[AlainB]

Hi,

What is your opinion about this signal?

The upper one is from a parallel port, from TurboCNC 4.1 (a DOS version) at a velocity of 1200 hZ with a pulse width of 200 micro seconds. The pulse is the high side (+5V) of the square wave.

The lower signal is the same signal fed twice into a 74HC14N chip (an Hex inverting Schmitt trigger); fed twice to keep it non inverted and from another 5V supply.

The blur is from my camera, the signal is crisp.

What concern me is the jitter that you can see in both the signals.

It is about the first time that I use an oscilloscope to really study a signal and I am wandering if this is a normal behavior.

[ame="http://www.youtube.com/watch?v=dl3ocmfS6Ds"]YouTube- 100_2033.MOV[/ame]

Thanks!

Alain

[irving2008]

Very normal in pulses generated by software running on a complex machine... the jitter is the result of non-deterministic interrupts to the software. Despite turboCNC running on DOS (IIRC) there are still many interrupts - timers, keyboard, etc. going on in the background.

[AlainB]

It makes sense, thanks! The computer used was an old 486 laptop with a math co-processor.

This morning, I noticed some infos printed on the opening screen of TurboCNC concerning jitter:

Mean 8254 timer latency: 13.65 microseconds, max is 512.00, min. is 13.00
Frequency where jitter approaches 10% is: 8738.7 Hz

This mean. I suppose, that a certain amount of jitter is acceptable, 10% being the very high limit.

[RomanLini]

Don't run it from a dos window, reboot the machine in dos. There is an enormous difference in performance.