I am gaining .0001 every time I move +X and -Z. If I move +X ten times my zero is off .001. Over the scope of my program X and Z get to far out.
I verified this by setting jog to .001 and slowly alternating X. After ten moves my zero is out .001.
-X and +Z do not gain the .0001 and work fine.

I am running Mach3, XP, Sherline lathe, Sherline stepper motors.
I have Sherline 1/2 pulse mode checked.
X axis step pin# is 3, Dir pin#2
Z axis step pin# is 7, Dir pin#6
Dir Low active and Step Low active are both checked
Step port and Dir Port are both at 1
Steps per is 16000, motor velocity is 37, acceleration is 3, step pulse is 2

Any help would be greatly appreciated.

What Sherline model lathe do you have?
How much backlash do you have?

RICH

The lathe is a 4400. The backlash in mach is not checked. There is not much backlash in the lathe I would guess around .001 and it doesn't bind at all.

Have you tried setting the Step Pulse active High?
Hood

I know my Sherline had more than .001"  Z backlash and that was using the backlash locks with new screw and nut etc.
I also did a number of tests of the screws and they were quite accurate in both directions.

Something you can do is run the axis calibration feature but do it for the longest distance possible distance and adjust
the steps per slightly. Make sure you take out the non-movment of the axis "before" you use the calibration festure. That would take care of a gaining or loosing screw pitch, but, like I said, actualy three Sherlines were checked and the screws were very linear and consistent. If you can I would check the screw just to rule it out.

RICH

Setting the step pulse active high fixed the problem.
Thank you Hood and Rich for your help

This only helped until I rebooted Mach3 then it went back to missing steps when direction is changed. I tried lowering acceleration and speed in motor tuning, but it just ran slower and still missed steps. I searched google for similar problems and read for hours. I tried different settings at random but nothing seemed to help.

I removed the Z axis motor from the lathe and reinstalled the wheel to see if the motor is still missing steps. I did this to eliminate the backlash nut, gibs, and any other form of binding. It was still missing steps.

Next I suspected XP and all of the programs and data I had on there. I had already disabled everything, but since I had a hard drive laying around why not start fresh.

I installed a 19GB hard drive into my tower, so I could format and reinstall XP and not loose my data.
I formatted the disk in ntfs, installed XP SP1 in standard PC mode. Before it was running SP2 so I thought SP1 might be better.
I followed the instructions on http://www.machsupport.com/downloads/XP_Optimization.txt except for #9 because I installed in standard mode.
I ran driver test. Preparation was successful, PPS ran between 25038 and 25039, pulse rating is system excellent. 
I downloaded and used the Sherline lathe (inch)xml found at ftp://machsupport.com/XMLs/Sherline%20Lathe%28Inch%29.xml

In bios
Peripheral config mode is set to ECP, ACPI Suspend State is S3(STR). I also set Peripheral config mod to EPP1.9, but it did not help.

I start Mach, selected Sherline lathe and select manual. It reads that the driver successfully initialized.
With the motor on the desk it still looses steps jog inc .01 feed 6 or feed 1 alternating direction.
It will not loose steps if I keep going in the same direction. It only happens with a direction change. After 10-12 direction changes it looses .001 on the dial of the wheel. My Z would get .012 out after peck drilling 1.5" when it was hooked up to the lathe

I moved the wire from the PC to the driver box and the driver box to the motor away from any other wires, which also did not help.

I can install the Linux software that came with my lathe and see if it is still losing steps. It is the only way I know to eliminate the software. I don't want to use Linux because I can write my own macros in mach using VB.


Any help would be greatly appreciated.

Lewis

 

I would not trust the dial reading. Use an indicator to check for backlash and you will have backlash in that Sherline. Also make note that
The best resolution that you can achieve would be 1/16000=.0000625" which you will probably never achieve, so doubt .0001" movements are actualy happening when changing direction.

Move in one direction to remove the backlash, then change the jog increment to 0.0001", now see how many jogs at that setting it takes to
move the indicator reading .001". Chances are it is not 10. You will see the DRO change by .0001" but there will be no actual axis movement.

RICH

My X backlash is ~.003 and my Z is ~.005. I adjust for this in my macro and is not a issue.
The dial only tells me that every time I change direction the motor is advanced slightly clockwise.

This code will advance my X axis ~.001, but Mach will still read X0. The code is not debugged but hopefully you will get the idea.
Code "G1 X0 F6"
dim i
for i = 1 to 10
 Code "G1 X.01 F6"
 Code "G1 X0 F6"
next
F1 will produce the same results

If I make the loop for i = 1 to 100 then I will be ~.01 advanced clockwise, but mach will still read X0.
In my macro I have to program a stop five times to adjust my zero so it will home correctly and cut to my +/- .001 tolerance.

Tonight in my macro I am going to start writing a sub that will keep track of X/Z movements and adjust the axis value to compensate for my problem. I am not to fond of making a band aid, but I can't afford servos and don't know what else to do.

Thanks
Lewis

I don't remember all the details but wasn't there something about slow optos in some drivers that caused missing steps immediately after a direction change. I seem to remember that happened one way only.

Ian