Hello Guest it is March 28, 2024, 09:25:16 AM

Author Topic: Motor tuning calculations  (Read 21479 times)

0 Members and 1 Guest are viewing this topic.

Re: Motor tuning calculations
« Reply #30 on: April 24, 2009, 10:51:10 PM »
Whew!  I can breathe again.

It sounds like you found that information a little useful.

I almost didn't post... I'm a noob... I know it... and I'm not looking to start posting lots of useless noob advice.

I find the problem you describe very interesting.

It is very strange that it worked for so long, and then seemed to change when you tweaked some settings.

It would seem to me, that it's one of two things:

1.  Either the Step signal coming out of your LPT port (if you are using an LPT port) or out of your motion controller, has changed.  Meaning it's not putting out the same number of steps for a given distance as it was previously.

Or...

2.  The number of step pulses coming out are correct... but somehow the stepper drivers are no longer operating in the same mode (Full, Wave, Half, some level of micro-stepping).

I'm an electronics guy... so my next move would be to hook up either a counter or a logic analyzer to my STEP signal for the problem axis... tell Mach3 to do a move that would generate a specific number of steps... and see how many pulses I actually got.

At that point... I would know something... as I would actually see the expected number of pulses... or not.

If the number of pulses were right... I'd look for what changed downstream... probably the mode that the stepper driver is running in (some can be changed).

If the number of pulses were wrong... it's got to be something in Mach3... almost certainly some configuration setting.

edit --- If you are using an external motion controller... it could also be that, as that's what would be generating the step pulses.

My take on this would be that the Mach3 software has worked for too many people for too long to have that fatal of a flaw... where it's flaky and only works sometimes... it's got to be something more definitive.

Your own attempts to solve the problem by re-installing and using different P.C.s seems to indicate that whatever it's doing... the result is very consistent... it's just not resulting in what you expect.

I wish you the best of luck.

Best regards,

Michael
« Last Edit: April 24, 2009, 10:54:34 PM by techsol »
Re: Motor tuning calculations
« Reply #31 on: April 24, 2009, 10:58:53 PM »
Tommorrow I am going to run TurboCNC from dos ... to see if the results repeat or are corrected ... if they remain then most definatly the controller and drives ... if not then it is the way the Mach3 is sending signal and they way the controller and drivers are interpetating the signals.

for what is is worth ....

www.soc-machines.com ... Mark4 and MM220's ... in the download section for the pdf's.
Re: Motor tuning calculations
« Reply #32 on: April 24, 2009, 11:57:55 PM »
Gawd... I typed up this lengthy reply... and then something went wrong with the posting and it got lost.

Ouch.

Sorry man, but I'm not up for typing it all again.

Here's the summary.

Cool stuff on that Soc-machines site, thanks for the link.

The docs on the MM220s pretty clearly indicate that they were shipping them configured for 1/2 step, but that they can be reconfigured for full, 1/4, and other modes.

With the earlier MM220 software... you had to download a new hex file to the boards to change the stepping mode.

I think you would probably know if you had done that, grin.

However... I also spent some time looking at the docs for the MM130s.

The MM130s have this interesting feature where they can be reconfigured via the same STEP and DIR pins normally used for motion.

You change the STEP and DIR signals in a particular way... and instead of generating motion... you send commands to the stepper drivers processor.

One of the things you can do with those commnads... is change the stepping mode.

This is possibly relevant... as one of the MM220 documents mentions a new release of the MM220 software which will give the MM220's features similar to the MM130s.

It doesn't elaborate on excatly which features... but hey... it's not a stretch to think that maybe "configuration commands over the step/dir interface" is one of them.

If your MM220 drivers have updated software, and now have that feature... then it suddenly becomes at least possible that the step/dir signaling coming out of Mach3 could... depending on your configuration... result in a configuration change of your stepper drivers.

A lot of speculation here... sorry for that... but I'm working with very little real knowledge of the stuff you are using... I'm just reading tech. documents and looking for clues that might help.

Best regards,

Michael
Re: Motor tuning calculations
« Reply #33 on: April 25, 2009, 01:01:57 AM »
Michael,

I have also written lengthy replays several times and lost them when posting on various forums. Now I always copy what I have written just before I post it. Has saved me from rewriting a long reply many times :-)
Re: Motor tuning calculations
« Reply #34 on: April 25, 2009, 03:00:48 AM »
Well that doesn't explain what I am seeing. My electronics don't have any provision for anything except full step.
Yet two of my axes are now running set at 2000 steps per inch and the third is set at the normal 1000 steps per inch.
There has to be a bug in Mach 3. I am going to get my O scope on it maybe tomorrow.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Motor tuning calculations
« Reply #35 on: April 25, 2009, 05:29:10 AM »
Hood,
Did your BOSS 6 have 3 SMD cards and an ACC card mounted on the right hand side of the main card cage cabinet? If so it was definately silgle step.

I honestly cant remember as I stripped and sold the boards and motors almost as soon as I got the BOSS. I did have pics of all the boards I had when I put them on eBay and will look to see if I still have them.
 You are probably right regarding the full step, it was 4 years ago since I read the pdf regarding the differences between USA and UK Bridgeports and probably it was the pulley difference that was making the steps per unit difference rather than the full or half stepping.

Hood

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Motor tuning calculations
« Reply #36 on: April 25, 2009, 06:59:54 AM »
Heres pics of the boards I had.
Hood

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Motor tuning calculations
« Reply #37 on: April 25, 2009, 07:01:31 AM »
And these

Offline bowber

*
  •  216 216
  • Kirkby Stephen,Cumbria, UK
    • View Profile
Re: Motor tuning calculations
« Reply #38 on: April 25, 2009, 08:00:57 AM »
Is it anything to do with the pulse length/duration?

I've never had anything like this at all, the only problem i had when first setting up was I assumed the lead screws were imperial so measured 5 tpi but they were actually metric and were 5mm pitch, very close but not quite right.
Doesn't help with your problem though, I curious that it's so consistant as well which indicates some setting somewere.

Have you checked the axis scaling?

Steve
Re: Motor tuning calculations
« Reply #39 on: April 25, 2009, 11:28:35 AM »
Newest update ...

I did not test TurboCNC (could not fing a darn floppy that would work LOL)  so I tested both KCAM ( http://www.kellyware.com/ ) windows driven and also Linux EMC both are doubling the distances as does Mach3 .... so now I am off to call SOC again to find out why this is happening.  I would have to say it would be with in the controller but this is specualtion at this point.

Machinemaster maybe we could get our systems drunk and let them mate take the offspring and we would be both set .. LOL