[How do I?] DRO troubleshooting

[ChazzC]

No, he's on batteries- I don't see how grounding would matter
Possible I suppose but I'm doubtful.
More doubtful than the usual amount- doubtful 2.0

Check Yuriy’s remarks on the linked page in post #7: there is internal grounding in the read head that can go wonky).

 

[slow-poke]

Good scales use differential quadrature signalling to provide good noise immunity.
One step down uses simple quadrature single ended TTL levels that are not as immune to noise as above.
IIRC those scales use single ended non quadrature serial data that is quite susceptible to noise, IMO they are not really suitable for an industrial environment. I think you will be chasing your tail until you move to one of the better signalling schemes. I would not be surprised the ones that appear to be working are actually intermittently dropping packets, you just don't notice it.

 

[ChazzC]

Good scales use differential quadrature signalling to provide good noise immunity.
One step down uses simple quadrature single ended TTL levels that are not as immune to noise as above.
IIRC those scales use single ended non quadrature serial data that is quite susceptible to noise, IMO they are not really suitable for an industrial environment. I think you will be chasing your tail until you move to one of the better signalling schemes.

I haven’t had any issues with my iGaging Absolute Plus scales, and while my home shop is not an industrial environment, I don’t believe @ARC-170 ’s shop is either. There are known issues with the iGaging EZ-View scales, which can be addressed as noted by Yuriy. The EZ-View Aluminum scales are also more vulnerable to how they are physically mounted than are the SS Absolute scales.

 

[ARC-170]

No, I'm a hobbyist. My mill is 220V, if that means anything. I'll get photos when I have some time to take it apart, maybe this weekend.

Is it important that the internal wires be grounded to whatever the scales are attached to? In other words, is the attachment connection for the scales really important?

I have a feeling this is a mechanical and mounting issue. That z-axis was hard to get everything attached to well. If it's not that, then I might have a bad reader, either a physical or electronic issue. I'll have to test it and see.

 

[ChazzC]

No, I'm a hobbyist. My mill is 220V, if that means anything. I'll get photos when I have some time to take it apart, maybe this weekend.

Is it important that the internal wires be grounded to whatever the scales are attached to? In other words, is the attachment connection for the scales really important?

From the TouchDRO link; this is how the scale is grounded to the read head:

I have a feeling this is a mechanical and mounting issue. That z-axis was hard to get everything attached to well. If it's not that, then I might have a bad reader, either a physical or electronic issue. I'll have to test it and see.

That is my thinking.

 

[ARC-170]

From the TouchDRO link; this is how the scale is grounded to the read head:

View attachment 510552



That is my thinking.

I saw that on the Touch DRO page that was linked. Looks like I might have to open my cases and see if the wires are connected.

 

[RJSakowski]

Looking at your numbers, it doesn't appear that you are losing counts on the DRO. After .020" by the DI, the DRO tracks fairly well with the DI reading. A dial indicator, especially a low cost one,can have mechanical issues, sticking, skipping a tooth, etc. I would bet on the dial indicator sticking. They use an internal hairspring to minimize backlash in the gears but if the mechanism is sticking, it can give erroneous readings.

I would do this:
1. switch display with one of the other scales to see if the problem rests with the display or the scale/read head.
2. Beg or borrow another dial indicator and run your comparison again to see if the problem still exists.

I have three of the iGaging DRO's on my lathe and have never seen any inconsistencies in the readings.

 

[Rex Walters]

DTI is from LMS

Sorry to be so pedantic, but that's a "dial indicator" or "drop indicator" (DI). The term Dial Test Indicator (DTI) to me means those with a lever arm like this one:



DTI's are usually more accurate (and pricier), with smaller measurement ranges, but allow you to reach into small areas otherwise inaccessible with a DI.

As Rick points out, you'll likely find it easier to get repeatable results if you measure over a MUCH wider range than 0.100". A pair of 1-2-3 blocks, gage pins/blocks, or micrometer standards work well as he suggests.

As he also points out, mechanical standards and hard stops are far more precise than any adjustable measuring device (indicators, micrometers, calipers, etc.). But quality indicators should be extremely repeatable: they excel at +/- comparative measurements. For example: zero the indicator dial on one reference, then compare the reading to something else — if it doesn't also read zero then they are different sizes. Indicators are less useful for quantitative measurements (they are fine for coarse measurements, though).

So my preferred technique differs from Rick's only slightly: I use a pair of 1-2-3 blocks to create a 4", 5", or 6" reference, then mechanically attach a DTI to the spindle and zero the indicator on the top of the stack. The DTI eliminates any need for "feel". Remove the stack and lower the quill (or raise the knee) until the indicator reads zero again. You can be confident the indicator tip has moved exactly, say, 5.000" (if you've verified the precision of your 1-2-3 blocks and stacked then for a 5" reference).

Since you said your results aren't repeatable and are all over the place, it's probably an electrical or quality issue with the DRO. Mechanical issues should at least be repeatable (except for sticking or floppy mounts, which should be somewhat obvious).

I'm better with mechanical troubleshooting than I am with electrical troubleshooting, though, so here are some thoughts to at least verify the mechanical aspects:

• As above, use hard stops, physical standards, and comparative indicating rather than measuring anything whenever possible.
  
  
• Avoid any backlash effects. Ensure you do all of your measurements traveling in just one direction. E.g. always lower the quill until the dial reads zero and don't back up. If you overshoot, back way up, then travel back down until it just reads zero.
  
  
• My pet peeve: some of my offshore equipment dials effectively estimate 1" as 25mm (an error of 1.5% — one inch is 25.4mm). They use metric screws with metric pitches instead of actual inch screw threads with true turns per inch. I doubt this has anything to do with your DRO readings, but I wouldn't be surprised if your mill dials have this problem. Use a 1-2-3 block on its long axis, zero the indicator and your dial, remove the block, then turn your dials until you've supposedly moved exactly 3.000". Don't be surprised if the indicator doesn't zero on the table (or vise top if that's where you placed the block).
  
  
• Angular travel and poor tramming. In theory, lowering the quill should have the same effect as raising the knee (or lowering the head on a bench top mill). This is only true if the spindle is perfectly trammed, though. Tram the spindle to the table as accurately as possible, then use an indicator mounted in the spindle against the the vertical faces of a good square reference to ensure there is no movement in X or Y as you move in Z (a 1-2-3 block or a 2-4-6 block works well). Do the same with the indicator tip agains the side and face of your DRO scale instead of the square to ensure it's running perfectly parallel to the spindle axis.
  
  
• Verify rigidity. Some of my most difficult problems to track down boiled down to a loose screw or something flexing even a tiny amount without my noticing. It's incredible the amount of havoc one loose screw can create. Glued, brazed, or welded attachments are significantly more rigid than screws and nuts/bolts (screwed and glued gives you strength AND rigidity). Use glues that harden when cured, not stuff that remains tacky/soft. A few dots of CA glue under your DRO gage mounts won't hurt anything, and you can still remove them with a sharp tap. A broken bond will also make it obvious if something is flexing or moving. (Use acetone to clean up CA glue detritus.)
  
  
• Understand your dependencies: verify your metrology gear! Get even a cheap small surface plate and some basic metrology equipment (DTI, surface gage, and at least a few calibration standards at a minimum — I recommend acquiring at least a 0.500" and 1.000" gage pin, but I find a full set of gage pins up to 0.500" in 0.001" increments incredibly handy) . Learn how to inspect a pair of matched 1-2-3 blocks: Verify all surfaces are parallel to the plate surface and that the two blocks match in each dimension. Then use reversal methods to ensure the faces of your 1-2-3 blocks are square/perpendicular (glue a ball or even belt-ground round front to the base of your surface gage to perform comparative squareness checks). Ultimately, you should only need to depend on the flatness of your surface plates, the accuracy of your calibration standards (gages), and the repeatability of your indicator. You should be able to directly (or indirectly) verify everything else.

 

[ARC-170]

I might have solved the problem. ALL the screws were loose! LOL! I've reinstalled them with Loctite. Pictures for your viewing pleasure:

View of the z-axis install without the cover from the left side.


Close-up.


View of the z-axis install without the cover from the right side. There are 2 screws holding the aluminum L that hold the DRO. These also hold the cover on. I will probably install separate screws for each, so the cover is independent of the L bracket and DRO.


This was how loose the screws were!


This is how loose they were when I took the cover off! Way too much movement.

I think this explains the readings being off at first then getting somewhat more consistent. The screws were loose enough to have a small amount of play that was taken up by travel.

I'm waiting for the Locite to cure, then I will test.

 

[ARC-170]

Sorry to be so pedantic, but that's a "dial indicator" or "drop indicator" (DI). The term Dial Test Indicator (DTI) to me means those with a lever arm like this one:

View attachment 510626

DTI's are usually more accurate (and pricier), with smaller measurement ranges, but allow you to reach into small areas otherwise inaccessible with a DI.

As Rick points out, you'll likely find it easier to get repeatable results if you measure over a MUCH wider range than 0.100". A pair of 1-2-3 blocks, gage pins/blocks, or micrometer standards work well as he suggests.

As he also points out, mechanical standards and hard stops are far more precise than any adjustable measuring device (indicators, micrometers, calipers, etc.). But quality indicators should be extremely repeatable: they excel at +/- comparative measurements. For example: zero the indicator dial on one reference, then compare the reading to something else — if it doesn't also read zero then they are different sizes. Indicators are less useful for quantitative measurements (they are fine for coarse measurements, though).

So my preferred technique differs from Rick's only slightly: I use a pair of 1-2-3 blocks to create a 4", 5", or 6" reference, then mechanically attach a DTI to the spindle and zero the indicator on the top of the stack. The DTI eliminates any need for "feel". Remove the stack and lower the quill (or raise the knee) until the indicator reads zero again. You can be confident the indicator tip has moved exactly, say, 5.000" (if you've verified the precision of your 1-2-3 blocks and stacked then for a 5" reference).

Since you said your results aren't repeatable and are all over the place, it's probably an electrical or quality issue with the DRO. Mechanical issues should at least be repeatable (except for sticking or floppy mounts, which should be somewhat obvious).

I'm better with mechanical troubleshooting than I am with electrical troubleshooting, though, so here are some thoughts to at least verify the mechanical aspects:

• As above, use hard stops, physical standards, and comparative indicating rather than measuring anything whenever possible.
  
  
• Avoid any backlash effects. Ensure you do all of your measurements traveling in just one direction. E.g. always lower the quill until the dial reads zero and don't back up. If you overshoot, back way up, then travel back down until it just reads zero.
  
  
• My pet peeve: some of my offshore equipment dials effectively estimate 1" as 25mm (an error of 1.5% — one inch is 25.4mm). They use metric screws with metric pitches instead of actual inch screw threads with true turns per inch. I doubt this has anything to do with your DRO readings, but I wouldn't be surprised if your mill dials have this problem. Use a 1-2-3 block on its long axis, zero the indicator and your dial, remove the block, then turn your dials until you've supposedly moved exactly 3.000". Don't be surprised if the indicator doesn't zero on the table (or vise top if that's where you placed the block).
  
  
• Angular travel and poor tramming. In theory, lowering the quill should have the same effect as raising the knee (or lowering the head on a bench top mill). This is only true if the spindle is perfectly trammed, though. Tram the spindle to the table as accurately as possible, then use an indicator mounted in the spindle against the the vertical faces of a good square reference to ensure there is no movement in X or Y as you move in Z (a 1-2-3 block or a 2-4-6 block works well). Do the same with the indicator tip agains the side and face of your DRO scale instead of the square to ensure it's running perfectly parallel to the spindle axis.
  
  
• Verify rigidity. Some of my most difficult problems to track down boiled down to a loose screw or something flexing even a tiny amount without my noticing. It's incredible the amount of havoc one loose screw can create. Glued, brazed, or welded attachments are significantly more rigid than screws and nuts/bolts (screwed and glued gives you strength AND rigidity). Use glues that harden when cured, not stuff that remains tacky/soft. A few dots of CA glue under your DRO gage mounts won't hurt anything, and you can still remove them with a sharp tap. A broken bond will also make it obvious if something is flexing or moving. (Use acetone to clean up CA glue detritus.)
  
  
• Understand your dependencies: verify your metrology gear! Get even a cheap small surface plate and some basic metrology equipment (DTI, surface gage, and at least a few calibration standards at a minimum — I recommend acquiring at least a 0.500" and 1.000" gage pin, but I find a full set of gage pins up to 0.500" in 0.001" increments incredibly handy) . Learn how to inspect a pair of matched 1-2-3 blocks: Verify all surfaces are parallel to the plate surface and that the two blocks match in each dimension. Then use reversal methods to ensure the faces of your 1-2-3 blocks are square/perpendicular (glue a ball or even belt-ground round front to the base of your surface gage to perform comparative squareness checks). Ultimately, you should only need to depend on the flatness of your surface plates, the accuracy of your calibration standards (gages), and the repeatability of your indicator. You should be able to directly (or indirectly) verify everything else.

Thanks for all of this, I really appreciate it. Once the Loctite cures, I will test using these methods and see what I get.

I thought your comment about offshore dials being metric was interesting. 1.5% of 3" is 0.045", I wonder if my mill dials are off that much? I would not be surprised if they are metric with inch markings. I'll test and see.
I'll have to check my dial indicators as well. 1.5% of 0.1" is 0.0015". I'm not sure if I can even read that. One of my DI's is from LMS and I have another from Fowler.