Another 7i96 Error finishing read

Follow up. I'm still unable to replicate the finishing read/write errors, but I've now had a chance to do some test cuts and they're coming out the wrong shape. Could this be a result of changing the servo period?

It seems that the Y axis (dual motor) is being shrunk. Perhaps missing steps? The issue progressively gets worse if the machine isn't re-homed between jobs.

Attached are two separate files. The first you can see how the bolt holes are oval/oblong, and the second is supposed to be 7" square. It appears this is only happening on the Y axis, but my brother was the one running these jobs and sending me the pictures so can't confirm 100%, but the one cut was dimensionally accurate to 7" along the X axis.

 
 

Any ideas?

Could this be a result of changing the servo period?

No, the number of steps and scaling are independent of the servo period.

Could be mechanical or marginal step timing
(try doubling the STEPLEN parameter and see if anything changes)

I'll have to try updating `STEPLEN` next time I'm out at the shop.

With this new symptom, in addition to the original issue, do you think this might be pointing to a bigger issue? Like a failing card?

The original issue was a host setup problem

Step loss is unlikely a 7I96 issue unless a step
output pin has been damaged so has marginal levels.

You can check this with a voltmeter at the 7I96 step pins
by reading the voltage when LinuxCNC is running and
not running. The step/dir outputs should be close to +5V
in one case and close to 0V in the other.

chadfawcett wrote:

With this new symptom, in addition to the original issue, do you think this might be pointing to a bigger issue? Like a failing card?

Not yet, but i am fairly certain it can come to pas as i am fairly certain the issue is with bad wiring or grounding or shielding.
Do some cuts in a straight line, in both directions X and Y, but do them as 'cut, no cut, cut, no cut' so 5 cuts of 5cm with 1cm in between, post pictures.

chadfawcett wrote:

Follow up. I'm still unable to replicate the finishing read/write errors, but I've now had a chance to do some test cuts and they're coming out the wrong shape. Could this be a result of changing the servo period?

It seems that the Y axis (dual motor) is being shrunk. Perhaps missing steps? The issue progressively gets worse if the machine isn't re-homed between jobs.

Attached are two separate files. The first you can see how the bolt holes are oval/oblong, and the second is supposed to be 7" square. It appears this is only happening on the Y axis, but my brother was the one running these jobs and sending me the pictures so can't confirm 100%, but the one cut was dimensionally accurate to 7" along the X axis.

Any ideas?

If it's consistent check the Y axis scale, if it's random check the smoothness of the Y axis and that there are no sticky spots on that axis however usually if a stepper misses steps due to interference then the stepper stalls and just sits there and hums.

Run a stepper test for each axis as outlined in the documents Configuration > General Info > Stepper Diagnostics.

JT
 

Thanks for all the responses! Sorry for the delay, I don't seem to be notified of followup responses so didn't realize there were more.

It ended up being a hardware issue after all for the Y axis stuff. The grub screw on one of the motor pulleys wasn't tight so it was slipping on the motor shaft. Tightened that back up and everything functions as expected now. 

Thanks again for all the replies.