looked up Scherr-Tumico's blurb for their master precision level. It claims to be precision ground and hand lapped. I don't know enough about lapping-is it possible to lap a center section 0.0002 hollow?

It certainly is. It is just a variation of the "hand" technique used on flat and spherical laps - in this case with a cylinder with a radius and chord to suit the base of the level.

Earlier I mentioned that the spec explains the 15" length. I have since discovered 12" and 18" models of some levels. But it is my understanding that a manufacturer's product must meet the published spec if it has any hopes of selling to branches of the US federal government. If that is wrong can somebody correct me please.

So far as I can see in the specific instance/case of the 199 level which is the subject of this thread, the military/government specification was the requirement to be met for US government use - and so far as I can see, Starrett met those requirements.

I would not try to relate the 199 requirement to any other document or requirement.

On measuring a distance shorter than 15": If a parallel is place in the center of the level and another at one end it follows that the sensitivity is doubled. (A 0.0002 hollow will be the same as a 0.0004 difference over the full length). Hence, the error caused by the hollow can be even more vexing than half a division.

Don't agree as a curve is not necessarily a regular or geometric curve as it can vary within its tolerance limits envelope.

If I were that concerned I'd probably secure one end and move a spacer along the curve at pre-determined interval between the ends and would record and/or "map" the actual rise and fall of the non-secured (opposite) end - and repeat for and from the previously "free" end toward the previously "fixed" end. I would use/consult the "map" when necessary when using the level under the same or similar conditions.

Tool maintenance: I'd guess that almost everyone on this board maintains their own lawnmowers and most guys have had a crack at cars and chainsaws too. Sharpening screwdrivers and drills and plane blades etc is common. Why not tool maintenance in the precision end of the shop? Since you can make a basic scraping set-up* for less than $100 + granite surface it is possible to maintain all of the iron surfaces to whatever degree of refinement desired/required reasonably cheaply.

Agreed but that is in Forrest's domain/bailiwick - and I will leave that to Forrest.

Even the cheap surface plates should have an area flat to 0.00005.

Agreed. But it depends entirely on the quality (Grade) and the size of the surface plate.

0.00005" is "50m millionths" and unless it is "mapped" and not worn it will be very difficult to say or find where that elusive 50 micro-inch "flat" is.

My mapped Grade AA plate has a requirement of accuracy of 75 micro-inches but the map points out an area where 50 or better is available - hence my buying a Grade AA mapped plate.

Unless I really need that plate it stays in its box. I find that my float glass plate (~ 0.0001"/") together with my mill table and the "ground-in" tops of my surface grinder magnetic chucks and the top face of the grinder slide is more than adequate for 99.9% of my work. That's what the grinders are for.

Using thin spotting ink, followed by an alcohol haze spotting can yield some crazy flat surfaces. So flat they just about wring to the granite. (Of course this exercise has drawbacks in a world full of 10 micron dust, but you get the idea).

Again - Forrest's domain.

There may be another surprise on this topic if Mike chimes in;)

*diamond grinding wheel on ebay + cast iron lapping wheel & diamond paste + carbide tipped scraper shop made + spotting medium

No comment.

.RC.
06-11-2010, 10:51 PM
I have discovered why Starrett 199 hand scraped levels are bowed in the middle and I have been able to reproduce this bow..

I got a Starrett 199 yesterday and while it is in good order the base has scratches and one small minor ding..

I thought I would scrape it flat... So I did...Now when I put it on a different area on my chinese granite 18X24 surface plate it comes up as being high in the middle, and yet another place shows it as being high on the ends...

So there you go, the way they have done it is by buying chinese surface plated and using them as scraping masters..

Only thing is I do not know if mine is concave or convex but it sure as hell is not flat...

http://img.photobucket.com/albums/v606/OzRinger/starrett199007Custom.jpg

http://img.photobucket.com/albums/v606/OzRinger/starrett199002Custom.jpg

Fasttrack
06-11-2010, 11:26 PM
So there you go, the way they have done it is by buying chinese surface plated and using them as scraping masters..



:D IIRC, Mcgyver ran into a similar situation. Makes me scared to do any scraping with a Chineese flat. I've been half-heartedly looking around for a used, but recently calibrated, surface plate for awhile now.

.RC.
06-11-2010, 11:39 PM
:D IIRC, Mcgyver ran into a similar situation. Makes me scared to do any scraping with a Chineese flat. I've been half-heartedly looking around for a used, but recently calibrated, surface plate for awhile now.

There is no such thing as a second hand surface plate in this country (well they are very rare)

A new chinese 24X36 00 grade is $800+ here...

oldtiffie
06-12-2010, 12:21 AM
Ringer.

I don't know if your Chinese surface plate is any good or not, or what condition it is in or what grade it purports to be - so the fault could be in either or both the surface plate or the level. I have no idea of the extent of any error or either.

I note that both are "pre-loved".

It is quite possible that the level has been "re-scraped" by "others" too.

I'd suggest that a good "known" surface plate of the right accuracy would a good if not the only realistic place to start.

A bit of necessary math revision as regards inches and metres/mm etc. will be a good start point for reasons that I hope will be evident further on.

1 millionth of an inch = 1 micro-inch - 1ui

1 millionth of a meter = 1 thousandth of a mm = 1um

1 um ~ 40ui

10um ~ 400ui ~ 0.00004"

So, lets have a look at a typical Grade B (or 2) "Shop" grade surface plate:
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplatetablecalcs1.jpg

That plate has an acceptable error of 0.000398" ~ 0.0004" ("4 tenths") which is in excess of that allowed for the Starrett level - and further as we don't have a "plot" for the "B" plate, we don't know where the errors are either.

This is the "plot" for my "Chinese" Grade "0" plate which is two grades better than a "B".
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplategradediag1.jpg

The maximum errors from flat are 1.2um x 40 ~ 48ui ~ 0.000048ui ~ 0.00005" (~ "half a tenth") and that is along the top/high edge.

The best flat spot for checking toy level is along the diagonal from top left to bottom right as there is no difference between the ends of the diagonal and only a 0.12um (0.12um x 40 ~ 4.8ui 0.0000048" ~ 4.8 say 5 millionths of an inch).

That's about as good as it gets.

I know, I know, I know - "but its Chinese so its $hit" etc. but I have no good reason to doubt it as I've given it what I consider to be a good "run through" and it came up "tops".

So, it really does need better than a "Shop"/"B"/"2" surface plate - with "plot" - to check and scrape that level to if you are to be confident and assured of meeting the original "Starrett" quality and limits.

I suggest that the situation your are now in asks more questions than it answers.

Forrest Addy
06-12-2010, 12:46 AM
I think this discussion has progressed to splitting hairs that have been previouly split.

Greg Q
06-12-2010, 01:01 AM
Ringer, if you want to be absolutely sure, send it to me and I'll scrape it in using my own and Mike's levels in a three-way.

Greg

oldtiffie
06-12-2010, 01:08 AM
I think you are right Forrest, but it should make the point that a "Starrett" (or any other) "Master" level needs a surface plate at least one grade and preferably two grades better than the level is or is to be made to be.

Calibration is crucial to get the best out of any plate.

Even in an ideal world, a perfect plate will only indicate where it touches. The amount and degree of "misses" is not known - or given - and I can see no normal way of getting it either - on a surface plate by "blue-ing".

If I recall correctly - without checking further back - the Starrett spec had a "rise" in the form of a notional arc of the order of 0.0004" or so over its length - but good conditions and great care were required.

Frankly, if it were me, I'd scrape it flat and be done with it.

But it is a casting after all that possibly has or has had residual stresses in it that have been or are being "relieved" even yet - hence the QA/QC case for having such "master" tools checked, made good if necessary, and re-calibrated and re-certified.

We are not to know that Ringer's plate and level have both been rejected and disposed of after they failed any required QA/QC check - or not.

Short of recalibration, I can't see how we would ever know either.

.RC.
06-12-2010, 01:57 AM
Ringer, if you want to be absolutely sure, send it to me and I'll scrape it in using my own and Mike's levels in a three-way.

Greg

Thanks for the offer Greg, but at this stage I will use it as is... I believe the concavity or convexity to be minimal, and I only noticed it by using the pivot point test as shown in Rich King video.....






I note that both are "pre-loved".



No the surface plate I bought new...However, it came with no certificates, paperwork, it has no identification on it nor does it even have marks on it where the three feet are supposed to go on it...

In any case when it comes to anything like this, unless you you have a certified traceable certificate whatever comes with your plate is meaningless... Because you have no way of testing if your plate is flat you have to trust the manufacturer... The question is... Do you trust most manufacturers???

Would I trust a certificate out of Brown and Sharpe? Yes I would as we know the companies history.. Does that mean they always make 100% accurate products....No...If the QA officer has a bad day then mistakes can be made..

Would I always trust a certificate out of a company in China? No, I would not as I have no idea of what the company is like... That is not to say the chinese company is making a bad product, just that I do not know their history..

You mentioned you got your surface plate from a chinese couple, the husband being a metrologist... I would say that would be a very good indication that what you have got on your certificate would be more likely to be accurate then inaccurate.. Do they have a website or contact number as I was considering a larger 24 X 36 plate anyway...

.RC.
06-12-2010, 03:31 AM
I have solved the mystery.... We can finally put the myth to rest of the concave base.. And every one here who owns a 199 can test my theory to either prove or fail it...

I was tearing my hair out trying to find out where the problem layed... I tested different parts of the plate with the level and was getting similar results all over the plate...

My next point of call was to blue the level and scrape it....When I blued it up I got points all over the base.... I scraped it with a 40mm radius blade to increase the ppi count (now over 30) :)

Back to the surface plate and it looked good, blue spots all over...

I then had a half hour break...

Back to the surface plate for a final blueing.... WTF!!!!!!!!! Yes it was convex again and high in the middle.... CLUNK went the brain.... Even though I was super careful with my handling of the level, just a minor amount of heat from your hands into the level will make it go concave..

It is a real bastard of a cold day here today as it is 16 degrees C and the heat differential of my body is enough to throw it out of wack with minimal handling...

So while everyones Starrett may be concave I bet it has something to do with the way we handle it, as even small amounts of body heat will distort it... I cannot think of any other reason for the difference...

Forrest Addy
06-12-2010, 04:37 AM
Ringer, Heh, heh, heh. Welcome to the "Snake Bit" club. Been there, been bit by the same snake - in spades. I could tell you stories. In my hot little hands slightly needy straight edges turned into rocking chair runners and surface plates into salad bowls. The world is a cruel place for the feckless.

oldtiffie
06-12-2010, 06:58 AM
Good experience well handled and well reported Ringer.

On that basis I'd be prepared to say that at 21*C your plate and the base of your level will be pretty good and could be classed a Grade 00 for the plate and flat for the base of the level.

I have yet to see one of those Chinese black granite plates that was not "00".

I think there is a link further back in this thread showing where the three supporting points should be with regard to the main axis (2) and the ends/edges of the plate. It is possible to get two results as you don't seem to have any indication where the single point is on the longer axis.

But even if you guess wrong I can't see it mattering much - if any - that needs to concern you.

The even mark-up you got consistently all over the plate seems to indicate a very high order of flatness.

I would certainly be prepared to take the chance and assume that it is Grade 00.

Your Old Dog
06-12-2010, 07:16 AM
Can we move on to knurling now?

JCHannum
06-12-2010, 08:08 AM
I have yet to see one of those Chinese black granite plates that was not "00".

How many of those black granite surface plates have you inspected, what means have you used to determine they were "00"?

oldtiffie
06-12-2010, 08:17 AM
I saw the certificates - what else?

https://www.machineryhouse.com.au/Products?stockCode=Q810

Our NATA (National Asssociation of Testing Authorities) = USA NIST

Forrest Addy
06-12-2010, 08:37 AM
$AU 795 for a 640 x 640 surface plate? Wow! Here's a US source for about the same thing.

http://grizzly.com/products/24-x-24-x-3-Granite-Surface-Plate-No-Ledge/G9656

$US 110 for a 24" x 24" (no stand). What would shipping and duty for 234 lb to Oz would be?

luthor
06-12-2010, 08:55 AM
I saw the certificates - what else?

https://www.machineryhouse.com.au/Products?stockCode=Q810

Our NATA (National Asssociation of Testing Authorities) = USA NIST


Tiffie, are you saying that the Hare and Forbes Chinese surface plates come with NATA test certificates? I doubt that they would as H&F are not known for importing high quality equipment and the documentation that comes with the equipment leaves a lot to be desired.

oldtiffie
06-12-2010, 09:00 AM
$AU 795 for a 640 x 640 surface plate? Wow! Here's a US source for about the same thing.

http://grizzly.com/products/24-x-24-x-3-Granite-Surface-Plate-No-Ledge/G9656

$US 110 for a 24" x 24" (no stand). What would shipping and duty for 234 lb to Oz would be?

That's a Grade "B" plate Forrest - two grades less than "0".

Your question seems to infer that it is likely that that probably Chinese plate is all that it says it is, but I notice that there is no mention of a certificate - or the mounting points, or the "plot" - being provided either in Grizzly catalogue page 661:
http://www.grizzly.com/catalog/2010/Main/661

The Grade 0 item in OZ is a good 4"/100mm thick as well.

But I am satisfied with mine and I guess that many who buy the Grizzly product will be too - I hope they know and take account of the accuracy limits and the seeming lack of the plot.

My guess is that most will use it for marking-out and comparator work and it will be more than adequate for that - and just about everything else in many HSM shops as well.

oldtiffie
06-12-2010, 09:19 AM
Originally Posted by oldtiffie
I saw the certificates - what else?

https://www.machineryhouse.com.au/Pr...stockCode=Q810

Our NATA (National Association of Testing Authorities) = USA NIST

Tiffie, are you saying that the Hare and Forbes Chinese surface plates come with NATA test certificates? I doubt that they would as H&F are not known for importing high quality equipment and the documentation that comes with the equipment leaves a lot to be desired.

Here is the text of the H&F "Specification" for that surface plate:


Granite Surface Plate

Granite surface plates have virtually replaced the cast iron plates which were used before World War II because of price and maintenance costs, corrosion (rust) resistance, superior accuracy, less response to temperature changes, no humping when damaged by a severe blow, warp resistance, non-seizing surface, and non-magnetic. Because of granite's lower coefficient of thermal expansion, granite will respond less than cast iron (steel) to temperature changes, which may distort the accuracy of the surface plate.

PRECISION GRANITE SURFACE PLATES are made of black granite which is stronger and denser than light colored granite, more stable, and wears longer. The granite properties determine the stiffness of the granite and accuracies can be guaranteed with less thickness than other types of granite. This translates to less freight and handling whenever the plate is shipped or moved.

MeasuMAX Granite Surface Plate include an adjustable levelling metal stand. The granite surface plate sits on 3 adjustable jacks for easy levelling
Each plate is calibrated in a controlled environment at 20°

Features

• Manufactured to Precision DIN standards Grade 'O'
• Flatness within 3.253/μm
• Hard wearing, non corrosive & non magnetic
• Test certificate supplied
• Does not include NATA certification

I did not say or infer that a plate from H&F came with a NATA certificate. I replied in the context of the last item(which I have coloured red) which said that a NIST certificate was not provided and that here in OZ our NATA is cross-accredited with the USA NIST.

It is clear that no NATA/NIST certificate is provided.

I have seen and am satisfied with that plate at H&F and the Chinese certificate provided and have no good reason to say or suspect that it maybe or is otherwise.

I did buy mine in OZ - but not from H&F as it was before they stocked them.

I have never claimed, nor have I had reason to claim or say, that H&F is less than perfect. It possibly is.

Based on what I've seen of that plate at H&F I can be confident that it is what it purports to be.

I have bought a fair bit of stuff over time from H&F and in just about all cases the stuff performed quite adequately in its intended use and purpose.

Others may differ as is their right.

JCHannum
06-12-2010, 09:47 AM
Granted that they are probably adequate for most laying out and HSM tasks, but they are far from being guaranteed as being "00". Chicom certificates issued with their products are relatively meaningless mass produced copies. Most contain no information such as serial number or other meaningful data relating to equipment used or it's calibration.

Adequate performance in just about all cases is hardly a ringing endorsement for H&F quality. Sounds like the typical crapshoot when buying Chicom goods.

oldtiffie
06-12-2010, 08:34 PM
Sure Jim.

As I am pig-ignorant and given that I would not know any better and was "conned" and "sold a pup" ("gold brick"??) and not able to see a bad deal if I fell over it, it seems that your are right - as always.

Ignorance is bliss.

I'm happy.

Fasttrack
06-12-2010, 10:27 PM
Can we move on to knurling now?


Trouble maker. :D

oldtiffie
06-13-2010, 01:24 AM
I thought I was moving towards knuckling.

gmatov
06-13-2010, 01:56 AM
Granted that they are probably adequate for most laying out and HSM tasks, but they are far from being guaranteed as being "00". Chicom certificates issued with their products are relatively meaningless mass produced copies. Most contain no information such as serial number or other meaningful data relating to equipment used or it's calibration."

I think most of your Mitus would be "mass produced", too, tho' they might come with a "mass produced" certification.

They surely are not hand made, and Sven signs all the certificates.

I don't THINK they are any more accurate, but what do I know? You could possibly split an electrical impulse closer with a Swiss instrument, 150 bucks, than a Chinese instrument, that MAY actually MAKE the SWISS instrument, today.

That you guys WANT to deny that a CHINEE caliper could EVER equal your BIG BUCKS tools is ridiculous. I know, the batteries die too quick, they gotta be junk. I buy them batteries for 8 for a buck. BFD.

And they caliper to the same OD as my Starett mikes do. My Lufkin's TOO.

Cheers,

George

.RC.
06-13-2010, 02:25 AM
Granted that they are probably adequate for most laying out and HSM tasks, but they are far from being guaranteed as being "00". Chicom certificates issued with their products are relatively meaningless mass produced copies. Most contain no information such as serial number or other meaningful data relating to equipment used or it's calibration."

I

Any certificate even from Starrett is totally meaningless unless it can be traced back to being from a certified lab..

Here in Oz we probably have more protection for buyers then in the US, as such our chinese made goods are probably of a higher quality... If within seven days the purchased goods to not perform as advertised we have a legislated guaranteed money back guarantee.

I have a mitutoyo 0-25mm micrometer, a Tesa 0-25mm micrometer and a no name chinese micrometer... All three show the same measurement when I test them against each other...

PeteF
06-13-2010, 03:36 AM
OT but I had a lot of trouble finding surface plates down here such as sold by Grizzly, Enco etc etc that you guys have cheap access to. Used pretty much never come up. In the end I bought one from here http://www.mcjing.com.au/searchresult.aspx?keyword=surface%20plate I got the largest granite, seems to be ok, and came with a map, unremarkably with the same "inspector's initials" as the one reproduced here. In fact I will have to drag out my map and compare to see if the map is just a photocopy :D Seriously though, I have no reason to suspect it is anything other than what they claim it to be.

Pete

.RC.
06-13-2010, 05:08 AM
Hi Pete, I would be interested to see if your surface plate map is the same as Tiffie's..

oldtiffie
06-13-2010, 05:13 AM
That's where I got mine Peter.

Top people, top service, no nonsense, and know their stuff. They are a Chinese married couple, both with (Chinese) University Engineering degrees - has one in Metrology. They go back to China at least once a year for "shopping" so I guess they know what they are after and how to get it at a good price. I've been quite happy with other stuff I've bought as well.

I first met them when they had a stall at an Engineering Expo/show at "Geoff's Shed" (Melbourne, OZ) and was very impressed. I asked the wife what the standards were for engineers squares and a couple off other items - and she reeled them right off - spot on.

I've been over my plate with a 3um (3 micron ~ 120 millionth of an inch ~ 0.000120") test dial indicator and I cannot see any movement in it - and I don't think that indicator is stuck either.
https://www.machineryhouse.com.au/Products?stockCode=Q223

I am satisfied that it is what it purports to be and is more than adequate for mu purposes.

It rarely gets used as I prefer a sheet of float glass - with an accuracy of the order of 0.0001"/" or better - on my mill table. I use a surface guage instead of my digital height guage as its easier to move around and can get into some tight corners that the digital height guage cannot.

I rarely work to a "tenth" but if the need arises I still use my "glass" plate and if in doubt I use the Chinese granite plate.

PeteF
06-13-2010, 07:24 AM
Hi Pete, I would be interested to see if your surface plate map is the same as Tiffie's..

I can't post a picture of it too easily but got the map out and compared it to the one posted here. At first I thought "here we go", as I mentioned, the A/B inspectors were the same, and the same DATE!!! Then I saw the serial numbers were 10 numbers different. Mine is a larger plate so it has more points. But no, the map is completely different. I know when I got the plate back here I was disappointed to see the high and low were right next to each other, if I was getting another I'd ask to go through and select the one with the best map, or at least reject one like this. Having said that, we're talking, in my case 3.3 um here so it's not exactly the grand canyon! BTW the spec to make grade is different on the larger plates, I think 3.5 um from memory, but I no longer have the map here to check. If I had the facilities to reliably map the plate myself I'd simply try not to use that point with the "dip", but I think that's splitting hairs that don't need to be split.

I also first saw them at a show somewhere, yeah they're good people, a whole family affair. I see precisely the same products when I'm up in Hong Kong/China and can assure you they're not exactly starving to death on the prices, but that's fine, they keep a very large range in their store and answer emails well if you use the net.

Pete

Pete

oldtiffie
06-13-2010, 08:04 AM
Ringer,

back to the topic of the base of your Starrett 199 level.

I think that you can take it that your level is OK. If it is as original, and given the Starrett manufactures them with an upward curve in them, you can only reasonably expect the "blue" to "show" on the ends of the scraped base-plate. If it is the same all over the surface plate it seems to indicates that the surface plate is pretty good too. Short of inverting the level and accurately setting the ends of the level scraped base parallel to the surface plate and going over it with a good indicator - perhaps on a good height guage - I can't see any other practical way of measuring that "bow" in the level base.

My guess is that Starrett made it that way so that provided that any downward distortion of the level did not exceed the bow the level would always bear on the ends. It would only "rock" if the distortion exceeded the bow.

I'd guess too that Starrett had to make a considered decision to limit the "bow" so that use of the level other than at its ends would not significantly affect the accuracy of it.

Given that all other things being correct, the warmth from the sun, lighting or heating etc. - or even the wind etc. can "draw" a bubble in a good level, a lot of the worry is pretty well needless.

I get by most times with a pair of good carpenters/builders levels - the shorter one is 1mm/metre (1:1,000) and longer one is 0.5mm/m (1:2,000).

1:2,000 is 0.5/1,000 or "half a thou per inch" which atan of 0.0573 degrees ~ 3.4 minutes of arc which is twice as good as a normal protractor which is calibrated at 0.1 degree = 6 minute intervals.

Levels of the order of accuracy of the 199 are a PITA as it seems to take forever for the bubble to "settle". The 1:1,000 and 1:2,000 levels settle straight away and don't have (m?)any of those PITA issues.

gwilson
06-13-2010, 12:54 PM
I suggest marking the edge of the base with a Sharpie. Then,when resting the 199 on 2 parallels across a lathe bed,place the parallels the same distance apart visually,and rest the level on the sharpie mark on one of the parallels. That's what I do. I could do without the concave sole,though.

Paul Alciatore
06-13-2010, 02:36 PM
Greg..... to get a difference, the TOP of the level needs to be at a different temp than the bottom.

Differences across the length will have minimal effect, I'd suspect.


"The TOP of the level needs to be at a different temp". This is a bad assumption that has been made here by several posters. Actually there are other modes of temperature warping, INCLUDING the far more likely scenerio of warming the CENTER as opposed to the ends. A careless user grabing the center firmly in a top to bottom manner, perhaps to not drop a costly instrument, would put heat in the center. This would cause the level to become barrel shaped due to the center portion expanding in all directions, including DOWN.

Sorry if this comment is somewhat late, I am just re-reading this thread.

Bill736
06-13-2010, 03:20 PM
Well, somebody needs to do the defining test; place your level ( wearing gloves) on a surface plate, observe the amount of gap in the middle, and then put your hot little hand on the middle of the level a few times and see if there's any flattening of the level. It occurs to me that if you only handle the level at the top , the amount of curvature might increase and not decrease. Probably the moral of this whole story is that, if you work within limits of
.0005 inches, lots of things can change from day to day, and job to job, and consistency is a more useful concept than absolute accuracy.

oldtiffie
06-13-2010, 08:05 PM
Well, somebody needs to do the defining test; place your level ( wearing gloves) on a surface plate, observe the amount of gap in the middle, and then put your hot little hand on the middle of the level a few times and see if there's any flattening of the level. It occurs to me that if you only handle the level at the top , the amount of curvature might increase and not decrease. Probably the moral of this whole story is that, if you work within limits of
.0005 inches, lots of things can change from day to day, and job to job, and consistency is a more useful concept than absolute accuracy.

Good reply Bill736 - and a very good "heads up".

Here are the Starrett specs for the 199 level:
http://www.starrett.com/download/338_p439_444.pdf

15" long and an accuracy of 0.0005/12" = half a thou per foot = 0.00004"/" ~ "half a tenth" per inch.

Unless you are in a very controlled environment there are far too many uncontrolled and unknown variables to be be certain of absolute accuracy, consistency and repeatability.

Checking the level and (re) adjusting it would be a good start.

While I can see the need for that level of accuracy as a challenge, and possibly during initial setting-up, I can't see the justification for the assumption that that level of accuracy will be maintained at all times that the lathe is used either.

Similarly, I can't see the implied need for it every time a machine - a lathe specifically - is used. I can't see that is even probable that once set up for the implied need for accuracy that it will be that accurate every time the lathe is used. And that in turn suggests that the lathe needs to be checked and if necessary readjusted each and every time it is used or the temperature of the lathe or the shop changes significantly enough to potentially cause a "set up error that will show up on the 199".

Get real.

I don't see people getting that concerned about milling machines or surface or cylindrical grinders which are just as and probably more accurate and therefore more in need of adjustment and checking than a lathe is.

I suspect a touch of herd mentality - or paranoia - here because "every body (else??) does it or says ........................... ".

Perhaps its because a lathe is - or is seen to be - more easily "tweaked" whereas the other machines are either more difficult (or all but practically impossible?) to "adjust".

When was the last time you "needed to" or actually checked your mill table for "out of flat" (twist, bow etc.) - and if you did, what did - or could - you do about it? And did it really matter?

I can see the need for an occasional check if a problem seems to exist that really does need checking or remedial action - but not otherwise.

I only use a carpenters level for setting up the balancing rails for my surface grinder - that's plenty good enough.

I very occasionally use my "good" level to set my surface plate level - on my mill table - if I have need to really check something for level but other than that, I rarely need to use it. While I have the surface plate set up for those sort of jobs, I re-check, and if necessary, re-adjust the level if required. But the level is there - all set up - if, as and when required for anything else. It rarely sees the light of day (or the shop).

I set up my mills and grinders with a carpenters 1:2,000 level as really all the "level" is need for it for ease of table movement (same both ways) and drainage of cutting fluids.

A random spot check with my "good" level shows that the very easy set-up with the 1:2,000 level is not too bad at all - and its a bloody side easier.

I have no problems with accuracy - when and where its needed, necessary and effective - but not otherwise. I sure don't get obsessed about it as it gets used only if and when needed and not otherwise.

Knowing when accuracy is needed and using it requires knowing when its not needed and giving it a "miss" as well.

The 199 is a nice instrument or tool - but that's all that it is.

Keep it in its box and treat it with the care a good instrument requires. Its box is not a tabernacle.

.RC.
06-13-2010, 08:28 PM
Tiffie you probably miss the point of accurate levelling of machine tools...

If you have a very accurately levelled machine when things go wrong it means you can disregard the machine being out of plane as a cause of the problem... With the multitude of variables that come with accurate machining isn't it better to be able to rule out at least one of those variables as being a cause of the problem.. Look at the expense gone into the installation of this floor mill, there is probably $50 000 gone into the foundation alone http://www.practicalmachinist.com/vb/cnc-machining/updated-cnc-floor-mill-install-pics-194142/ all so the machine stays level.

Also the likes of the Starret 199 and other high presicion machine levels are used for mapping out errors in slideways so you can ascertain wear and the best way to go about the rebuilding process...

The difference I see is that you make money out of your machining work, therefore you have to be productive, for you to chase microns is pointless as you want to get the job out the door for the minimum amount of time and outlay and get paid..

Most of us here do it for the love of removing metal and as a hobby, and if we want to chase microns for the fun of it even though there is no rational reason to then that is our choice...

oldtiffie
06-13-2010, 11:12 PM
Tiffie you probably miss the point of accurate levelling of machine tools...

If you have a very accurately levelled machine when things go wrong it means you can disregard the machine being out of plane as a cause of the problem... With the multitude of variables that come with accurate machining isn't it better to be able to rule out at least one of those variables as being a cause of the problem.. Look at the expense gone into the installation of this floor mill, there is probably $50 000 gone into the foundation alone http://www.practicalmachinist.com/vb/cnc-machining/updated-cnc-floor-mill-install-pics-194142/ all so the machine stays level.

Also the likes of the Starret 199 and other high presicion machine levels are used for mapping out errors in slideways so you can ascertain wear and the best way to go about the rebuilding process...

The difference I see is that you make money out of your machining work, therefore you have to be productive, for you to chase microns is pointless as you want to get the job out the door for the minimum amount of time and outlay and get paid..

Most of us here do it for the love of removing metal and as a hobby, and if we want to chase microns for the fun of it even though there is no rational reason to then that is our choice...


Ringer.

I don't think I've missed it at all.

I have no problems with an as accurate an initial installation as you can reasonably get.

I've said that.

The condition/s at the time of installation is/are the datum or basis for every subsequent change of temperature, wind heat/cool chill and any changes in the ground conditions under the slab etc.

The example you linked to at PM pretty well makes my point.

In less than optimal conditions or change of conditions, it is a "big ask" to expect that there will be no changes that will or may not substantially change the levels etc. in the lathe in a normal or usual shed or shop on a normal slab . There is no realistic point in checking "in case there is a change" unless it is really necessary - which implies that change requires the machine to be re-set up to "as original". You'd be doing it with the change of seasons. Anything that causes the slab surface to move in other than a non-differential way will cause the, lathe supports and the lathe to move differentially. In most cases it just doesn't matter but if you are concerned about any differential movement of the order of the limit of accuracy of that 199 level you potentially have a pretty severe "leveling" problem.

Using the level requires that there be a datum set-up (which also needs constant checking) to refer the level back to at regular periods or change of conditions.

Most just accept that even if an initial leveling and set-up is "spot on" that it is quite possible that it may not always stay that way.

PT Doc
03-01-2019, 11:13 PM
Very interesting discussion about this level and the concavity. I bought a basically new 199 and contacted Starrett about proper calibration. They forwarded me some i formation. I have a photo of the doc but can’t post it here. Anyways, they state in spiriting the concavity is less than 0.0002”. I can forward the phot of the doc if someone want something it of can post it here. It could be helpful to others. I have to admit I have not read the 12 pages of this post so maybe my info is already here some place.
Here is a cut and paste job.

LEVEL

Cat. #199​

Gages and Equipment Used -
​Surface Plate with Leveling fixture
​.0002” Onion Skin Paper

Accuracy Requirements -
​Vials shall be set and adjusted to center and reverse
​Base surface shall be hollow with maximum of .0002” clearance
​Base surface scraped with minimum of 12 bearing points per inch.


​Vial Sensitivity:​ Elevation D / foot / 1/10“ of bubble movement
​Main​7.5 - 9.5 seconds​ .00044' - .00055"
​Cross Test​ 6 - 7 minutes​ .0210" - .0245"


Procedure -

​1. Appearance - Inspect for quality of painting, defects in casting such as blow-holes,
​ porosity, etc., and marking on top plate. The vial shall be clear with distinct
graduations. The scraping of the base shall be sufficiently fine as to produce a
minimum of 12 bearing points per square inch.

​2. Inspection on Surface Plate - Using a leveled surface plate the level shall be placed on
the surface plate and both vials shall indicate a level condition. The level shall be
reversed 180 degrees with no significant change in the position of the bubble in the
vials.

​3. Base Flatness - With the level resting on a surface plate there should be a slight drag
on both ends. The hollow in the middle should not exceed .0002”. Inspect the
concavity using the .0002” onion skin paper.

​4. Check vial sensitivity by moving dial on leveling plate until bubble in main vial
moves 1 graduation (1/10"). Take reading off of the dial to determine how much
change (in seconds) in elevation it took to move the bubble 1/10".

softtail
03-02-2019, 08:26 AM
I didn't read through all 13 pages of posts so this may be redundant, but Tom Lipton (oxtool) did a vid or two on scraping in a level and talked about the concavity.. makes perfect sense. If memory serves he showed how the blue should look when pulled from the plate.