Random Loss of Position

[Hood]

That should be ok.
What kernel are you running?
 Does the Frequency on the Diagnostics page sit steady at or close to the frequency your kernel is at?
Hood

[retro-AC-Servo]

Yes it's close,

On the motor tuning and set up page there a setting for "step pulse" 1-5  and  "dir" 0-5.   Out of desperation I decided to tweek the "step pulse".  Mine was originally set to "0" I changed it to 1.   My problem hasn't re-occurred since I changed it.  What might have I done?   Keeping my fingers crossed...

[Hood]

That could well be the problem.
What frequency are you running the kernel at?

Hood

[Jeff_Birt]

Out of desperation I decided to tweek the "step pulse".  Mine was originally set to "0" I changed it to 1.

Mach likes to put out little teeny tiny step pulses. Since the parallel port driver timing and hardware vary so much this can lead to pulses that are missing or way to short for the stepper drive to 'see'. Often times this problem presents itself as a 'tick' sound in the axis. Lengthening the step pulse or selecting the 'Sherline 1/2 pulse mode' are two ways to force Mach to increase the pulse length giving a better chance for a good pulse to be generated.

[RICH]

Building on what Jeff said, you may want to look at your drive information as sometimes there is a recomended min setting.
RICH

[retro-AC-Servo]

Well it ran great again all day today .  I'm cutting grooves in aluminum.  My feed rate is set at 55 ipm and my program is about a hour long.  So theres plenty of time for it to screw up..   I'd like to turn up my feed rate and see what happens.  Maybe I'll get brave tomorrow. 

Hood, its set at 25000 Hz

Thanks again for all the input.

[Hood]

Thats fine, I was thinking you maybe had the kernel cranked up more than you needed. Its human nature to see that you can set it to 100KHz and think that must be the best but often it makes things worse. Obviously in your case thats not true so looks like the pulse width was the issue for you.
Hood