Unable to calibrate axises

[mchasal]

Just to clarify, you're calling the carriage movement the X axis and the cross slide the Y, right? I believe the "standard" is a bit different with the carriage being Z (the spindle anchors the coordinate system) and the cross slide being X: https://www.hobby-machinist.com/threads/x-y-z-on-a-lathe.44499/
It seems a bit confusing and I'm not 100% sure , but as long as we're all on the same page; what you're seeing is that the cross slide travel on the DRO isn't matching reality (as defined by your 2 calipers), correct? Can you detail what your calibration process was and also what you're test procedure was for making the cut?

Are these the same brand/type of scales that you put on the mill?
How are you tweaking the calibration numbers when you try and correct it?
Did you try setting it to 5080 and seeing if the measurements match then?
Are you getting the same wrong measurement repeatedly or is it varying?

My initial concerns are to ensure you're using the same reference point for all the measurements and to ensure that you don't have some issue with mounting or interference with the scales.

 

[redvan22]

Just to clarify, you're calling the carriage movement the X axis and the cross slide the Y, right? I believe the "standard" is a bit different with the carriage being Z (the spindle anchors the coordinate system) and the cross slide being X: https://www.hobby-machinist.com/threads/x-y-z-on-a-lathe.44499/
It seems a bit confusing and I'm not 100% sure , but as long as we're all on the same page; what you're seeing is that the cross slide travel on the DRO isn't matching reality (as defined by your 2 calipers), correct? Can you detail what your calibration process was and also what you're test procedure was for making the cut?

Are these the same brand/type of scales that you put on the mill?
How are you tweaking the calibration numbers when you try and correct it?
Did you try setting it to 5080 and seeing if the measurements match then?
Are you getting the same wrong measurement repeatedly or is it varying?

My initial concerns are to ensure you're using the same reference point for all the measurements and to ensure that you don't have some issue with mounting or interference with the scales.

Mchasal,
I attached a document of every step I took to calibrate and test. (Didn't like the way the Word document was acting so i pasted the text below.)
As for your questions... two different brands of calipers, this time I didn't tweak the calibration numbers, didn't try 5080, several cuts give me off measurements each time.

Setup…
Axis changed for carriage from x to z and the cross slide changed from y to x.
CPI for both axis set to 10,000.

Z axis…
Hard stop setup on ways.
Two 123 blocks placed long end to long end against hard stop.
Incremental and absolute set to zero on z axis.
Removed one 123 block and moved carriage up to remaining 123 block.
DRO reading shows – 1.5256
Divided by 3 = 5085.
Set CPI for z axis to 5085
After reconnection, zeroed z axis.
Replaced 123 block and moved carriage up to it for a snug fit and display showed 3.0002

X axis…
Setup two 123 blocks 1” side to side against back of cross slide micrometer dial.
Moved cross slide up to blocks.
Zeroed x axis.
Removed one 123 block.
Moved cross slide up to remaining block and display showed 5084.
Set CPI for x axis to 5084.
Moved cross slide back until display read 1.0000 and inserted 123 block for a snug fit.

Test…
Chucked a piece of ½” steel stock in vise. 1.5” stick out.
Faced and zeroed z axis.
Took .005 cut ending at .7374 on DRO.
Measured cut distance with digital caliper which read .7390
Measured cut distance with non digital caliper which gave reading of .738

Thank you,
Michael.

UPDATE:

Went back to try new cuts and this is what happened,
First cut DRO showed .5425, DC was .5395 adjusted the CPI until I got as close as I could which was .5391 @ 5102 CPI
Second cut DRO showed .5640, DC was .5675 adjusted the CPI until I got close and that was .5674 @ 5097
Third cut DRO showed .6131, DC was .6185adjusted the CPI until I got close which was .6187 @5089

Please explain.

 

[mchasal]

Mchasal,
I attached a document of every step I took to calibrate and test.

It looks like you're seeing less than 2 thou of difference between all 3 measurements.

Your calibration procedure looks right as far as I can tell, but the "snug fit" of the carriage against the 123 block could vary enough to explain it. You could try it with a dial indicator to get a more repeatable position.

Or it might just be the tolerance/rigidity of the lathe. I don't know what machine you have or what condition it's in, but you could get some flex or slop that would explain this.

Or it could be measuring tools. Looking at some random digital calipers, they list their tolerance of +- .0012 so even in that one caliper there's enough tolerance to explain this difference and then there's the tolerance of the scales themselves that stack on top of that.

So I don't think you're doing anything wrong, but it just might be as good as your setup can do. Chasing these last few tenthou's down is probably above my paygrade, but if it were me, I'd turn off that 4th digit(except for calibration) and accept that level of tolerance. Maybe Yuri or some others that are in those paygrades above me can offer some other suggestions.

 

[Zertsman]

some thoughts,

in the absence of pictures of what you are doing... if the side of the carriage is cast and painted like mine, the 123 block will touch at a different dimension, because of paint thickness, each time as it's not a machined and predictable surface to locate from. a .002 difference may be spot on.

If you have a bolt head or other small feature that would be a more predictable contact point, use that and you may get better numbers.

 

[redvan22]

some thoughts,

in the absence of pictures of what you are doing... if the side of the carriage is cast and painted like mine, the 123 block will touch at a different dimension, because of paint thickness, each time as it's not a machined and predictable surface to locate from. a .002 difference may be spot on.

If you have a bolt head or other small feature that would be a more predictable contact point, use that and you may get better numbers.

I understand and I've milled the saddle to fit a carriage locking mechanism and the 123 block actually touched part of that mechanism so it was a solid touch-off. However, .0002 is nothing in my book to worry about since I'm only a hobbyist and have no military contract orders to fill.

 

[redvan22]

It looks like you're seeing less than 2 thou of difference between all 3 measurements.

Your calibration procedure looks right as far as I can tell, but the "snug fit" of the carriage against the 123 block could vary enough to explain it. You could try it with a dial indicator to get a more repeatable position.

Or it might just be the tolerance/rigidity of the lathe. I don't know what machine you have or what condition it's in, but you could get some flex or slop that would explain this.

Or it could be measuring tools. Looking at some random digital calipers, they list their tolerance of +- .0012 so even in that one caliper there's enough tolerance to explain this difference and then there's the tolerance of the scales themselves that stack on top of that.

So I don't think you're doing anything wrong, but it just might be as good as your setup can do. Chasing these last few tenthou's down is probably above my paygrade, but if it were me, I'd turn off that 4th digit(except for calibration) and accept that level of tolerance. Maybe Yuri or some others that are in those paygrades above me can offer some other suggestions.

I think you missed the overall picture. Every cut requires tweaking the CPI to get it to display an accurate measurement - this should not be happening. It should be calibrated and used with some degree of accuracy repeatedly without the need to adjust the CPI after each cut. What would be the point of having it then?

It's been said by you and Yuiry that 5080 is spot on but look at the CPI numbers I've had to use to get close on my cuts, they're nowhere near 5080. I'll take your advise and eliminate the fourth digit, not necessary for the work I do anyway but I did expect it to be closer than 3-5 thousands.

I have a Micromark lathe and mill. For the lathe, I've lapped and scraped both the cross slide and compound respectively to get as much rigidity out of them as possible. I even added the brass carriage adjustable gibs from a contributor here at this site to the carriage. I can take a .025 cut in steel with no chatter. The machine could probably be a little tighter but its 300% better than it was out of the crate. It doesn't get used every day, I'm a hobbyist, but it gets a regular cleaning and oiling and I've even installed button oilers for quick lubrication of the dovetails.

As for the mill, I lapped every dovetail, including the headstock, until it's smooth and tight. I can take a .020 - .025 cut in mild steel with no chatter. (Headstock locked of course)

My thoughts are that it's the tablet I'm using since I've experienced this strange behavior on both machines using the same tablet for both but I do not know for sure if that could have a bearing on this or, as you said, it could just be the limits of the machine. Someone with more technical knowledge would have to speak up.

Thank you again,
Michael

EDIT:
Had the lathe doc wrong.

 

[mchasal]

I think you missed the overall picture. Every cut requires tweaking the CPI to get it to display an accurate measurement - this should not be happening. It should be calibrated and used with some degree of accuracy repeatedly without the need to adjust the CPI after each cut. What would be the point of having it then?

Sorry, I hadn't seen your updates when I posted that reply. Right, you should not be needing to adjust the CPI between cuts and it would not be very useful if you did.

Do these variations only happen when you actually cut? Like can you setup the 123 blocks, do the calibration, then keep repeating that same movement and always get 3.000" (or whichever side) over and over? And try different locations in the travel and different sides or other measurements.

If it does repeat reliably with no force, maybe the stress of cutting is actually causing something to move? Like the scale or the read head isn't quite stable or even the tool or compound slipping?

I don't know if the tablet could cause something like this. I guess it would have to be missing pulses from the scales, or the scales themselves could be flaky. Yuri would be better to elaborate on possibilities there.

 

[JFL4066]

My 2 cents...
Calipers are not very accurate and repeatable.

You should buy or borrow a Mitutoyo or Starrett micrometer and depth micrometer with the standard. Use those to measure the diameter and length of part. I have personally seen calipers off 1.5 to 2 thou. They really should only be used to get a ball park size when machining up to 5 thou then check with a mic.

 

[redvan22]

Sorry, I hadn't seen your updates when I posted that reply. Right, you should not be needing to adjust the CPI between cuts and it would not be very useful if you did.

Do these variations only happen when you actually cut? Like can you setup the 123 blocks, do the calibration, then keep repeating that same movement and always get 3.000" (or whichever side) over and over? And try different locations in the travel and different sides or other measurements.

If it does repeat reliably with no force, maybe the stress of cutting is actually causing something to move? Like the scale or the read head isn't quite stable or even the tool or compound slipping?

I don't know if the tablet could cause something like this. I guess it would have to be missing pulses from the scales, or the scales themselves could be flaky. Yuri would be better to elaborate on possibilities there.

The instructions that came with the kit said to mount the reader close to the scale but not touching. Only the wipes should be touching the scales. Is there a specific distance required between the reader and scales to get an accurate reading?

 

[mchasal]

The instructions that came with the kit said to mount the reader close to the scale but not touching. Only the wipes should be touching the scales. Is there a specific distance required between the reader and scales to get an accurate reading?

Not sure, but the specs for your scales should provide one if it's important. I don't know if that could cause inaccuracies or just not reading at all though.

I'd try an isolate it to cutting vs not cutting first to try and narrow down where the issue could be as I posted above.