PDA

View Full Version : Flatness criteria for surface plates.



Forrest Addy
07-18-2009, 03:18 AM
Someone was asking what the flatness tolerance for the different grades of surface plates.

Here's the flatness tolerance formula:

• Laboratory Grade AA: (40 + diagonal [in inches] of surface plate squared / 25) x 0.000001 inches.
• Inspection Grade A: Laboratory Grade AA x 2
• Inspection Grade B: Laboratory Grade AA x 4

Source: http://www.mahr.com/index.php?NodeID...Language=EN-US

oldtiffie
07-18-2009, 08:47 AM
These further informations may assist:
http://www.tru-stone.com/pages/faq.asp

http://www.tru-stone.com/pdf/Fed%20Spec%20GGG-P-463c.pdf

Don't get "thrown" by millionths of an inch as there are 100 of them in a "tenth" (0.0001") and 1,000 of them in a "thou" (0.001").

These "small" numbers can be just as over-awing as the large number of digits after the decimal point on a calculator. If working in "inches", its rarely that you'd need to go past the fourth significant figure after the decimal point - ie 0.0001" or 6.1234" as most times the third significant figure - ie 0.001" or 6.123" is quite adequate.

Working in millimetres (mm) and micrometers (1,000 per mm) is similar. The normal significant figure after the decimal point is two - ie 0.01mm (~0.0004") or if "finer" use 3 significant figures - ie 0.001mm (~0.00004").

To use the formula:

Laboratory Grade AA = (40 + diagonal squared/25) x .000001" (unilateral)
Inspection Grade A = Laboratory Grade AA x 2
Tool Room Grade B = Laboratory Grade AA x 4.

The AA plate accuracy over say 4" is:

(40 + (4^2/25)) X 0.000001"

= (40 + 16/25) X 0.000001"

= (40 + 0.64) X 0.000001"

= 40.64 X 0.000001"

or approximately 0.00004" per 4"

and for a Grade A plate it is 2 X 0.00004 = 0.00008" per 4"

and for a Grade B plate it is 4 X 0.00004 + 0.00016" per 4"

So the "Grade B" plates that are readily available are pretty good and should be more than adequate for Tool Rooms and most HSM and "Pro" shops.

Your Old Dog
07-18-2009, 09:21 AM
My 12x18 from Enco at $25.00 and free shipping must have been the much sought after patio grade.

oldtiffie
07-18-2009, 09:25 AM
Don't think so YOD.

I'd bet it would be Grade B - Tool-Room.

Goes hand in glove with Grade B slip guages (aka "Jo Blocks").

loose nut
07-18-2009, 04:01 PM
Still better then a piece of plate glass and probably cheaper.

DICKEYBIRD
07-18-2009, 05:39 PM
My 12x18 from Enco at $25.00 and free shipping must have been the much sought after patio grade.LOL:) Dangit! You made me blow a swig of adult bevarage all over my monitor with that one.

I just got the same one from Enco as well. I never thought about it being 3" thick and over 70 lbs. Quite a lump to move around! I think they're a bit better than "patio grade" though. Plenty good for my needs.

oldtiffie
07-18-2009, 06:08 PM
Still better then a piece of plate glass and probably cheaper.

Probably so.

I use ground float glass as thick as I can get it from my friendly local Glazier - straight from his scrap pile -and the thicker the better. He cuts it as required and grinds all faces and "breaks" all edges (arrases) or bevels them as required just for beer money.

He says that a lot of local people (motor cycles etc.) who need to "do up" a cylinder head will use it as a lap with "wet and dry" emery paper. Some small home and jobbing shops use it for setting out or setting up.

I use it to set up as a surface plate.

Its much easier to handle and store than a surface plate - mine lives on it edge on the floor at the side of my bench.

Here is mine. I have not attempted to measure it to the n^th degree but it is more than adequate for any job I've used it for. Other than that I use the top face of one of my magnetic chucks or the top face of my universal grinder - and the table of my mills.
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Dig_Ht_Gg1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Dig_Ht_Gg2.jpg

Granite surface plates have been generally unobtainable at pretty well any cost let alone cheaply in OZ until fairly lately. These are all I've seen - although I haven't looked too hard for one lately. An accuracy of 6 micrometres (~ 240 millions of an inch ~ 0.000024 ~ 1/4 of a "tenth" is pretty good.
http://mcjing.com.au/searchresult.aspx?keyword=surface%20plate

Thomas Staubo
07-18-2009, 06:24 PM
How thick are your glass plate Tiffie? (in mm)


.

darryl
07-18-2009, 09:06 PM
Patio grade, ha ha. For some time I had been using the ATVB grade- a valve body from an automatic tranny. I built a wooden box, then put a layer of concrete in the bottom, set the valve body into that, then filled in with concrete. The milled surface of the valve body sits higher than the concrete so I ended up with a 'flat' that's full of tapped holes to mount stuff to.

Now that I have an actual granite surface plate I should probably check the flatness of that old one. I found it most useful when assembling housings etc that needed to be accurately bolted together without twist, and for fixing sandpaper to when needing somenting sanded perfectly flat.

I actually built two of them. Just today I used them both, but not for the intended purpose. They're holding down some rubber belting material. I'm hoping I can get most of the warpage out of them so I can use them in front of the machines as mats.

oldtiffie
07-18-2009, 09:15 PM
Thomas.

The glass must be "float" (and not "drawn") glass. Float is the "norm" now anyway. Get it as thick as is reasonable at the time - for obvious reasons

The piece shown is 12mm thick. I have another smaller bit that is 10mm thick that I use on the mill table. I will post a pic or so later.

I will also address some "issues" about the use of a surface plate in a HSM shop.

lane
07-18-2009, 09:34 PM
Heck I just use my mill table. Like a old boss used to say. It`s as flat as piss on a plate.

oldtiffie
07-19-2009, 03:41 AM
Yep - I agree with you Lane as many times there is a lot of undue concern about and reverence toward the common shop surface plate. Same applies to digital and vernier height guages. And too little to the common surface guage and Machinist's (6" and 12") rules.

And all too often with no justification other than someone - or a lot of someones - said so.

Following on from my previous post to Thomas Staubo, I thought I'd provide the glass "surface plate" sizes:
- 12mm: 600 X 600mm;
- 10mm" 600 X 300mm.

And some pics and comments:

Here are my digital height guage and surface guage, each with a 0.01mm (~0.0004") test dial indicator in them and the TDI's resting on my 2-3-4 blocks and all of it on my 600 X 300 X 10mm float glass "surface plate" on my HF-45 mill table. Also on the glass table is a stock standard 200 X 50 X 25mm (8" X 2" x 1") fine and smooth (and very flat) oil-stone (China). The scriber - sitting on the base of the digital height guage base - belongs to the surface guage.

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Surfaceguage1.jpg

The oil-stone is rubbed over the mill table to remove any burrs or "dings" so that the table is restored to its "flat" condition for the glass plate to rest upon.

Both the height guage and the surface guage are just as accurate as each other for use a comparators as shown. The surface guage is much easier to use as both a comparator and for marking out than the height guage. I can put a second extension "cross" mast on the surface guage which is also easier to get into tight spaces. All of my DTI adaptors will fit both the height guage and the surface guage. I have a range of different length (8mm) masts for the surface guage.

I can very accurately set the DTI on the front of my surface guage to measure backward/forward motion but I can't do it with my digital height guage.

Generally speaking the surface guage is at least as if not more useful than the digital height guage.

Here are a couple more pics of the same set-up:
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Surfaceguage2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Surfaceguage3.jpg

Here is the surface guage doing what it really does do best - setting up a scriber against a rule for marking out. Sorry about the poor focus. The metric carpenters rule was used a good contrast.
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Surfaceguage4.jpg

When a comparator is used on a surface plate, the comparator should stay where it is - and NOT move over the job to be measured. The job should move under the comparator. A better result will be more assured and easier to get.

A comparator does not measure the height of anything really. What it does do is to compare a "master" (to which the DTI dial is "zeroed") with a job to be tested. Provided that the arm of the DTI is not more then 2 (3 or 4 at a stretch) degrees from the horizontal the "dreaded co-sine error" is pretty well insignificant.

Similarly, if the height guage or surface guage moves off vertical and provided that the "error" is not more than 2 degrees the "co-sine" error is insignificant.

Cos 1 degree = 0.99985 and so the error is 1.00000 - 0.99985 = 0.00015" (1.5 "tenths") per inch height.

So to minimise "cosine error", use a flat surface (surface plate, glass sheet, or machine table) as well as keeping the height differentials as short a possible and you will be on your way rejoicing.

If you are using slip-guages, use as small a stack and as few guages as possible and you will minimise any likely real individual and collective errors there too. But even the garden variety (Class 2) slip guages that you can buy almost anywhere are very accurate indeed for HSM purposes - they are with millionths of an inch. As previously, there are 1,000 millionths in a "thou" (0.001") and 100 of them in a "tenth" (0.0001").

Don't get "hung up" about surface plates, height guages or slip guages - as some do - as there is no need for it. Just be careful as they are relatively precision tools - just as a good DRO (which is accurate to "Two tenths" = 0.0002") and nobody seems too concerned about that - nor should they be.

Just be realistic in your needs and expectations for the job in hand, the tools you make it with - and the tools you measure it with.

In short, Lane is dead right.

Forrest Addy
07-19-2009, 05:13 AM
Yup. The punishment has to fit the crime. This is an alligory for tolerances have to suit the end requirements the parts are destined for. Those of you who poo-poo accuracies closer than you feel you need to work should bear in mind that some home shop machinists work to closer accuracies when their work requires it. The people seeking these accuracies naturally concern themselves with equipment refinements that makes it possible to attain them. This isn't a competition for status or glory. We're all working towards our respective goals.

Plate glass, even a thick piece of glass 1/2 or 12 mm) is thin when flatness is considered over an area. Glass is one of the stiffest of common materials but it will flex in rough as a function of its thickness to length ratio if not supported on a surface known to be flat. A mill table used as a flat substrate (a mill table in good shape is pretty flat) or a table saw (seldom very flat at all even if new) is fine if it does the job. Glass works pretty slick but there are hazards like it's easy to scratch, chip, and break. It's also too smooth to use to blue stuff on its surface.

I omit the glass myself when I use a machine table for general layout. I work directly off the machine table. Explanation: I hate glass where it has to be handled. I was in a car wreck when I was a sprout. I had to pick glass fragments out of my hide for years afterwards. Thus, I can't abide using plate glass as a shop tool. Break it and there's a zillion little pieces of glass everywhere.

Those of us who are developing skill at precision scraping have legitimate need for flatness and straighness accuracy approaching the limits of granite surface plate specifications. That's why I posted the formula so them that needs it can benefit.

Finally work only to the accuracy you need to ensure part function. Accuracy is generally expensive in time and money so strive for it with reason and the larger concern in mind. I've built crab pots and boat trailers, spindles and machine slides and never got hung up chasing for inappropiate accuracy in either.

oldtiffie
07-19-2009, 07:03 AM
Thanks Forrest.

Well said and timely too.

The reason I gave the "formulas" a working over was to put them in terms that some who may not have the math can relate to.

I have nothing against the use or having a surface plate at all.

Its just that on many occasions - as you say - the degree of accuracy may not require it.

A surface plate is neither a crutch nor does it guarantee accuracy of a very high order unless used skillfully and correctly.

I only use a surface plate when it is all that will do the job required. I would only use "blue" on a surface plate when the flatness and finish of the job achieved else-where warranted or demanded it. In my experience, it is the flat surface of last resort - not the first or anywhere else in between.

One point I over-looked as regards using dial test indicators and their "feeler" or "finger" or arm or stylus etc. and that is to move either the work or the indicator such that the relative motion between the work and the DTI arm is such that it is along and not across the longitudinal axis of the feeler. This is to get the best "feel" and result whilst minimising excessive lateral force which may lead to increased risk of damage to the DTI.

The two TDI's I showed are identical. One is used as a "working" TDI (the older one) and the other (newer one) where I can be more assured of accurate and consistent results. They are both calibrated to and accurate to 0.01mm (~0.0004").

I have another that is calibrated to 0.001mm ( 1 micrometer) but which is accurate to 0.003mm (~0.00012") - which I have never had occasion to use - and not very likely to either. It's there "just in case" (which is bull-$hit as I just couldn't resist it - and don't ask about the cost - even I am embarrassed!!).

Forrest mentioned "tolerances" which is the difference between the limits that the job requires which are in two parts which are largely inter-dependent:
- physical size, straightness, roundness, concentricity and parallelism etc. etc. and;

- surface finish: the quality or "roughness" of the finish.

The degree of accuracy required for measurement will exceed the degree of accuracy required to make the job.

Surface plates play an integral and vital part in meeting many - but not all - of these criteria.

wierdscience
07-19-2009, 08:39 AM
Another factor here-

Is your basement or garage climate controlled and kept at a balmy 68*F constantly?
Do you wear wool gloves when handling your measuring tolls so heat from your fingers doesn't affect accuracy?

Some will claim accuracy to+/- .0001" in they're home shops and I have no doubts some of those do actually achieve that.But it isn't cheap,easy or common.

There is a definite reason why super high tolerance work is done in shops that look like clean rooms.

As mentioned before-

.01"=$$
.001"=$$$
.0001"=$$$$
.00001"$$$$$$$$$$

Not at all linear is it?

Forrest Addy
07-19-2009, 04:12 PM
Amplifying Wierd's post: 0.0001 over an inch part size can be a little touchy in the home shop but it can be done provided thermal equalibrium of part and apparatus is considered and thermal expansion is calculated if the part and apparatus are of materials have different expansion rates. 0.0001" in 10" is almost impossible to certify outside a climate controlled measuring room. So is 0.001 in 100". Accuracy isn't just about tolerance. It's about the size of the part feature in proportion to tolerance.

I once machined a seal ring prep on a reactor vessle weld mock-up. The papers held me to 0.0005' in 78" a little internal ledge about 1/8 high. This is nuclear stuff where everything is on paper even the mock-up. I hollered for the inspector and told him I couldnt reliably measure that accurately and asked for a tolerence change to something more attainable. There was a meeting of white hats at the machine. I challenged them to measure the semi-finished ID using the 8 ft vernier caliper we had to work with. Yes the it was calibrated for the dimension.

Naturally the white hats to a man handled the caliper with bare hands and it was like the poem about the seven blind men and the elephant. They got readings all over the map as the caliper warmed up even though most of them did a creditable job of measurement and read the vernier afer some study. Animated discussion like chickens disputing posession of a crippled grasshopper. Then I explained about temperature and the problem was how do I ensure the caliper is the same temperature the full length as well as that of the enormous part? I suggested they look at the tolerance for the seal ring, note its out of roundness, refer to the weld prep data and note that the clearance between seal dia and prep surface was 0.015"" to 0.025, then work up a tolerance that reflects these clearance "call-outs" (technical English is a dialect all its own).

A day later I got new paper superceding the paper I had. The critical paragraph gave me "...machine ledge to 0.030 to 0.050 clearance with the measured diameter of the seal ring." Cool! I could live with that. About 20 minutes later I was done and the feature was ready for sign-off.

wierdscience
07-19-2009, 10:37 PM
Exactly right Forrest,high tolerance machine work,heavy construction and crane work often require a group brain jerk before any move is made.Better to blow a weeks wages in planning than FUBAR a multi-million dollar assembly.

oldtiffie
07-19-2009, 11:03 PM
Another factor here-

Is your basement or garage climate controlled and kept at a balmy 68*F constantly?
Do you wear wool gloves when handling your measuring tolls so heat from your fingers doesn't affect accuracy?

Some will claim accuracy to+/- .0001" in they're home shops and I have no doubts some of those do actually achieve that.But it isn't cheap,easy or common.

There is a definite reason why super high tolerance work is done in shops that look like clean rooms.

As mentioned before-

.01"=$$
.001"=$$$
.0001"=$$$$
.00001"$$$$$$$$$$

Not at all linear is it?

Good and very topical post WS.

Forrest was pretty well saying the same thing.

Its surprising how many either don't know about or don't use "Tolerance Tables" which will give the limits to work between for both the "shaft" and the "hole" for a given type of "Fit" over a given ranges of "diameters".

They are there to be used - and they are VERY useful.

There is no point in working to a "tenth or so" if the difference between the limits is say "3 thou" (0.003").

My "rule of thumb" is to leave at least half to go to get to a limit so that in case of a "boo-boo" I still have metal to remove.

Its generally easier to turn a shaft to size (and measure it) than it is to bore a hole. I am inclined to bore the hole first and make the shaft to suit.

If I am making one part to suit an existing mating part (which is "fixed" dimensionally) I will combine limits of the shaft and hole ("Limits and Fits" tables) and "use up" the unused limits on the "fixed/existing" part that I am making a part to fit. All of a sudden the limits are MUCH wider, the "Fit" is still as good as it ever was and I have a MUCH easier time of it.

I see very little mention of the "Class of Finish" either as quite often that is more expensive or necessary than the class of fit.

If I think of it later in the day, I will post a list of time/cost versus finish and the range of finishes that are commercially reasonable for different types of machines/machining.

oldtiffie
07-19-2009, 11:13 PM
Someone was asking what the flatness tolerance for the different grades of surface plates.

Here's the flatness tolerance formula:

Laboratory Grade AA: (40 + diagonal [in inches] of surface plate squared / 25) x 0.000001 inches.
Inspection Grade A: Laboratory Grade AA x 2
Inspection Grade B: Laboratory Grade AA x 4

Source: http://www.mahr.com/index.php?NodeID...Language=EN-US

That's Forrest's OP.

I would think that the Grade "B" surface plate (size to suit yourself) would be quite adequate for just about any HSM shop.

Grade "B" is also known at the "Tool-Room" surface plate and they are readily available.

In my case, I'd seriously have to question as to why I'd ever need anything better as I don't have and can't see me ever needing anything of a higher grade or better.

My "glass plate" does pretty well all the "day to day" stuff that I need a surface plate for.

J Tiers
07-19-2009, 11:26 PM
Next time someone starts talking about measuring to tenths (or the famous "microns") with anything, there is something for them to do.

Set a tenths reading 5" mic to a standard. Then check it 30 minutes later.

That small and oh-so-simple task is probably going to cause a wee bit of trouble.

However, a granite flat isn't precision in the same way. it has ONE characteristic, it is "very flat".

Consequently, it is quite useful when measuring different parts of an item from a plane, like a gasket surface etc... any set of points defining a plane, that is at or outside the body of the part.

NOT ONE THING says that the measurement must be to those "microns", you can measure to the nearest 0.005" if you want.

it is the one case where the precision of the device can be used regardless of the precision of the measurement you want. All you need to know is that it is enough better than your needed measurement to not bbe a problem.

That is something you do not know about float glass. You can roughly INFER it, you can guess it is true, etc, etc, etc, but you can't KNOW it. At least not unless you have something else to measure it against, which by definition you don't, assuming you need to use float glass.

If float glass is what you can get, so be it. Use what you have, and accept the problems and unknowns with it.

BTW, the person quizzing (perhaps somewhat sarcastically) about the temp controlled 68 degrees may need to remember that for like materials, the temp drops out. The mic and the piece of steel are going to both expand/contract together, and the measurement should be good at any reasonable temp.

Not so if your piece is aluminum, or plastic, etc and mic is steel. Then there will be some difference the farther off of 68 degrees you are.

oldtiffie
07-20-2009, 12:08 AM
.................................................. ..............

Set a tenths reading 5" mic to a standard. Then check it 30 minutes later.

That small and oh-so-simple task is probably going to cause a wee bit of trouble.

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

If float glass is what you can get, so be it. Use what you have, and accept the problems and unknowns with it.

BTW, the person quizzing (perhaps somewhat sarcastically) about the temp controlled 68 degrees may need to remember that for like materials, the temp drops out. The mic and the piece of steel are going to both expand/contract together, and the measurement should be good at any reasonable temp.

Not so if your piece is aluminum, or plastic, etc and mic is steel. Then there will be some difference the farther off of 68 degrees you are.

I am not at all certain that the statement about similar linear expansion/contraction with a micrometer is true.

If it were just a micrometer spindle and thimble, I'd probably agree.

But the micrometer in its frame is another matter.

Heating/cooling the "C" (or "G"??) frame will cause differential movement due to the differential distances around/along the opposite sides of the frame neutral axis. This movement will or may probably cause the axis of the micrometer spindle to no longer co-incide with that of the anvil bit the mating faces of the spindle and anvil may no longer be parallel.

No different to what I'd expect when clamping/unclamping a "C" or "G" clamp (such as my Chinese micrometers??).

Now given that all of my measuring stuff is Chinese, it surely would not matter if I used "float glass" for (my) "super accuracy" as I guess that if I bought a Chinese "Grade B" surface plate it would probably be a mediocre paving slab anyway (plain "wood-float" finish as "Standard" but extra of "steel-float" or "polished" finishes).

I don't think I'd fancy moving a sizable granite surface plate too often and they do not only take a lot of bench-space and need to be covered (with nothing else stacked on top) between uses.

Using a glass plate or just a "stoned off" machine/d surface for most work certainly opens the option of only needing a much smaller granite plate if any at all.

A surface plate is only as good as its last calibration. I am pretty sure that most here won't go to that process and the cost as I'd guess that they'd just trust to luck and "wing it" - just a I do with my "float" (so-called??) "surface plate/s".

Its entirely up to the owner/user as to which way he chooses to go for what he has to do.

There are many "adequate" options that may suit individual cases and circumstances.

Perhaps its different "down here" south of the Equator.

Forrest Addy
07-20-2009, 12:25 AM
The effect of temperature on precision measurement is well known to me. I can take readings at elevated temperature in bronze brass and aluminum and by referring to the alloys coefficient of linear expansion correct the raw reading back to 68 degrees F. I also handle precision apparatus with over mitts if need be to eliminate heat transferrence. I've miked small bushings in liquid N2 (gotta be quick) prior to inserting them into their intended bore.

All this is established skill gained over the years I've taught my apprentices and younger journeymen for longer I'd believe. It's not rocket science. It's bonehead physics tied to practical shop measurement.

And by the way I'd win a beer and a pizza on JR's challenge if he was proposing a bet. No, the raw reading may not be the same as the compensated reading I'd calculate and record on the cert but my mikes (I'll use any mike if it will repeat), thermometers, and other conditions will be carefully defined. Yes, I can make reliable measurements in an open shop holding 0.0001" within that determined in a metroloy lab. Here that JR?

J Tiers
07-20-2009, 12:29 AM
I am not at all certain that the statement about similar linear expansion/contraction with a micrometer is true.

If it were just a micrometer spindle and thimble, I'd probably agree.

But the micrometer in its frame is another matter.

Heating/cooling the "C" (or "G"??) frame will cause differential movement due to the differential distances around/along the opposite sides of the frame neutral axis. This movement will or may probably cause the axis of the micrometer spindle to no longer co-incide with that of the anvil bit the mating faces of the spindle and anvil may no longer be parallel.



I don't think that is right. Expansion isn't around corners, it is linear..... at any infinitesimal section, it expands a certain amount perpendicular to any line though the section. The net result is that it all gets bigger or smaller.

A ring gets bigger or smaller all as one, inside and outside diameter together.




Now given that all of my measuring stuff is Chinese, it surely would not matter if I used "float glass" for (my) "super accuracy" as I guess that if I bought a Chinese "Grade B" surface plate it would probably be a mediocre paving slab anyway (plain "wood-float" finish as "Standard" but extra of "steel-float" or "polished" finishes).



Actually, a flat surface is the easiest thing to produce..... to very high accuracy. three pieces lapped against each other do it.... they can't help it.

it's cheaper to make it right than to try to cheat it, probably.

Of course if you stop it short, or never check, you don't know.

But a flat surface is self-checking if you scrape another substantial-sized surface 'to" it..... move that piece around, and if it marks up the same in all positions, it confirms the flat surface.




I don't think I'd fancy moving a sizable granite surface plate too often and they do not only take a lot of bench-space and need to be covered (with nothing else stacked on top) between uses.

Using a glass plate or just a "stoned off" machine/d surface for most work certainly opens the option of only needing a much smaller granite plate if any at all.

A surface plate is only as good as its last calibration. I am pretty sure that most here won't go to that process and the cost as I'd guess that they'd just trust to luck and "wing it" - just a I do with my "float" (so-called??) "surface plate/s".


Not so. The "calibration" does not change the part (unlike a mic, etc), it merely confirms its status. It is still just as flat (or bowed)at 12:03 AM as it was at 11:57 PM, 3 minutes before the "calibration" expired, despite the fact that the 'calibration" is out-of-date.....

Granite is specifically used because it is stable, and hard. if you don't wear grooves in it, it will stay what it is to accuracy better than what you need..

I'd trust an older granite flat, if I was fairly sure it had not been worn un-flat, much farther than I would trust NEW float glass of 1/4th the thickness........

Use what you have. Just don't claim unjustified "microns" for it.

Doggie
07-20-2009, 12:30 AM
Believe it or not I have a friend that knows another friend, ETC. That engraves tombstones. And I talked them out of one that had a mistake on it! I just put the letter side down, and to me it is as flat, polished and smooth as you would want. It is 8" thick marble, square on the 3 sides (the top is curved). I can lay a strait-edge on it and cannot put a .001 feeler gage under the strait edge anywhere, no matter what direction. This place says that they have to occasionally throw one away because of a mistake or break. To me this might be a good source for everbody to get a "poor boys" bargain surface plate.
So the question is "whatta you want on your tombstone" :D

Your friend, Doggie

oldtiffie
07-20-2009, 12:47 AM
Doggie,

I had to smile at that post of yours.

And as you asked (I've posted these before but they are worth a repeat):
http://i200.photobucket.com/albums/aa294/oldtiffie/Funnies/morbidbastid.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Funnies/puter.jpg

Forrest Addy
07-20-2009, 02:14 AM
Jerry, "I don't think that is right. Expansion isn't around corners, it is linear..... at any infinitesimal section, it expands a certain amount perpendicular to any line though the section. The net result is that it all gets bigger or smaller." You are absolutely correct for a steady state situation.

Thing is temperature change takes a while to diffuse through-out the micrometer frame because steel is not the best conductor of heat. Thus a micrometer may move and shift on a 1 part in 10,000 scale for some time after it was last handled. So if a micrometer was handled without regard to heat input it may take a half hour or more for the temperature to diffuse enough that it has no material effect in the frame's geometry. It may still be warmer than the environment and thus affect a close tolerence measurement. However, if the micrometer frame has reached thermal equilibrium and its temperature is known, its readings can be compensated for.

Knowing that I seldom use a micrometer's as-calibrated accuracy in very close tolerance work. Instead I make up a stack of Jo blocks to the measurement desired and use the micrometer as a mechanical comparator. If the Jo blocks, the work, the micrometer, and a thermometer are sequestered under some form of insulation and protected from drafts and radiant heat a man with a good touch can take high confidence measurements accurate to 1 part in 50,000 with little fuss or feathers.

Cut a single corner in this linked chain of care and procedure and you might as well measure the work with a yard stick.

So a reading from the Jo block, a reading from the work, and a confirmation reading on the Jo-block leads to a measurement one can have confidence in - provided the readings have been adjusted for thermal expansion and material coefficients back to 68 degrees F or 20 degrees C.

For some things there are no shortcuts and 1 part in 50,000 accuracy mechanical measurements in the open shop is one of them. Fortunately most work is not this fussy and bare mikes and ordinary care are plenty accurate. But Jerry posed the situation.

But we're a long way from the flatness of surface plates.

oldtiffie
07-20-2009, 04:21 AM
Originally Posted by oldtiffie
I am not at all certain that the statement about similar linear expansion/contraction with a micrometer is true.

If it were just a micrometer spindle and thimble, I'd probably agree.

But the micrometer in its frame is another matter.

Heating/cooling the "C" (or "G"??) frame will cause differential movement due to the differential distances around/along the opposite sides of the frame neutral axis. This movement will or may probably cause the axis of the micrometer spindle to no longer co-incide with that of the anvil bit the mating faces of the spindle and anvil may no longer be parallel.



I don't think that is right. Expansion isn't around corners, it is linear..... at any infinitesimal section, it expands a certain amount perpendicular to any line though the section. The net result is that it all gets bigger or smaller.

A ring gets bigger or smaller all as one, inside and outside diameter together.


I don't think so. The expansion is omni-directional - not uni-directional. Expansion is assumed to be equal per degree rise/fall in all directions - which it is and only is if it is fully unconstrained. For example, if it is constrained and it increased in length by 1% it would have increased by 1% in all directions if it were a cube (or an oblong box) such that its volume would increase 1.01^3 = 1.0303.

Get a ring with a substantial thickness compared to its inner or outer diameter, cut it with a hack-saw at room temperature, then heat it slowly and informally and see what the slot does.

My guess is that the slot will open due to nominally equal enlargement per unit volume but as there are more units on the outer part than the inner the gap will (should??) open. Also assuming equality through-out there will be more expansion - including parallel to the neutral axis - on the outer parts than the inner of a part of that circle which can be likened or equated to the "C" frame of a micrometer - the more so with those that have large inner and outer flanges and a thinner web. (Like a "H"-iron aka RSJ).


Actually, a flat surface is the easiest thing to produce..... to very high accuracy. three pieces lapped against each other do it.... they can't help it.

it's cheaper to make it right than to try to cheat it, probably.

Of course if you stop it short, or never check, you don't know.

Now if you want to go to the expense of buying three good "shop-size" granite surface plate blanks ready to the final cut and then use them to lap each other in, then you are ahead of me.

I couldn't justify the cost of those three blocks not the space required and I sure can't see me lapping them. My guess is that if you were to take that on you'd have more muscle and grunt than Serena and Venus Williams (and a few others!!) all rolled into one!!!

While I have not had my (laminated) glass plates checked for flatness, I am quite confident that they are and will remain adequate for my purposes. I just either put them directly onto a good machine surface or else I put strips of Painter's masking tape at about 4" intervals across the machine surface.

I am quite coincident that my surface grinder tables and the top of my magnetic chucks are sufficiently flat as well.


But a flat surface is self-checking if you scrape another substantial-sized surface 'to" it..... move that piece around, and if it marks up the same in all positions, it confirms the flat surface.

That's so - no dispute there - providing the "flat" is not a very large concave or partial (hollow) sphere.


Granite is specifically used because it is stable, and hard. if you don't wear grooves in it, it will stay what it is to accuracy better than what you need..

No argument there.


I'd trust an older granite flat, if I was fairly sure it had not been worn un-flat, much farther than I would trust NEW float glass of 1/4th the thickness........

Just how sure would you be - or need to be - to be as sure as you seem to infer that you need to be? After all being "sure" (subjective?) is less definitive than being "certain" (objective).


Use what you have. Just don't claim unjustified "microns" for it.

I do and I don't.

I have no reason to think that my accuracy is any more or less than I've needed or will need.

I micron is actually a "micrometre" - ie 1 millionth of a metre (metre is correct - meter is US/American).
http://en.wikipedia.org/wiki/Micron

http://en.wikipedia.org/wiki/Micrometre

But as there are 1,000 of them in a mm and given that I ordinarily work to not much closer than 0.01mm (which ~ 0.0004") has 10 of them elusive micrometres I have enough of them to go round. If I work "closer" it's usually no closer than 0.005mm (~0.0002") or still 5 of them pesky micrometres.

As an example, the co-efficient of linear expansion for "steel" is 7.5 microinches per inch per degree F change at 68 deg F. (Machinery's Handbook 27).

So even for a 10 degree change in degrees F the total movement over that inch is 75 micro-inches which is 0.000075" aka 0.75 or 3/4 of a "Tenth" (of a "thou").

So realistically, while it may be either significant of not depending on the job, it is good shop practice to have no more than 10 or 20 degrees F temperature differential when measuring precisely or when fitting critical parts together. "Temperature equilibrium" to within a couple of degrees F is a big ask bit not as big an ask as having the ambient temperature accurately maintained at 68 degrees F.

Lane and Forrest have got it pretty right as regards the usual levels of "flatness" accuracy in any shop "flat plate" depending entirely on the job and the user.

Oh yes. As regards the "thin" 12mm and 10mm laminated float glass sheet "off-cuts". They are quite common sheet thicknesses. 15nn and 18mm are a bit harder to get off-cuts from and 20mm and upwards are pretty well impossible here. If I went closer into the Capital City here I daresay I could run around and try my luck but I've had no need to even think about it let alone try it to date.

I have no objections to surface plates at all - none. If I needed one I'd get one - pronto. If others have and/or need them - fine - as they know what's best for them in that regard and I am damn sure its not my place to tell 'em.

J Tiers
07-20-2009, 09:10 AM
Forrest: Note the original statement of the "test" for temperature effect knowledge......


Set a tenths reading 5" mic to a standard. Then check it 30 minutes later.


Tiffster:

1: I specifically said expansion is in all directions.... "perpendicular to ANY line through the section"

Your ring will expand all over, and become proportionately larger... as will the slot. What will NOT happen is the slot somehow becoming all distorted.


As for three flats... whatever would have suggested that you needed to get three and then lap them? Where did THAT come from? Seems to be a "magnification ad absurdam"

The point was that the manufacturer can make them flat very easily, and that the flatness is checkable in ways that do not "require" any advanced "micron measuring" devices.

if the manufacturer simply grinds them as flat as he wants, and then laps them in rotation groups of three, he will get to any flatness he cares to.

If you think a flat is uneven, you can check it by scraping another flat or a straightedge to it, and then applying that all over the original flat in various orientations. if the 'flat" is flat, it will blue up* evenly everywhere. if not, it will not.

Note that two will check almost every case, but two are NOT sufficient to actually generate a flat surface... or tell you who is wrong.

As for an older flat and not trusting it...?

One thing you KNOW about it...... it WAS flat at one time, to standards for whatever grade it is.

Your assessment task is to determine as best you can whether there is any condition which would have changed that.

With float glass, or a tombstone, or the roof of your 1924 Chevrolet, you have no particular reason to believe that it was ever more than "generally flat". Nobody ever intentionally MADE it flat, and any flatness it has is completely secondary to the process of making it...... In other words, it amounts to an "off data sheet spec", basically something you kinda figure must be pretty much true, but have no proof of.

probably fine for making 4 chair legs sit flat on a floor. After that you are guessing. if you are happy with that guess, more power to yah.

if you want to generate a flat surface yourself, you must make "B" fit "A", and then make "C" fit "B", and finally make "C" fit "A", and repeat until they all blue against each other. Only if all are flat will they all blue up* against each other.

*for the fussiest fitting, you wouldn't use actual "blue", as it is too thick

Falcon67
07-20-2009, 09:46 AM
Can't add anything to the thermal expansion posts, except that my bathroom scale says I have expanded about 2 lbs in the summer heat.

From my inspection days, if you need to test .001 then your equipment needs to be an order of 10 better - so your mic should be good to .0001. And on from there. For an HSM, patio grade surface plates and the inspection equipment we buy is in most cases quite good 'nuff - depending on the tolerances. A decent set of jo blocks and some reasonable cost calipers, mics, etc will work to around .001~.0005. After that, if you are trying to maintain .0001 or better for your own satisfaction, go for it - but unless you're instrumentation is certified to at least 10X better (by 10~1000X better equipment preferably traceable to NIST) and you have the certs on file, and it's 70F in the shop you are likely very "bench racing" your tolerances. That's OK - I know what my engine makes but if you ask it's "500 HP". ;)

The tightest tolerances I've see were in a shop I audited in the 'burbs of Chicago. They had a side job working with a university medical lab. They were working with these items because no one else wanted to mess with it and they were doing a little "giving back" by taking on the work. I'm not sure they were even paid for it. Some small rings for use in heart valves - 20 millionths was the required tolerance spread. In titanium. One guy, one very good machine, in a box, in a controlled environment. All the inspection tools were in there too. When he went in, he had to putt around for about an hour for the temps to stabilize in the workroom. The machinist did some low tolerance work while he waited - titanium catheter sample needles, .005 dia, .002 bore about .500 long. "You can hold the parts, but don't sneeze." A handful weighed nothing.

Forrest Addy
07-20-2009, 11:30 AM
Jerry, for 0.0001" in 5" I'll still bet you a beer and a pizza measuring a standard or a part size. When you handlle standards at that accuracy, working bare handed is like navigating with your enameled steel coffee mug on the compass binnacle.

When I'm done you might have reason to put an "actual size" sticker on your standard/ When working that close I insist on using a mike as a mechanical comparator in conjunction with a set stack of Jo blocks. That way the mike is a mere indtrument for comparason and the actual size is taken from the jo block stack assembled to null the mike reading. Yes, you can wring Jo blocks when wearing nitrile over cotton pall bearer gloves. Its not that I'm such a red hot inspector. It's metrology SOP for the open shop.

oldtiffie
07-20-2009, 07:10 PM
I agree with Forrest.

Here is one I use occasionally - 3-directional - to 0.01mm (~0.0004") which works pretty well but is expensive!!
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Taster-3D-1.jpg

And here is Forrest's point about using Jo-blocks (slip guages) as a reference for a digital caliper (accurate to 0.01mm ~ 0.0004") and a micrometer (metric) which are used as comparators. ie the caliper and the micrometer are referenced to the slip guage and then used to "measure" the job.
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Digital_caliper1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Digital_caliper2.jpg

The "reading" on the caliper and the mic. when used on the block is the "right size" and any difference between that and the measurement on the job is the "difference" or "error" or the "amount to go yet".

The caliper is much better thermally insulated than the mic. but I have to be very careful with thermal insulation when handling the slip guage.

As for "flat" and "right angles" this is what I use:
http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Machinist_Square1.jpg

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

It is flat and straight to within the "bubble" accuracy of 0.02mm/m = 1:50,000. While it is definitely NOT a surface plate it does very well when is used as in the "edges and stripes" on the (UK) "Union Jack" flag. Along each side and across the part at the centre "both ways" and diagonally. If required, I reverse the guage. If it "comes up OK", I am pretty right and can proceed with confidence.

I work to pretty well normal shop tolerances - and that's all. If I am only allowed a thou or a (cuppla??) tenth/s, I use it. If its a clearance hole for a bolt, it will get anywhere from plus 1/32" to plus 1/16" depending upon application.

Some may work to higher standards for size and finish and there is nothing wrong with that.

If I require real accuracy, that's when the shop roller door is closed (wind drafts and "wind chill") and the temperature as stable as I can get it on the day. But its rarely that degree so care is needed - but its there if needs be.

If I really am fussy about size and/or finish, then that is when the surface or universal (surface and cylindrical)grinders get a "run" - slip guages and all. But they hardly ever get used.

But back to the "glass surface plate": it might surprise some as to just how well it comes up with the "Union Jack" test - quite well in fact- but while it may get used as a "sort of" surface-plate it quite definitely is NOT a surface plate. But it is - in my shop anyway - quite adequate for testing and marking-out. It may or may not suit others. It is very portable, can be stowed almost anywhere, and being laminated and toughened, it won't shatter easily (I've dropped the smaller one onto the concrete floor and there was no damage). It is from a door/window sheet and has had to meet all the Standards hazards for that as well so there will be pellets and not shards if it breaks.

Its all a matter or having and using what you need when you need it that counts.

Every shop and every person will have different needs and tools and ways of using them.

It really doesn't matter what methods are used provided that the user is satisfied that his standards and that of the job are met.

Its all about "limits".

I was talking about "tolerances and limits" and "surface finish" earlier on.

Here is a pic (metric) that gives a co-relationship between machining process, tolerances and surface finish.
http://i200.photobucket.com/albums/aa294/oldtiffie/Limits_and%20_Fits/Limits_and_Fits14.jpg

Limits and fits (metric):
http://i200.photobucket.com/albums/aa294/oldtiffie/Limits_and%20_Fits/Limits_and_Fits11.jpg

All of this is pretty well dead in line with what Forrest had been "banging the drum" about all along.

Continued at the next post ......................

oldtiffie
07-20-2009, 07:11 PM
This post is a continuation of the previous post.

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

Here is an article I did here some time ago on that subject. It may help some see the relationship between "fit", tolerances, limits and sizes:

http://i200.photobucket.com/albums/aa294/oldtiffie/Tolerancing/Caliper_accuracy1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Tolerancing/Caliper_accuracy2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Tolerancing/Caliper_accuracy3.jpg

J Tiers
07-20-2009, 11:39 PM
Jerry, for 0.0001" in 5" I'll still bet you a beer and a pizza measuring a standard or a part size. When you handlle standards at that accuracy, working bare handed is like navigating with your enameled steel coffee mug on the compass binnacle. ..................


I see you got the point....... as I figured you would.....

Which was that most folks would be chasing the setting all over the place, since 5" is enough to start showing thermal effects in fairly short order.

While they have their "hot little hands" on the mic re-setting it to read correctly, it is expanding vs the standard. Then the setting shows up wrong, and they may figure they went too far, and re-set, etc, etc, etc. Since the standard won't be held in the hands as much, it will stay closer to ambient temp, and the fun begins.

The fact that some mics require significant torque to set, and so they need a hand wrapped over them holding the frame steady enough to set doesn't help much!

Tiffie, I quite agree that the glass is fine for marking out and the like. It seemed that you were claiming better flatness than that, but I must have misunderstood your intent.

I need better flatness because I use the flat for a scraping reference, both for direct use, and to scrape auxiliary reference tools to. So the better flatness is actually useful, and I can indeed notice if it is "off". Float glass would be worse than useless to me.

tony ennis
07-20-2009, 11:47 PM
The idea that a piece of plate glass wouldn't be flat enough for me is laughable. It is far beyond my skills to lay out something so precisely that I'd need something more.

.001 is still tiny in my world.

Forrest Addy
07-21-2009, 12:25 AM
Glass is OK, Tony, but unless its quite thick it will be flexible under the weight for lay out tools and the work. You need a good flat support under it like a machine table.

My objection is a pathologival anxiety associated with broken glass. Most people have no problem with. Ok I carry some baggage. But if you ever work with a nice thick matte finished granite surface plate of known accuracy you might not fully appreciate the accuracy but the smooth glide of surface gages and other layout tools outperforms the feel of glass. And a small granite surface plate is not that expensive - $35 for a 12 x 18; $42 for a 18 x 24. But work with what suits you. Personally glass out where it can break gives me the creeps.

oldtiffie
07-21-2009, 01:02 AM
Well Forrest and JT.

surface plates are cheap in the US.

Here is the best I've seen in OZ (multiply X 0.80 for AU$>US$):
http://mcjing.com.au/searchresult.aspx?keyword=surface%20plate

I realise that the description leaves a lot to be desired, but I suspect that the "Webmaster" has a bit of catching up to do with English. The owners (husband and wife with kids working there too) are both Chinese and both have Chinese University degrees majoring in Metrology. Nice people too.

I first ran into them at an Engineering Trade Show here. All of their stuff on sale was from good>very good>excellent. They know their stuff. I asked them about the grade, standards (US, UK and ISO) for machinists squared and slip guages etc. and they reeled it right off - or went and got the details - and discussed it. That just about blew me away as I've never seen that level of competence anywhere else here.

I've bought a fair bit of stuff from them.

Anyway, back to the surface plates.

Thanks for the discussion as I have re-thought a lot of stuff and points made and have decided that I really don't have a good reason not to have a good surface plate and a good few as to why I should.

I will opt for the larger granite plate 630mm x 400mm x 100mm x 33 Kg (25" x 16" x 4" x 70pounds - nominal) - cost AU$260 x 0.80 ~ UD$210 + postage and handling.

Its big enough to cover all that I am likely to use it for and reasonable as regards size/storage and lifting etc.

There are things that I really will need a good flat surface for as well as for any scraping or "blue-ing" or any "good" measurements with feeler guages etc. that I need to do.

I anticipate using the glass plates on the mill or other machine tables for 95%+ but the mill may be otherwise "tied up".

I am anticipating that the surface plate will be "Grade B"/"Tool Room".

I need to confirm that as well as whether the dimensions of the flat surface are across the centres or across the diagonals.

I will keep you informed.

This has been a good thread - thanks for starting it Forrest and thanks JT - sincerely - for contributing to it in the spirit in which you have - same to Forrest.

While this may have been my journey on the "Road to Damascus" it will surprise no-body when I say that St. Paul I ain't!!

DICKEYBIRD
07-21-2009, 08:14 AM
I will opt for the larger granite plate 630mm x 400mm x 100mm x 33 Kg (25" x 16" x 4" x 70pounds - nominal)Mick, I bought a 12"x18"x3" on sale from Enco a couple weeks ago and it weighs over 70 lbs. I reckon your 25"x16"x4" wil go at least 170 lbs, eh?

I still can't figure how they can sell the 12"x18" for less than $25.00, INCLUDING SHIPPING! Must be a huge overstock.

J Tiers
07-21-2009, 08:58 AM
I agree, be prepared for that slab to weigh twice what they say. That size is, volume-wise, almost 1 cubic foot and most stone is about 140 to 165 lb per cubic foot.

My W.A.G. on weight is about 155 lb. A sanity check with Grizzly Industrial online shows an 18" x 24" x 4" flat to weigh 154 lb.

Your target flat is somewhat larger in length, but narrower and slightly thinner than 4", so will likely be in that same range or a bit more.

oldtiffie
07-21-2009, 10:51 AM
Dickiebird and JT.

The mass of granite is about 171 pounds/cubic foot and I reckon that plate will be about 160 pound.

My guess is that they've cocked up the mass/weight big time.

Thanks for the heads up.

I guess that it live on my 500Kg (1/2 ton/tonne) lifting table until I sort it out.

If I lift it with my ar$e hanging out and slip, step back and step on it, I will cause myself quite some inconvenience!!!

Here is the text of my email to them. I will post the reply for information:


.
.
.
I am interested in purchasing a granite surface plate.

Your item GSP-630.

The "Accuracy "0" 6 micro" is a bit confusing.

What is the grade of the plate ("B"/"Tool-Room" - or better).

Is the plate certified - and if so by whom and to what grade?

Is the "6 micro" 6 micro-inches or 6 micro-metres?

Are the measurements of the top of the plate the lengths of the sides or the diagonals?

What is the postage/couriers + handling cost to Bittern Victoria, Post-Code 3918?

What is the total cost?


Your early advice will be appreciated.


Regards

Forrest Addy
07-21-2009, 04:48 PM
Granite covers a wide range of ingious rocks that vary in densuty according to the mix of the rock from which its comprized. There/ That's all I know about granite. If you need to know more hunt down a minerologist or a geologist.

My Standbridge straight edge is made from "tertiary black grabbo" whatever the hell that is. Thing is granite is a naturally occuring material like wood or pizza. I'm told that the rock will vary considerable from on side of the fuarry to the other. We take what we get meaning it weighs what it weighs. So don't calculate too precisely.

DICKEYBIRD
07-21-2009, 05:27 PM
Don't forget, the density of Chinese granite may be measured using the same metrology as their electric motors.;)

oldtiffie
07-21-2009, 06:33 PM
Granite covers a wide range of ingious rocks that vary in densuty according to the mix of the rock from which its comprized. There/ That's all I know about granite. If you need to know more hunt down a minerologist or a geologist.

My Standbridge straight edge is made from "tertiary black grabbo" whatever the hell that is. Thing is granite is a naturally occuring material like wood or pizza. I'm told that the rock will vary considerable from on side of the fuarry to the other. We take what we get meaning it weighs what it weighs. So don't calculate too precisely.

Thanks Forrest.

As you say, there are lots of granites that have an average density of 2.75 gram/cc^3 which translates/converts to about 172 pounds per cubic foot.

http://en.wikipedia.org/wiki/Granite

The plate in question is about 0.63 x 0.40 x 0.10 = 0.025 cubic meters which is about 0.88 cubic feet x 172 ~ 151 pounds ~ 77.5 Kg which is a lot heavier than the 33Kg on offer = about 2.5 times heavier.

There are 2.2 pounds per kilogram. I suspect a mass or weight conversion mistake.

This is turning into a Shakespearean "Comedy of Errors"
http://en.wikipedia.org/wiki/Comedy_of_errors
http://en.wikipedia.org/wiki/The_Comedy_of_Errors

And is somewhere between a French farce and a Greek tragedy.
http://en.wikipedia.org/wiki/La_Cage_aux_Folles_(play)
http://en.wikipedia.org/wiki/Greek_tragedy#Greek_tragedy


I think I will have to re-think this and go for a smaller plate.

I will await the reply from the vendor and will let you know.

oldtiffie
07-21-2009, 06:41 PM
Don't forget, the density of Chinese granite may be measured using the same metrology as their electric motors.;)

Thanks DB.

There is a tenuous relationship here.

Chinese motors are noted for "smoking" and there is a bit of "smoke and mirrors" here.
http://en.wikipedia.org/wiki/Smoke_and_mirrors

And I may be caught between a rock and a hard place.
http://en.wikipedia.org/wiki/Between_a_rock_and_a_hard_place

lazlo
07-21-2009, 08:13 PM
Don't forget, the density of Chinese granite may be measured using the same metrology as their electric motors.;)

TexasTurnado bought one of the bigger surface plates from Enco, and when it arrived, it had clearly cracked during manufacturing, and they epoxied it back together and finished the lapping. :rolleyes:

I've actually bought two of the $25 with free shipping Enco specials, and they both measure well within the Grade B tolerance. When you're buying surface plates this cheap, you're not paying for quality control...

oldtiffie
07-22-2009, 07:18 AM
The saga continues - bit all is sweetness and light.

I rang my Chinese supplier twice today, the first to query the web and the second to place the order.

He was quite surprised at the errors on the web page. We discussed sizes, weights and accuracy. That info had been there for quite while apparently and no one had queried it before.

The larger plate I initially required which was previously 33 Kg is now 90Kg (boxed - ready to go).

I settled on the next down 400 x 300 x 70mm (~ 16" x 12" x 3" nominal) which is only 46 Kg ~ 100# (boxed) and just a bit over half the price.

The accuracy is 0.001mm/m ie 1:1,000,000 or 1 in 1 million which is good enough for me.

I was amazed to see that the cast iron plates are about double the cost of granite.

I wonder what that is due to.

Any ideas or comments Forrest or JT or anyone else?

The web page is here:
http://mcjing.com.au/searchresult.aspx?keyword=surface%20plate

I will advise further when it arrives and I have it uncrated.

oldtiffie
07-22-2009, 07:45 AM
As a matter of interest, I converted the mm sizes to inches and applied Forrest's "Grade B" formula and both it and the Chinese "1:1,000,000" standard are the same within very small limits that can be disregarded.

I suspected that when I noticed the "40" and the "25" in Forrest's formulas as:
- there are approximately 25 mm in an inch (25.4 to be precise); and
- a mm is approximately 0.04" (0.03937 to be precise).

My guess is that the US standard is now metricated and the "inch" standards are (re?) derived from the ISO metric standard. It was just too neat to be otherwise.

I will re-do the calculations to "tidy it up" and will post it tomorrow.

oldtiffie
07-23-2009, 09:35 AM
My new granite 400 x 300 x 75mm surface plate arrived today at 11.00am. It was ordered at 3.00pm yesterday. Distance was over 600 miles!! They musr have flown - both by air and courier!!

Very impressed all round. Flatness was 1.2 micro-metre (~ 480 micro-inches ~ 0.0000480 ~ "half a tenth" with diagonals 29.4"

It is graded as "00" which has an error range of 3.0. "0" has 6.0 and "1" has an error range of 12 micro-metres.

It has the Mouton points marked on the back.

I will post details tomorrow.

Very impressed.

For Forrest Addy:
Forrest, I used your Grade "B" formula (4 x Grade "AA") and it allows 0.000298 (say 3 "tenths") - so it's not bad at all. I will post details tomorrow - with the test

oldtiffie
07-24-2009, 07:25 AM
OK.

I think I have it all together now.

Cost was AU$155 ~ US$124 plus AU$36 ~ US$30 freight - which I thought was pretty good. Not a problem anywhere.

Here is a pic of it as it arrived in its box:
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplatebox1.jpg

This is what it looks like from the top:
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplatetopface1.jpg

These are the support positions (under-neath):
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplateset-uppoints1.jpg

This is the test sheet. Note that the calibration is in units of 0.001mm/metre = 1 : 1,000,000 = 1 micro-metre ~ 40 micro-inches.

http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplategradediag1.jpg

The highest reading (from the reference zero) was 1.22um ~ 49micro-inches ~ 0.000049 ~ "half a tenth" (of an inch).

The plate is classifies as Grade 00

Grade 00 is allowed 3.0um ~ 120 micro-inches ~ 0.000120 (1.2 "tenths").

Grade 0 is allowed 6.0 um (2.4 tenths).

Grade 1 is allowed 12.0 um (4.8 tenths ~ almost "half a thou").

This is my calculation/check sheet using the formulas that Forrest Addy provided that the US uses. I worked it out for "Grade B" - for both the formula and the Chinese 1:1,000,000 system for my plate and they are as near as identical that any difference doesn't matter:
http://i200.photobucket.com/albums/aa294/oldtiffie/Surface%20plate/Surfplatetablecalcs1.jpg

Forrest Addy
07-24-2009, 09:33 AM
Don't just stuff that cert in a file and forget about it. It's a resource you can exploit. Keep a copy handy near the surface plate for ready reference.

Lemme give you a few ideas. That's a very good plate. Consider the error map. You notice the left edge is almost perfectly flat (0.00, 0.25, 0.49). The upper left to lower right diagonal readings are 0.00, 0.12, 0.00 (error 0.12 high in the middle) and the other 0.49, 0.012, 0.49. (error 0.37 low in the middle). There is a hump of 1.22 at 12 oclock, and the lower edge is 0.49, 0.35, 0.00 (error 0.10 high in the middle but assymetrically distributed). As you look over the error map you will note other error groupings that can be exploited for enhanced accuracy work while the calibration is fresh. Quantified error data to three significant figures is so far superior to pass/fail I don't know why binary acceptance criteria exists. If I know the error precisely I can use it to compensate collected data - not every time you understand. You can't simply take things for granted. You have to keep your wits handy or you will wind up asserting data that canot be verified.

Note there is a high place best used as a preferred work area where apparatus like surface gages, hight gages etc can be slid over the reference. The wear thus generated has the effect of improving the plate's accuracy by reducing the height of the hump. Also note there are "lanes" of greater than grade accuracy should you wish to do anything fancy like scrape in the base of a master level for example.

The "lanes" give you a check on just how sensitive your scraping and bluing is. Vertcal left lane is almost flat. The lane from upper left to lower right is convex by 0.12 in 29" or so. Scrape the stipulated level to left edge until you can't improve the bearing. Then blue up the top left - bottom right diagonal and check the level for flatness there. If the bearing peters out in the center you know your technique detects inconsistency in the flatness of the plate to 0.13 micrometers (5millionths). If your technique will pick up that you're 'way craftier than me. If you get the same readings on separate days, you are paying close attention to handling and environmental factors .

The lower left quadrant is about the flattest place on the plate. It's almost a perfect plane if sloped. You can use this to set a "repeatometer" should you wish to conduct a localized wear survey between calibrations. Equipped with a LVDT and a gage amp a repeatometer will detect localized departures from initial settings to sub-millionths.

That cert is a powerful tool if you need better than mere net grade accuracy for any reason. Except for the 1.22 hump the plate predominates 'way better than "AA" so it's there if you happen to need it some day. Don't expect this kind of accuracy in general use. The plate is thin for Grade AA ervice and heavier weights will deflect it. FWIW, You can roughly estimate the weight capacity of the plate given variable loading: the differential levels used to check it weigh about 3 lb apiece so take it from there. The readings shown on the cert are dirived. A level will give you slope data not heights. The raw level readings have to be crunched to yield height data for the error map on the cert and a nice 3D representation of the error.

Finally, remember you have the cert. You can use it to re-calibrate the plate using home-brew "planikator" technique. Lastly, I would orient the plate on its stand so the hump is to your working front where it will take most of the sliding wear.

All the above is 'way, 'way fancier than anyone in his right mind will ever need. I just used Tiffie's cert and his application to illustrate just how handy a calibration cert can be and how it can be used in the shop to enhance reliability of metrology technique.

oldtiffie
07-24-2009, 08:51 PM
Forrest.

Thank you sincerely for that excellent response and exercise in concise, focused and detailed professional technical reporting.

I really do appreciate the time, expertise and effort that went into it.

This is a courtesy note to say that I've read it - and re-read it - in depth - as it deserves to be.

I will reply further later - in the next day or so - in a manner which your post warrants and deserves.

oldtiffie
07-26-2009, 09:32 PM
OK Forrest.

Back again.

I had picked up most but all all of the points you mentioned.

That Test Report has been copied for shop and office work and the original filed safely away - as you suggest.

Only problem with using the plate to the levels of accuracy that you quite rightly advise is that there is no indication of what are the "front" and "back" longer edges of the plate - so its a 50-50 punt there. If the standards don't require the reference corners and edge to be identified for future reference, then they don't have to do it - but it would have been nice(r) if they had.

My guess is that as the plate is "00" that's its almost too good to be used in the shop!! A USA-class "B" would have suited me better!!!

But all that aside, I will use it for comparator/reference work only as I will continue to use the glass plate and machine top faces for "rough(er)??" setting out and marking-up with the granite plate, finer indicators and my "Grade 2" slip guages used to reference and comparator work.

Thanks to you and JT and your focused advice, I bought it at all - but I am now much the wiser and better off - even if a bit poorer - for it all.

As a matter of interest and with reference to a recent discussion on practical levels of accuracy in the average HSM shop, I decided to up-grade my internal measuring capacity. I wanted 25>150mm (1">6") micrometer capacity, but all that was available was 25>50mm (1">2") and 50mm>300mm (2">12") so I bought them. And very nice they are too. I went to get my 0>150mm (0">6") micrometers out and noticed that they were still very cold from the over-night temperatures here, but the the internal micrometers were relatively warm (no definitive temperatures taken). The first to be checked was the 0>25mm inside micrometer with my 25>50mm (1">2") outside micrometer. The outside micrometer was quite accurate using its 25mm setting guage but had quite an error (~ 0.015mm ~ 0.0006") when applied to the inside micrometer. So I let them settle down to the ambient shop temperature and the result was "spot on". That was just a "spot" check. Ordinarily, if I was after as good an accuracy as I could get, I'd use my slip guages.

I will post pics of the internal micrometers later.

But it is all relevant to the accuracy of the surface plates - both granite and cast iron - at different temperatures as individual items and when used with other tools with say both the same at ambient temperature and then with a temperature difference that may be significant. I would expect just about every tool I'd use on a surface plate (or a glass plate) to "settle down" to shop temperature quicker than the large mass surface plates.

I notice that in this table (metric) of co-efficients of linear expansion:
http://www.engineeringtoolbox.com/linear-expansion-coefficients-d_95.html
that:
cast iron is: 10.8
granite is: 7.9
steel (average) is: 14
glass plate is: 9
brass is: 18.7
aluminium is: 22

So there is quite considerable sources and causes of error due to "temperature".

This will also apply to stuff that has been machined and not cooled - through-out - "core" included) as it may be the cause of quite some WTF remarks!!

So that plate and your advice has led to a lot of re-thinking and deciding what the options (expanded!!) now are and what is appropriate in the circumstances for each job.