PM25 DRO and Hand wheel off by 1 part in 1000?

Bill Kahn

Registered
Registered
Joined
Mar 19, 2017
Messages
181
I just got my PM25MV. Lots of great beginner experiences. (Amazing how hard it is to indicate the head rotation to 1 mil over 18". But I got it. The nod on the head over 8" is 3 mil, but seems like that adjustment is not for a beginner.) But it is all set up and I am enjoying it way too much. Life is good.

A question: I spin the X handle 210 times. Exactly. With backlash taken out. DRO says I have moved the table 21.025". So, the DRO and the mechanical handle differ by 25 mil in 21 inches. About a mil per inch.

Which is right? More trustworthy?

I tried to indicate, but over the 1" of my dial indicator the 1 mil difference is in the noise of indicator (guess now that I have spent $2000 on the mill I really should spring for more than $15 for a proper indicator. But it is what I have now.)

Any thoughts as to why they differ? (I have taken readings every inch. And the effect is cumulative. Not like all 25 mil comes in at one end. Right through the center of the X-travel you can see the mil difference every inch slowly accumulating.

(And, I have no idea at all if I will ever do anything requiring .1% accuracy. On my drill press with HF cross-feed vise I have been almost fully happy with 1% accuracy. )

Ahh, a man with one watch knows the time. A man with two doesn't.

Thanks for any thoughts.

-Bill
 
  • Like
Reactions: Guv
Asian mills and lathes often use metric threads on their lead screws but have Imperial dials. They use the approximation of 1mm = .03937" is approximately .040". However, In your case the error is much closer than that.

At .1000"/revolution of the lead screw, this has to be an Imperial lead. There are no metric thread leads close to it. My guess is that the lead screw was cut on am etric machine and they approximated the pitch, figuring that .1% error was good enough.

I would trust the DRO over the dial. Particularly if you have the magnetic or glass scales. The scales are precision made and, in my experience, come with a calibration document.

There are a variety of ways that you can verify the DRO accuracy. An edge finder and a micrometer calibration standard come to mind. You can machine the ends of bar and measure with a micrometer. The diameter of the end mill will need to be taken into consideration but that could be accommodated by making two cuts, one at one inch and another at five inches and calculating the difference between the two measurements.
 
DROs have a specified accuracy, but usually with glass and magnetic scales it is more noise or fluctuations as opposed to cumulative. So I would not expect to see a fixed change for every X revolutions of the handle. I had something similar in my previous bench top mill, and the error was very close to what you have. The problem was they used metric pitch feed screws and imperial dials. This was for all axis. But it could also be a error in the feed pitch. Per above, I would trust the DRO.
 
Another vote for trusting the DRO. I typically check the accuracy of a new DRO with a gauge block, or a 1/2/3 block of a known (and measured) dimension, and a edge finder.

I don't even read dials on handwheels anymore.
 
The problem would have to be a lead screw that has the incorrect lead per revolution, or the DRO is not parallel to the axis, causing cosine error. I understand that some DROs are adjustable for inconsistencies, though I have never looked into that. The real question is whether either of the two is correct, and the way to test that is to use gage blocks or other trustworthy measuring standards to check against both the lead screw and the DRO.
 
Asian mills and lathes often use metric threads on their lead screws but have Imperial dials. They use the approximation of 1mm = .03937" is approximately .040". However, In your case the error is much closer than that.

At .1000"/revolution of the lead screw, this has to be an Imperial lead. There are no metric thread leads close to it. My guess is that the lead screw was cut on am etric machine and they approximated the pitch, figuring that .1% error was good enough.

I would trust the DRO over the dial. Particularly if you have the magnetic or glass scales. The scales are precision made and, in my experience, come with a calibration document.

There are a variety of ways that you can verify the DRO accuracy. An edge finder and a micrometer calibration standard come to mind. You can machine the ends of bar and measure with a micrometer. The diameter of the end mill will need to be taken into consideration but that could be accommodated by making two cuts, one at one inch and another at five inches and calculating the difference between the two measurements.
RJ, Thank you for your thoughts. Yes, the hand wheel accuracy is way better than saying 25.4mm per inch is about 25 mm per inch. But I took your suggestion about how to directly measure given .1% error is still in the noise of my 1" $15 indicator. I have a 1-2-3 block. Cheap but my $25 digital caliper says it is 3.000" long. So, I have some confidence in it. I used an edge finder on it. My reproducibility was .5 mils. Over the (presumably correctly known) 3.2" (.2 extra for the edge finder diameter) the DRO was off by -.4 mils (3.1996", which is in the domain of reproducibility noise) and the hand wheel scale was off by 4 mils (3.204" which is way outside reproducibility noise).

I find this simply fascinating. I was vacillating on whether to get the $500 DRO upgrade on a $1500 mill. How hard is it to read dials--that seemed just fine for me in high school 45 years ago. I had no idea that the hand wheels were off by so much. I am most glad I went with the DRO--doing a .1% correction on every precise location would have gotten very old very quickly.

It strikes me .1% for real machining is actually a lot. OK, I am coming from a drill press with a $50 HF cross feed vise. So .1% is way better. But why would they make something so close but clearly wrong? (Feels like the "uncanny valley" of almost human robot faces--quite disturbing.)

I don't see how this could be a bad apple--the screw seems to actually have the wrong pitch.

Yet the rest of the machine seems perfectly well built and designed. And the company Precision Matthews seems very responsible--they have been very good not just with pre sales support but also post sales support.

I wonder if others with a PM25MV with DRO can replicate my findings.

Also, given I am a beginner, I am still reserving some probability for this is all just neophyte, not machine, trouble.

Thanks again for your edge finder suggestion.

-Bill
 
It is a common practice when cutting threads where you don't have the proper gears for an exact pitch to pitch one that is close. As an example a 10-32 thread is fairly close to a M5 - .8mm (4.83mm dia. by.794 mm) and the 10-32 screw would engage an M5 - .8mm nut. Not the best practice as threads are only engage for a short distance and prone to stripping under load but if there is no other option, it will work. But over 100 threads, the error would be .63mm or .025". If the manufacturer couldn't cut Imperial threads with their lathe, they just might fudge it and cut something close. The lead screw nut, being relatively few threads, would fit even if its pitch was correct. For all intents and purposes, the difference would be unnoticed by the end user as it is unlikely that anyone would measure the length by counting dial rotations. For my part, when I got past ten, I usually lost count.:( Then along comes someone like you who counts out 210 turns.;)

I would guess that if the lead screw were cut on a CNC lathe, someone could have errored in the programming. With Chinese manufacturing who knows?

At any rate, you have a DRO now and as others have said, you will soon ignore the dials completely. My Tormach mill is entirely by the DRO as there are no dials and virtually all moves on my mill/drill are by the DRO.
 
Back
Top