Ballscrew Compensation

[mick41zxr]

Having an odd issue in setting up ballscrew compensation.
I have mounted a glass scale on the table of a BF20 style mill using closed loop steppers.
Machine runs smoothly & backlash dialed in. Went to try & setup screw compensation as per Alan's excellent videos. Compensation required is minimal +-0.015mm, however as soon as I enable screw compensation (in motor section of setup) the errors increase dramatically & any changes to screw compensation table do not make any difference.
These errors are large enough to be measured with a tape & verified with scales (Easson) +-1mm.
I am sure I have missed something, any help appreciated.
Michael

report_38D2693E210E-0103180443_2018-10-16_21-56-40.zip

(194.77 KiB) Downloaded 138 times

[AMHKnives]

Ill look over your compensation tables as soon as i get off work to try and help you out bud.
Alan

[AMHKnives]

Just looked through your report, the tab files for the ballscrew comps for all 3 axis dont have any comp changes in your report file. This is what they should look like when you have changed them.
NC10 TAB FILE VERSION 3
AXIS X
UNITS INCH
INTERVAL 0.5
ENTRIES 32
+0.00000000 +0.00000000 +0.00000000
-0.50000000 +0.00089000 +0.00153300
-1.00000000 +0.00043700 +0.00164900
-1.50000000 +0.00137000 +0.00246000
-2.00000000 +0.00111000 +0.00233000
-2.50000000 +0.00172000 +0.00313000
-3.00000000 +0.00203000 +0.00341700
-3.50000000 +0.00286000 +0.00439000
-4.00000000 +0.00350000 +0.00491400
-4.50000000 +0.00437000 +0.00577000
-5.00000000 +0.00537000 +0.00674100
-5.50000000 +0.00620000 +0.00768000
-6.00000000 +0.00718000 +0.00858000
-6.50000000 +0.00797000 +0.00938000
-7.00000000 +0.00884000 +0.01013000
-7.50000000 +0.00910000 +0.01047000
-8.00000000 +0.01002000 +0.01128500
-8.50000000 +0.01012000 +0.01129000
-9.00000000 +0.01023000 +0.01165000
-9.50000000 +0.01036000 +0.01155200
-10.00000000 +0.01082000 +0.01221800
-10.50000000 +0.01120400 +0.01217000
-11.00000000 +0.01160700 +0.01316000
-11.50000000 +0.01176000 +0.01289000
-12.00000000 +0.01205000 +0.01362000
-12.50000000 +0.01267000 +0.01358500
-13.00000000 +0.01313000 +0.01442600
-13.50000000 +0.01392000 +0.01509000
-14.00000000 +0.01432000 +0.01524000
-14.50000000 +0.01518000 +0.01627000
-15.00000000 +0.01573000 +0.01618700
-15.50000000 +0.01670000 +0.00000000
Hope this helps some.
Alan

[mick41zxr]

Thanks Allan,
That's the problem, with comp off I am +- 0.015mm, however as soon as comp is turned on I am out by +-1mm.
No matter what changes I make to tables the error does not change.
Regards,
Michael

[AMHKnives]

Would you be able to save a table and send it over in a report, just want to see the corrections your making on it.
Thanks
Alan

[mick41zxr]

Ok had another go at comp & I think I may have it sorted.
I misunderstood how to correct the corrections so to speak.
However I still don't understand why the measurements varied between comp turned off to comp on before any data entered into the comp tables.

Below are measurements with comp off.

Measurements with comp on but no data entered.

Michael

[mick41zxr]

1st iteration of comp table with data entered.

Measurement after X axis homes with comp table.

[mick41zxr]

After a couple of iterations of corrections, getting closer but a little more work needed before moving to next axis.
Comp table.

Measurements

System report

report_38D2693E210E-0103180443_2018-10-19_22-31-52.zip

(196.87 KiB) Downloaded 140 times

Regards,
Michael

[AMHKnives]

Im not sure why it would change the moves with comp on with no data entered. It looks good through, getting it dialed in nicely bud.
Alan