PDA

View Full Version : wiggler accuacy problem



loose nut
10-25-2009, 06:18 PM
When trying to test the accuracy of my mill DRO's I placed a 4" gauge block in the machine vise so that the "good" surfaces where facing the sides of the vise, picked up one end of the gauge with the wiggler zeroed the X scale and repeated several times to check the reading, always getting 0.000" (DRO readout set to 3 decimal places for this test). Next I moved to the other end and picked it up, should read 4.100" (gauge block plus diam. of wiggler end) but consistently got a 0.004" to 0.006" (usually 0.004") undersized reading while repeating this several times.

I checked the size of the block, a cheap set admittedly but still on size, as accurate as my mic could tell and the end of the wiggler (Starret) which was within .0002" of the 0.100 diam it is suppose to be. Size errors here would not account for the degree of error showing up.

Thinking that there would be something wrong with the DRO I repeated the experiment with the .5, 1 and 2" blocks and every time I got the error same, .004 to .006" undersize. If the scales where off the error should be linear, say .001" /inch of travel not always the same.

I thought that there might be some flex in the mounting of the read head but every time I moved to the other end and picked up the original starting point it is always 0.000", if there was any flex in the mounting it should show up there.

If it is my ability to use the wiggler that is at fault why am I able to get consistent readings in any one place.

Any thoughts, I'm stumped. Thanks in advance.

tattoomike68
10-25-2009, 06:28 PM
Years ago I had a big ol fight with my brother. he said I bored some holes 12.002" apart when they wanted 12.000"

I told him to stick them 16" x 2" x 4" blocks of steel up his butt.

Once again it went on a farm machine and with the old beater mill it was fine.

dont ask too much from an old machine or a small china machine. +.002 is damn good for a old beater. so I told him to kiss my ass. he later admited he was being an ass and let it go.

;)

GadgetBuilder
10-25-2009, 07:08 PM
Are you using a wiggler?
Or an edge finder like this one:
http://www.starrett.com/pages/1512_t_and_r_edge_finder.cfm

There are various techniques for each, of course.

I generally get better accuracy and repeatability with an edge finder. I happened to check repeatability of an edge finder vs my DRO a couple weeks ago and it was mostly 4 tenths but occasionally 7 tenths. Surface roughness, oil on the surface and eyesight plus judgment affect the result. My repeatability with a wiggler was 1.5 thou (and it takes more concentration).

One thing that might help determine where the discrepancy originates would be to set up a long travel indicator on table position. This should prove the DRO is accurate over modest distances and allow you to then check your technique on using your finder.

If you get really into this you might want a Moore edge finder:
http://mooretool.thomasnet.com/item/all-categories/edge-finder/3070-a?&plpver=10&origin=keyword&by=prod&filter=0

But check the price, it might be a consideration... :D

John

Carld
10-25-2009, 07:18 PM
Are you using a wiggler or an edge finder? I would not trust a wiggler to find an edge with much precision. An edge finder would be best.

If the end of the "wiggler you used was .100" and you "0" the starting end and go to the other end it should show .050" longer than the 4,000" block not .100". That is because your measurement is taken from the center of the spindle to the edge of the "wiggler". If you did not "0" the DRO at the start end and went to the other end then you would have 4.100".

When you "0" an edge finder or what ever your using to find the edge you use half the diameter of the edge finder, wiggler or dowel pin to move the table and "0" the DRO.

As to why your getting a .004"-.006" error the only thing I can guess is it's the "wiggler" and you should try using an edge finder.

loose nut
10-25-2009, 07:22 PM
It is a "old fashion" wiggler not an edge finder as in your link and I can usually get a repeatability of less then a thou. with it. As for the equipment it is a fairly new mill and brand new Fagor DRO's.

I know for most things I don't need super accuracy but having an error like this is a bit unnerving, it shouldn't be there and I can't find the source.

loose nut
10-25-2009, 07:28 PM
If the end of the "wiggler you used was .100" and you "0" the starting end and go to the other end it should show .050" longer than the 4,000" block not .100". That is because your measurement is taken from the center of the spindle to the edge of the "wiggler". If you did not "0" the DRO at the start end and went to the other end then you would have 4.100".
.

I zeroed the X scale but did not move it over half the diam. (0.050") on either end hence the 4.100" readings.

loose nut
10-25-2009, 07:33 PM
One thing that might help determine where the discrepancy originates would be to set up a long travel indicator on table position. This should prove the DRO is accurate over modest distances and allow you to then check your technique on using your finder.
John

I have an edge finder, that works with an offset indicator, similar to the Mooretool one in your link, I'll try it next. As for using an indicator as a check, if you get an error which one is wrong.

Carld
10-25-2009, 07:35 PM
Ok, but it is strange that your getting an error when you say you can get within .001" with the wiggler. I just don't trust a wiggler except for locating crosshairs to drill a hole.

Call Fagor or the people you bought it from and ask them about what you are getting.

lane
10-25-2009, 07:39 PM
Make sure the JO block is dead parallel . Pick up the end 0 move to other end in one direction DO not back up . Hold another JO block tight to the face on the 4 inch block and continuer move until edge finder clicks off. That is your reading Should read 4.100.
The way I do it is with a set of micrometer standers and 2 v blocks clamped to the Table . And you stop one end of the slandered against a 123 block and you can keep changing out to longer standers and check the whole table travel for error.

Glenn Wegman
10-25-2009, 07:39 PM
Carl,

I believe he states he gets repeatability within .001". Not accuracy.

It could be 1/4" from the true edge and still repeat within .001" every time.

loose nut
10-25-2009, 07:46 PM
Carld, I agree with you on the error, it doesn't make any sense. If the Dro was a fault then after moving the table back and forth several times, as many as 20 times, if it was at fault, bad scale or loose mountings, when I go back to the original starting point it should show a discrepancy but it is always 0.000, never varies and as I stated earlier the error is always the same regardless of the distance travelled, that doesn't sound like a scale error.

I would expect some small error but not 5 or 6 thou. Must be me or the wiggler, I'll try the edge finder next.

loose nut
10-25-2009, 07:54 PM
Make sure the JO block is dead parallel . Pick up the end 0 move to other end in one direction DO not back up . Hold another JO block tight to the face on the 4 inch block and continuer move until edge finder clicks off. That is your reading Should read 4.100.
The way I do it is with a set of micrometer standers and 2 v blocks clamped to the Table . And you stop one end of the slandered against a 123 block and you can keep changing out to longer standers and check the whole table travel for error.

The vise was indicated parallel to the ways. I'll give it a try but I thought the whole point of using DRO's was so that you didn't have to worry about backlash and such things, the scales are independent of the leadscrew, why would moving it backwards be a problem.

GadgetBuilder
10-25-2009, 08:08 PM
I have an edge finder, that works with an offset indicator, similar to the Mooretool one in your link, I'll try it next. As for using an indicator as a check, if you get an error which one is wrong.

You seem to have gauge blocks so proving the long travel indicator correct should be possible for, say, the 0.5" range without too much effort.

If you have something similar to the Moore edge finder try reading both sides of the finder, then use the mid point. This should cancel finder or technique error.

John

Glenn Wegman
10-25-2009, 08:12 PM
Just use a DTI and Lane's method without the math.

Lay the long gage block on the table with another gage block against one end that sticks up just enough to create about a .100" step and indicate it parallel with the X travel. Touch one end with the DTI and note the reading. Zero the DRO. Move to the step and touch it until the same reading is obtained and the reading on the DRO will be the distance traveled.

I don't understand the reluctance on this site to using simple tools like DTI's :confused:

Tasks such as indicating vises, centering over bores, checking squareness, finding edges, etc. really should not be such daunting tasks as all can be accomplished easily with a simple DTI!

loose nut
10-25-2009, 08:20 PM
You seem to have gauge blocks so proving the long travel indicator correct should be possible for, say, the 0.5" range without too much effort.
John

That sound is me hitting my head against the table, I should have picked up on that before. So the whole day wasn't wasted I did learn something new.

beanbag
10-25-2009, 08:22 PM
Some other possible sources of error:
The DRO can have some backlash in it.
Looseness in the y gibs can cause errors in x, and the DRO will not pick that up.
If possible, mount an indicator to the spindle housing and use that to measure distances to your workpiece when testing out your DRO's and wriggler.

mechanicalmagic
10-25-2009, 08:30 PM
Clamp another flat to the 4.000" block. Approach both edges from the same direction. You should get 4.000". If not, I'm stumped.

Since you seem to have hysteresis, it could be either mechanical or electrical.

I have a cheap chinese quill DRO that has .003" of electronic hysteresis. Fagors should not have the problem.

I also have .002" mechanical hysteresis in the Y of my BP knockoff. I believe my dovetails are bell mouthed (abused as a child). Mine repeats, and is not a dragging read head. My X is right on.

I tested "0" on my mill by approaching a zero edge, using a .0005"/div indicator. Then overshoot and approach from the opposite side, trusting the indicator.

loose nut
10-25-2009, 08:44 PM
Some other possible sources of error:
Looseness in the y gibs can cause errors in x, and the DRO will not pick that up.


Good point.

loose nut
10-25-2009, 09:22 PM
EXTRA EXTRA READ ALL ABOUT IT

OK,I just ran out to the shop, tightened up the Y gibs and tried it again with the edgefinder on the 4" block.

3.9996

and the edge finders magnet didn't stick to well to the thin block so it may be able to tweak it a bit closer.

Still don't see why the wiggler was so far out or it was just me.

Thanks for the ideas.

beanbag
10-25-2009, 09:44 PM
Clamp another flat to the 4.000" block. Approach both edges from the same direction.

This may be the key point. You need to approach the edges from the same side for the measurement to be totally relevant.

I've only used an edge finder and not a wriggler. But let's say that hypothetically, your wriggler "goes off" at .048" from the actual edge. If you approach another edge 4" away, but facing the other direction, you will measure a distance of 3.996" between the edges.

Carld
10-25-2009, 10:41 PM
loose nut, I have done this a long time and I can tell you that personally I have never had good results using a wiggler to find an edge within .001", It just has not worked for me and I think you just proved it won't work for you.

I know, I know, people will say they get a good reading using a wiggler but it just don't work for me. I always use a edge finder or a dowel pin and cigarette paper but the edge finder works within .001" every time for me.

uncle pete
10-25-2009, 11:29 PM
GagetBuilder,
I'm amazed that no one commented on the price of the Moore edge finder, Holy crap $5100, I just bought Moore's 2 books and have had an interesting time reading them. Thought their equipment would be very pricey but????

Pete

Paul Alciatore
10-26-2009, 02:05 AM
EXTRA EXTRA READ ALL ABOUT IT

OK,I just ran out to the shop, tightened up the Y gibs and tried it again with the edgefinder on the 4" block.

3.9996

and the edge finders magnet didn't stick to well to the thin block so it may be able to tweak it a bit closer.

Still don't see why the wiggler was so far out or it was just me.

Thanks for the ideas.


I think you missed the key element in what you said above. "...tightened up the Y gibs". I found early on when working with import milling tables that you can easily get 3 to 5 thousanths variation if you don't pay a lot of attention to properly adjusting all gibs and LOCKING DOWN any that are not actually in motion. What happens is the table can rotate a slight amount and, depending on where you are on the table, this can produce differences in almost any direction. Changing the direction of table motion will defenitely produce such a rotation. This type of rotation will also produce small errors in right angles you produce with the X and Y feeds.

I seriously doubt that your DRO is off by more than a few tenths and the error would not be constant for different distances, it would more likely be proportional to the distance. Lane's method is excellent as the motion is in the same direction when checking both ends of the block.

Adjust the gibs as tight as you can stand having them and LOCK DOWN any that are not actually in motion during the cut or measurement. I have actually been tempted to design a motorized lock for the gib screws. "Click", lock X: "click" unlock Y.

GadgetBuilder
10-26-2009, 10:41 AM
Hi Pete,

I wondered if the $5100 price was a misprint. Plans for the Moore edge finder are on the net, apparently they released them because even with the plans one can't easily make one - tolerance is +/- 25 millionths on some dimensions....

I don't really understand how the unit is used, would be nice if Moore made the user manual available too.

A friend has an LVDT probe with rated repeatability of 4 millionths so we're thinking about trying to make a few just to see the difficulties encountered. Best guess is they must be made in small batches to help maintain alignment while clamped together in the jig.

"Foundations of Mechanical Accuracy" makes it clear that Moore's products are made to a higher standard than most.

John

Glenn Wegman
10-26-2009, 12:52 PM
Here it is John!

It's not a misprint!

http://img.photobucket.com/albums/v647/Fighter1/edgefinder005.jpg

You simply hold it against the edge of the part as depicted in the lower right of the drawing, and with a DTI sweep the two inner surfaces that are parallel with the edge being "found" until the DTI reads the same sweeping both surfaces. When that is done the spindle is centered over the two flats you are dead center over the edge of the part.

A very good finish on the edge it is held against is the key to accuracy. It's primarily a Jig Grinding tool where wxtreme accuracy is paramount.

loose nut
10-26-2009, 07:45 PM
It appears that the potential for error creeping in from many sources is a big problem that I'm going to have to work on.

I tried some more tests tonight and there where some errors but they where much smaller at about 0.001 to 0.002", Part of the problem seems to be that the magnet on the edge finder that holds it in place doesn't want to hold against the block as good as I would like so it tends to move slightly. Of course I didn't pay $5100.00 for it either, more like $60.00.

I will accept that the DRO's are probably accurate enough for my needs.

GadgetBuilder
10-26-2009, 10:06 PM
Hi Glenn,

Thanks for the info on how it's used, my notion was to use it with an edge finder but a DTI makes more sense.

I like Moore Tool's style, not like Coke where they hide the formula.

John

Glenn Wegman
10-26-2009, 10:54 PM
Thay aren't as painful on ebay :)

http://cgi.ebay.com/MOORE-TOOLS-EDGE-FINDER-EXCELLENT-CONDITION-TOOLMAKER_W0QQitemZ360201579245QQcmdZViewItemQQptZ LH_DefaultDomain_0?hash=item53ddafeaed

Rustybolt
10-27-2009, 08:27 AM
loose nut, I have done this a long time and I can tell you that personally I have never had good results using a wiggler to find an edge within .001", It just has not worked for me and I think you just proved it won't work for you.

I know, I know, people will say they get a good reading using a wiggler but it just don't work for me. I always use a edge finder or a dowel pin and cigarette paper but the edge finder works within .001" every time for me.

My exact same experience. Thanks

Carld
10-27-2009, 08:44 AM
Personally I don't see where using a Moore edge finder would making finding the edge any better. What your doing is adding another surface to the surface you want to measure and then measuring that false surface.

Why not be innovative and learn how to measure of the real edge rather than a false edge? I just don't see how you can more accurately find the real edge by locating a false edge. Think about how easy it is to add an error using the Moore edge finder.

It makes no sense to me at all.

Glenn Wegman
10-27-2009, 08:52 AM
One method I use to find an edge is to set a DTI to a ring gage and zero it. Then move to the edge and touch it until it reads zero. The true edge is then 1/2 the ring gage dimension. I use this method to check my B&B edge finder for accuracy.

If the surface finish is good, perpendicular to the spindle axis, and it's clean, the Moore is right on. It's the standard of edge finders.

Glenn Wegman
10-27-2009, 09:23 AM
The edge finder will actually wring to a good surface so there is no more error introduced than there is with wrung gage blocks.

The slot is .400" wide. Pic up one side...
http://img.photobucket.com/albums/v647/Fighter1/Moore-2.jpg

Then the other with equal DTI reading and you are excatly centered over the edge. Very quick and easy.
http://img.photobucket.com/albums/v647/Fighter1/Moore1.jpg


This is the method I was trying to describe to Loose Nut.

Square up a 123 block and secure it to the table and set a Gage Block against it. Touch the edge of the Gage Block with a DTI and infeed a thousandth or two and zero the needle and the DRO.
http://img.photobucket.com/albums/v647/Fighter1/Moore2-2.jpg

Pick it up a little and move to the edge of the block and feed in until the needle zero's. Read the DRO for exact travel. No error introduced through wigglers etc.
http://img.photobucket.com/albums/v647/Fighter1/Moore3.jpg

If your spindle is not perpendicular you may be off a little by raising the quill slightly, but most likely immeasurable on the DRO.

Carld
10-27-2009, 10:59 AM
Glenn, I guess there may be times when you just can't read the edge directly with an edge finder or dial indicator as you showed but I just can't envision it. In about 20 years of using a Starrett edge finder I have never had a situation where I couldn't use the edge finder either on the edge of the work or the edge of the vise jaw. I have never had an issue of not being .001" or less right on the edge of the work.

I just don't see any advantage to reading off a surface that is laying on the surface your trying to locate. There is no reason I can see to add the posibility of an error from the tool slipping or dirt. It just don't make sense to me but if it works for those that use it and they are willing to pay for the Moore tool that's fine with me but I just want to express a diverse opinion that I think has merit.

I do like your version of measuring the DRO travel with the indicator, that would eliminate a lot of error.

Glenn Wegman
10-27-2009, 11:27 AM
No argument here Carl. :) The Starrett is accurate enough for the average shop. I've tested mine and found it to be within a couple tenths at 1000 rpm. Much depends on the machine and operator as far as accuracy.

Just pointing out how the Moore is used.

Along with a suggestion for Loose Nut for a more accurate measurement.
.

loose nut
10-27-2009, 08:33 PM
Glenn, thanks for the pictures, it makes it a lot clearer and I will try it the next time I get to the shop.

Got to agree with Carl on the edge finder, I though it would be the cats a$$ but found that it can move slightly just by the pressure of the indicator finger, still a lot better then a wiggler. Mine has a magnet in the side to hold it onto the edge that is being indicated but it may have gotten weaker with age and isn't as strong now. The Moore edge finder isn't in the budget but maybe the Starret edge finder is.

loose nut
10-28-2009, 08:11 PM
Went out to the shop after work tonight and tried Glens method of checking out the DRO scales.

X scale was consistently -0.0002" over 4" :D

Y scale was consistently +0.0004 over 2" :D

I repeated each test about 10 times in each axis, can't say why the Y axis was off more then the X axis but still seems accurate enough for anything I would do and it's not to hard to add or subtract a few of those little things without even knowing it, actually I doubt I have the skill to use that level of accuracy, most of my work is on the large side of a thou at best.

It's more a case of "I paid for all them fiddly bits and I wan' sem' "

My next problem is getting better at picking up an edge, I thought I was pretty good at it and now it's a bit of a rude awakening to find out I'm not.

It would appear though that I am very good at getting the same amount of error over and over and over ???? It's never easy is it.:(

Carld
10-28-2009, 09:28 PM
Yep, if you try for .0002" it is always hard to get.

spope14
10-28-2009, 09:38 PM
DRO's are NOT the end all. I calibrate mine each year, they end up fine in the end, but a tool and die maker friend of mine showed me that they are not exactly perfect. We tested all of my DRO's with 2" "best class (I am not at the shop) gauge blocks and found .001 out over 6 inches. Then he showed me a set of dip switches in the DRO that allows for calibration. All of my DRO's were off, even checking with my other four sets of gauge blocks. (yes, I have tools to spare in some categories, including a set of 10 inch to 20 inch gauge blocks, for whatever I may need it for).

You may want to consult your DRO literature before saying DRO's are perfect. I just calibrated an old J&L DRO that was donated to the shop years back, replacing a beaten Anilam DRO with it. It was refurbished, but still did the job, Four checks with 6 inch blocks, one with two different 12 inch stacks from two sets, .003 off on 12 inch. used the DIP switches, now things are great. Confirmed it with my microheight gauge reaming holes 12 inches apart before and after with my students as well. 4 decimal place .0005 resolution on the DRO, same on the microheight electronic probe.

Glenn Wegman
10-28-2009, 09:41 PM
Loose Nut,

Looks like you just discovered what I meant when I stated earlier that there is a difference between "accuracy" and "repeatability" as far as your edge finder! Just because it repeats, does not always mean it is the true edge.

What is the scale resolution? Does the DRO even count in .0001" increments, or does it count in .0002" increments? Usually standard garden variety Lathe or Mill scales are not as fine a resolution as they are when used on grinders.

loose nut
10-29-2009, 07:26 PM
Glen, the resolution is .0002, the manual says it is accurate to .00012 (that might be under perfect conditions in a test lab, they didn't specify) but it is rounded off because the limited display range, pretty sure it a lot more accurate then I am, should be able to hold a thou if I can get pick up the edges better, I don't have a lot of need to read tenths.

spope14, I didn't say it was perfect, just good enough for me. My reason for spending money I couldn't really afford on DRO's wasn't for the increased accuracy but because "me eyes ain't wha' they used to be" and seeing the lines on the dials, even with the glasses, at the end of the day was getting a bit hard to do. Getting old sucks.

Tonight I tried the Y axis again with the 4" block just to see if the error was linear and increasing and it was between .0002 and .0004 over 4", better then last night. So either I was doing something wrong or my DTI isn't as accurate as I would like (the old Mitutoyo is getting along in years) or the gremlins where playing games again, it's still a lot better then reading dials.

I still have to get used to using them properly and learn how to identify all the potential points where errors can creep in. Learned a lot this week from you guys. Thanks

Glenn Wegman
10-29-2009, 09:08 PM
Glen, the resolution is .0002, the manual says it is accurate to .00012 (that might be under perfect conditions in a test lab, they didn't specify) but it is rounded off because the limited display range, pretty sure it a lot more accurate then I am, should be able to hold a thou if I can get pick up the edges better, I don't have a lot of need to read tenths.


That is why I asked. It will only read .0000, .0002, .0004, etc. so you may be well below .0002" off or .0004" when you check the travel error.

Carld
10-29-2009, 09:10 PM
The reason I finally bought a DRO for my shop at home is I had used them at work so long I fell in love with them. When I started doing paying work at home I bought a DRO and since my mill is very tight I feel reasonably sure I can hold +/-.0005" on locating points. After that it depends on any wobble, runout or cutter movement and there's not much I can do about that.

My display reads in .0002" increments and I get as close as I can on the numbers and I try the best I can, without getting ridiculous, to machine close.

I never try for perfection, I go for in tolerance work that way I am satisfied when I get done instead of always being pissed because I couldn't get it perfect. I guess it wouldn't hurt to check my DRO to see if the travel is in tolerance.

loose nut
10-29-2009, 10:07 PM
That is why I asked. It will only read .0000, .0002, .0004, etc. so you may be well below .0002" off or .0004" when you check the travel error.

So if it rounds up and it shows a .0004 error, the actual error could be from .00012 to .0004"

I think I will just switch the display to the 3 decimal place mode and forget about it.

Better still, where did I put that fractional ruler, 64ths should be close enough. Like I could see the lines on a ruler any more.

Glenn Wegman
10-29-2009, 10:19 PM
Actually, if it reads .0004", it would be between .0002" and .0004". I'm not sure exactly where the rounding off point is though.

A good wooden yard stick takes a lot of stress out of close tolerance work!

lane
10-30-2009, 07:45 PM
I have found most mill`s are not even ridge enough to use a DRO set to read .0002 it will flash to .0004 and .0000 I keep mine set to read .0005 . Close enough for any thing I ever built . If I need closer I make my move off a stack of JO blocks laying on the table. And like Carld about the only time I use a wiggler is with the needle point to pick up a line are punch mark. And just to let yall know learned something to night . That thing you showed as a Moore edge finder Never new it had a name but made one 20+ years ago and used to use it a lot when I ran a Sip Jig borer and a Knight Jig borer.
Something else I agree with is how accrate a dro is and checking it at different places along its length . Table warp will not show up and you wont be where you think you are.