Are glass scales affected by vibration?

[ChazzC]

ToAuto DRO and scales from fasttobuy2012

ToAuto Digital Readout DRO Display 5µm Linear Scale Glass Sensor Lathe Mill EDM | eBay

Milling Machine Digital Readout 2/3/4/5Axis DRO Display Linear Glass Scale 5µm TTL Sensor CNC 1, Application : This kind of DRO is commonly fitted to machines, especially for metal working — lathes, cylindrical grinders, milling machines, surface grinders, boring mills and other machine tools —...

www.ebay.com

I don't recommend it. Wrong manual, no support. Seems to mostly work but there's no way to combine Z and quill.

I don’t have glass scales, but per the TouchDRO website while good Chinese scales are available (Ditron & Easson), the lower cost scales can have issues. Since your complete setup cost about the same as just the two (6” & 12”) iGaging SS Absolute Plus capacitive scales I bought 6 months ago, I have to assume you have a lower cost system/scales; this is most likely the root cause of the issues.

Sorry

 

[WobblyHand]

Suggestions so far:

1. Plug scale into different channel on DRO (mark)
2. Bad read head (Wobbly)
3. Installation/alignment (Wobbly, OT)
4. Adjust sensitivity (radm1)

Tried plugging X scale into Y input on the DRO. Good suggestion but no change. Probably the DRO isn't the problem.

The other issues are addressed below.

Here's the test setup:
View attachment 472884
I'm indicating in the pocket, getting inconsistent readings in one direction but not the other. When I move the work left, I get 0.372,2 When I move right, I get a variety of values, mean 0.380, standard deviation 0.001,2. So one side repeats, the other has random error, +/- 0.002,5. I've tried this several times and sometimes I get a +/- a tenth on the left side but the right side changes wildly.

If the error was random, I wouldn't able to return to the same spot. If the error was the same in both directions, it would have to change slowly so that I get the same size error traversing right and when returning. If that were the case, how would I get different error for each cycle of measurements?

I think the scale or read head is shifting position. There's a hard stop as I move the work left but something is moving when I move the work to the right. While fiddling around, I noticed that slow movements resulted in much less error. So if I back off 10 thou and slowly return, the reading repeats. If I turn 10 thou past 0, the reading changes but then repeats. This all sounds like backlash and somewhat reluctant sliding somewhere in the system.

The alignment was set with the spacer installed (see photo). Then I aligned the scale with the dovetail (right) before fastening the ends. So alignment should be within 0.010 over 18" of travel.

@WobblyHand If you're right about the failing read head, it's going to be a lot of work to replace it. I had to do quite a bit of modification to get the read head to fit under the table.
View attachment 472885

I hope it isn't the read head, but that's what it was for my glass scale. It would miss or skip counts in one direction vs the other. Never would have found it without simply going back and forth over 3 to 6 inches repeatedly using the DRO, the 123 block and the DTI. In the beginning the error was so subtle I blamed myself for some machining errors. As time went by, the error got worse. It was somewhere around 0.050" over 3". Then it got so bad, I could clearly see the error on the table x scale. After replacement of the head, the error was 0. Wishing you luck on this.

Check that the scale alignment is still good over the travel. Not sure if your scales can tolerate your alignment error. Did that number of 0.010" come from the scale manufacturer?

 

[rodm1]

Check that the wiring on the scales are the same as what the DRO requires. I had to rewire my scales do to incompatibility (skipped counts) . This shouldn't be a problem since you used the same vendor for both. My scales was getting erratic reading.

Are you positive the scale and read head are tight?

How to Find Glass DRO Scale Pin Functions

Chinese Glass DRO Scales have become very inexpensive, and can be found on eBay, AliExpress, etc. for under $100/piece. Unfortunately many sellers are unable or unwilling to provide accurate documentation, so it's often up to the buyer to figure out what the pinout of the particular scales is...

www.touchdro.com

On page 11 has the pin outs one of those should match the DRO.

http://www.yihaodro.com/uploads/soft/20210525/1-210525225235P2.pdf

 

[durableoreo]

I removed the scale, set some hard stops, and checked if the drift is coming from the scale. I'm moving it on the spacer so alignment should be OK. It's touchy but if I really concentrate on "feel" (without watching the DRO) I get very consistent readings. So it must be the installation that is the problem. Before disassembling, I pushed on the scale while watching the DRO. There was no detectable lash. I couldn't access the read head---maybe it is not held as securely as I thought.

Yesterday, I tried to rig up a little vibration test. Ridiculous. This class of milling machine would need a lot more to cause vibrations.

 

[durableoreo]

I took everything apart and put it back together. Problem solved.

When re-assembling, I took extra care to seat the read head in its mount. There is an undercut in the bracket (right) and the read head has been modified to mate with it. Attaching the read head to the bracket must be done with the table on the mill, so it's meant to slide in and be easy to attach. A cap goes on the near end (left) which also has an undercut. I tightened everything a bit more than feels "right". After monkeying around and fiddling with this and that, I got repeatable measurements.

While I was stewing about the problem, it occurred to me how few people are interested in this sort of thing. Most people don't even know what a thousandth of an inch is, much less tenths. I checked the US Census and found that there's about 350,000 machinists in the USA. That's 0.1% of the country. And how many of them will answer your questions? Thanks for your help.

 

[WobblyHand]

I took everything apart and put it back together. Problem solved.

When re-assembling, I took extra care to seat the read head in its mount. There is an undercut in the bracket (right) and the read head has been modified to mate with it. Attaching the read head to the bracket must be done with the table on the mill, so it's meant to slide in and be easy to attach. A cap goes on the near end (left) which also has an undercut. I tightened everything a bit more than feels "right". After monkeying around and fiddling with this and that, I got repeatable measurements.

While I was stewing about the problem, it occurred to me how few people are interested in this sort of thing. Most people don't even know what a thousandth of an inch is, much less tenths. I checked the US Census and found that there's about 350,000 machinists in the USA. That's 0.1% of the country. And how many of them will answer your questions? Thanks for your help.

Glad you were able to fix it!

I suspect that there's more than a few hobby machinists that aren't counted in the census. The census is work oriented, not hobby centric. I got the privilege of answering additional census forms, this last go around, they didn't ask me if I knew what a tenth was