Re: [CAD_CAM_EDM_DRO]CNC Pulse jitter?

RichD wrote:

> I have seen talk of computer pulse jitter causing axis drive problems
> on stepper driven systems.
> This may be a difficult question but, how much is bad enough to be
> on the threshold of being a problem?
> Is there a rule of thumb or easily perceptible symptom to look for?
> What is the std method/spec to measure jitter?
>
> I'm trying to troubleshoot a mill and apparently I have a bad case of
> resonance but, in looking at the incoming pulse train on the scope
> (Tek 2236) really makes me wonder of other bugs as well, not helping
> the situation.

10% timing jitter at the high speed end will probably cause you to greatly
reduce the max speed. On larger machines, even a few percent jitter
might make the motors lose steps. In other words, if you are trying to
move at 1000 steps/sec, and the jitter is 100 uS, that is 10 % jitter.
To get the jitter down to 5%, you have to slow down to 500 steps/Sec.
So, that has cut your top speed in half!

You really need a timing analyzer to measure worst case jitter, but
newer digital storage oscilloscopes can make this measurement, too.
You can make rough estimates with an analog scope, just reading
thw worst outliers on the pulse, with the machine running at a constant
speed. At a constant speed, set the scope to display 2 step pulses
across the screen, and see how much the leading edge of the 2nd
pulse wanders around.

Jon