Mach2/3 question

[planeflier]

Does running a faster cpu computer allow stepper motors to run faster? I have been running my CNC router with a Hobbycnc board on a 950 mhz pentium and have an oppertunity to change the computer to 2.5 ghz machine. By doing this will it allow my steppers to run slightly faster before losing steps compared to the slower machine?

[ger21]

If your losing steps, then a faster machine probably won't help. The only way it might help would be if the faster machine gave a smoother pulse train. But the difference would probably be minimal.

[Salty72]

continuation of this question is .... what pulsewidth is better for step and direction 1uSec or 5uSec

[Salty72]

Anyone ???

[NC Cams]

That depends on how "clean" of a digital signal you generate.

Essentially, digital signals want to be OFF or ON. Trouble is, the electronics has a range of OFFness and ONness that can lead to potential high speed switching problems.

Moreover, it all depends if you have isolation on your break out board. If so, it depends if the isolation IC is 'fast' enough to respond to a 1uS versus 5uS pulsewidth. This switching problem then is further accentuated by any downstream capacitance in your wiring and its ability to change state.

Thus, you computer may be able to send out signals at light speed rates, HOWEVER, your downstream circuitry may not have the response capability to resond.

The optimum pulse width to use is a byproduct of the entire system hardware and software response capability - not a vaporware spec that you dial in to the PC and expect the system to merely respond to at your bidding....

[Salty72]

thanks for that.

Well based on the information presented.....I was wondering if there is anything like a speed test (in much the same way as you would test out a graphics card ) this way we would be able to set the best pauls width for each of our own machines
Salty

[NC Cams]

Think about this: you're trying to drive a highly inductive load - namely a stepper motor. What makes you think that some blinking led based tester is going to accurately represent that sort of loading/driving capability???

Ultimately, you want/need to drive a motor.

Driving a test motor at ultra high speed really doesn't prove much either unless it is representative of the inductive properties of the motors your'e trying to drive.

Just some things to think about while you're trying to come up with a "tester"....

[One of Many]

What is missing in the loop is feedback for a step and direction system. Pushing it to its limit may "over drive" the physics of the mechanical limits.

Without some form of results data to known input, you know not where the limit gets exceeded, nor any way to correct for it unless there is a feedback compensation.

Scrap parts is one form of feedback..........not a positive result, though, but you can back off the input rate until it stops that (s)crap!!


DC

[Salty72]

Originally Posted by NC Cams Think about this: you're trying to drive a highly inductive load - namely a stepper motor. What makes you think that some blinking led based tester is going to accurately represent that sort of loading/driving capability??? Ultimately, you want/need to drive a motor. ....

That’s a little harsh.. Who mentioned driving some blinking light ??? what I was trying to get across it that a video card generates a signal this signal has no F/back but as the user we are able to determine if the signal is too fast too heavy or too crappy etc...

I agreed with your earlier statement .....

"The optimum pulse width to use is a byproduct of the entire system hardware and software response capability - not a vaporware spec that you dial in to the PC and expect the system to merely respond to at your bidding....and that signal is tested for the required output.

AND thus I posed my Question.... is there a test???

Can you knowing the steps per REV, size of pulse, speed of drive and other such information,.....taking all this into account ....then....drive a stepper, at a speed, to a point, and expect a 'known result' if you fail (don't achieve you expected result) change some / all of the variable till you PASS (achieve your expected result) .. If you PASS on the first run repeat the test after changing some of the variables till you FAIL then back it off one notch....

It is obvious to me that if it is possible to stipulate a pulse width to counter-act curtain factors. You must be able to determine when the pulse is wrong...THUS; it must also be possible to "measure" these factors and consequently calculate an "optimum" setting.

It seams I have all the answers, I know, but this is the ramblings of a newb..

[One of Many]

The test would be to couple a rotary or linear encoder system to detect an equal output pulse to the relative input pulses. Lost pulses of the output would be the feedback for various limit in forces for each axis.

DC

[NC Cams]

The point remains - you're trying to drive a HIGHLY INDUCTIVE DEVICE.

If you're going to test for speed of function, you HAVE TO BE ABLE TO DRIVE A CORRESPONDINGLY INDUCTIVE LOAD to see if the circuit will work as you desire.

You merely stated an essentially undefined "tester". Lots of testers merely blink LED's to see if electrical continuity is working. I did not take your question as being self evident that you wanted to run a motor.... only a "tester" as opposed to a "test motor".

If you're testing, you have to drive SOMETHING, right? What would you drive to,

"....PASS on the first run repeat the test after changing some of the variables till you FAIL then back it off one notch...."

Yes, you can measure inductance of a motor or coil but that sort of equipement is not typically found in the hands or tool boxes of most DIY tool boxes.

[ger21]

Originally Posted by Salty72 AND thus I posed my Question.... is there a test???

Most drive manufacturers will tell you what to set the pulse width to. But actual results may vary.