Finally got to move my working PC setup from the old table to the new-build one today, seems ok apart from one odd issue.

The settings were copied from the test pc i was using to debug the table build, not transferred, i just manually copied each setup screen from one to the other.

The Drives are digital DSP drives from CNC4You.UK and are new, the BOB is part of an MP3000-DTHC2 setup from CandCNC, this was taken from the old table.

The oddity is this - when testing the jogging I find Z and X axis work perfectly but the slaved Y&A axis sounds and runs very notchy and rough. My concern is that all drives are the same, all motors are the same and therefore i set all "Step Active Low" settings the same.

However, upon changing the step active low settings for Y&A seems to fix the issue.

But, I cannot see why these two axes need opposite step signal levels???

Any ideas here?

Can't edit now but..

One difference is that the test PC was underpowered and only ran at 25khz, the work PC runs at 45khz, step and dir pulse length is 2us on both PC's

From a quick scan of the data sheets I don't see that they actually say whether the drives step on the rising or falling edge so there's no immediate telling which is correct. The fact that your Y/A axis run better when you flip the setting would suggest that they are now set correctly and you're just "lucky" that X/Z work with the incorrect setting. What happens to X/Z if you change the settings to match?

Re: your 45KHz kernel. You should set the kernel to the MINIMUM required for your steps/per & max-velocity regardless of the "power" of the PC. There is zero advantage and potential disadvantage to using a higher kernel speed than needed.

thanks Stirling,

i have sorted it now i think, CandCNC advised me to flip the settings back and up the pulse width to 5us this seems to have done it.

I will ask them about the clock rate as it was their THC upgrade that switched it, i originally had it on 25 but after the upgrade it was running 45 no idea why, will ask.

Dave

Hmmmm OK, but a drive that claims a max pulse rate of 200Kpps can not require a pulse width as long as 5us so one of the two is suspect. (You can't have 200Kpps each 5us long).

Just noticed that on MYCNCUK you stated that all was well with a different BOB. I think there's your answer. It sounds as though your current BOB (not the the drive) is the thing that needs the longer pulse width. This usually down to a BOB that uses optos that aren't as fast as they could/should be.

It also raises the issue of driving optos (in the drive) with optos (in the BOB). Not generally a good idea.

Interesting point there, its a well known setup though, might have a chat with them.

It does seem to have fixed the issue though which is the important part.

I have also asked them about why they set the speed to 45k when i upgraded my THC to their system.

It does seem to have fixed the issue though which is the important part.

True, but then so did your original scheme apparently.

However, upon changing the step active low settings for Y&A seems to fix the issue.

But, I cannot see why these two axes need opposite step signal levels???

Any ideas here?

I've simply tried to answer the WHY bit.

1 - Yes, the original scheme worked, that was with different drive boards and the faster pc.

2 - This scheme worked with the new drives and a slower pc and cheaper BOB.

3 - This scheme failed with the new drives, a faster pc and the same BOB as 1.




Thanks for the help though, much appreciated as always.

If you really want to get to the bottom of it I'd scope it all but that's just me. I personally would want rid of the opto chain though. Anyway - have fun.