PM25mv, how to sort out X axis discrepancy between DRO and handwheel

[DJMaker]

Forgive me for any ignorance, I'm new to machining.

I have a pm25mv that displays an 11 thou difference between the DRO and handwheels over the length of the table (if I match them up at one extreme of travel, there'll be 11 thou difference at the other end of travel). This isn't much, roughly 1 thou every three inches of travel.

Right now I'm assuming this is simply the level of accuracy I should expect from an inexpensive machine, but since I'm new to this, it doesn't hurt to ask.

Questions:
1) Is this typical? Over 27" is it expected that 11 thou of error could accumulate in the screw or DRO accuracy?
2) Are there any tips on how to accurately calibrate my DRO? I rate my measuring ability at +- 0.0005 at best, so measuring over just a few inches isn't enough to account for this error.

 

[burtonbr]

I'm no expert, but I think you could verify the DRO accuracy with tools you may have or can obtain inexpensively. Gauge blocks or 123 blocks should be able to be used with a test dial indicator set up on the table and zero at one point and measure to another known point 4"-6" with a TDI and see if the DRO is different, not sure my explanation is clear but I think I remember a your tube video showing something similar I may can find.

 

[DJMaker]

I'm no expert, but I think you could verify the DRO accuracy with tools you may have or can obtain inexpensively. Gauge blocks or 123 blocks should be able to be used with a test dial indicator set up on the table and zero at one point and measure to another known point 4"-6" with a TDI and see if the DRO is different, not sure my explanation is clear but I think I remember a your tube video showing something similar I may can find.

Thanks I did just that, measured a 123 block with an indicator and my DRO did seem more accurate than my handwheels. The handwheels consistently measured about 1 or 2 thou off over the three inches, which would account for the difference. DRO was nearly spot on, within .0005.

I've noticed that my handwheels have about +/- .0015 random inconsistency, likely due to machining imperfections, so I could be measuring that instead of something that would lead to the much larger accumulative error I'm seeing over the length of the table. I guess the next thing is to do test it with the table in different positions to see if the error is consistent.

My bigger question is if this amount of error in the leadscrew is typical? a variation of 11 thousands over 27 inches may just be normal. That's more due to my curiosity on the limitations of the accuracy of leadscrews.

 

[burtonbr]

I had a G0704 I added glass scale Dro to it and it was fairly accurate I don't remember if I ever checked the error between the DRO and the dials since they became irrelevant once watching the dro scroll, it is probably more important that it is repeatable, that is if you set an axis 0 & crank the table 6"-10" in one direction then return the the original 0 are you in the same location or relatively close within error margin, that can be checked with dial indicator set to 0.
I remember being able to repeat within .001" and felt that for the class machine and DRO it was that was amazing .

 

[rwm]

Is the leadscrew error consistent and cumulative? I have seen this with metric lead screws that had Imperial handwheels attached. Jet 920 lathes do this. I believe the PM 25 is advertised to have Imperial leadscrews.
Robert

 

[benmychree]

Any machine leadscrew should be more accurate than what you are seeing.

 

[DJMaker]

Is the leadscrew error consistent and cumulative? I have seen this with metric lead screws that had Imperial handwheels attached. Jet 920 lathes do this. I believe the PM 25 is advertised to have Imperial leadscrews.
Robert

The error is fairly consistent and linear, I can predict it by dividing 0.11 by the percentage of table travel. Generally I only have about +/- .001" repeatability with the hand wheels, the DRO is reatable to +/- 0.0002"

lead screws are imperial---0.10 inch per turn, the metric/imperial mashups are typically a less round number per turn due to the conversion. If that was the issue the error would be far greater.

 

[DJMaker]

Any machine leadscrew should be more accurate than what you are seeing.

I do notice that table is harder to move with the handwheels at one extreme vs. the other, could be a slightly bent lead screw.

 

[burtonbr]

I’d think that it being stiffer at the ends for travel is more gib getting tighter than lead screw issue. If it is repeatable within tenths over 6” or more inches that would be acceptable for most work for that mill I’d think. Acme lead screw with brass nut and not a ball screw if the DRO os holding true that’s pretty good.

 

[rwm]

I agree. A metric lead screw will typically be off .004 per 1". That is a weird problem. I don't think a bent lead screw would account for it. Maybe you just got a bad lead screw?
R