SOLVED Why does a Mesa-5i25(6i25)-card system lose steps? [A. Settings.]

  • clunc
  • clunc's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
More
14 Sep 2022 01:46 - 20 Sep 2022 18:16 #251863 by clunc
I'm the "hero" of another thread, where I try to "copy" a software-stepper system over to a Mesa-6i25-based system.  It worked in the end, but "oh-the-humanity."

I thought it was all done and working, and so started a long complicated model.  After only a few minutes, I could tell from the other room that something was amiss. When I checked, I found the mill cutting air several inches away from the stock.  Luckily--for once--the 1mm bit had not, yet, encountered anything immovable.

I do remember reading that "latency still matters" for Mesa-based systems, but I can't find where I read it.  I am reluctant in the extreme to fire up PncConf to run a test and risk writing over my hard-sweated-out custom .hal and .ini files. (I realize I can tar up the configdir, and will.)

I wanted to ask first WHY latency matters when a Mesa card is handling the step-generation.

Suggestions for changes to make and a methodology for testing them will also be greatly appreciated.

My ini and hal are attached.

And thank you.
Attachments:
Last edit: 20 Sep 2022 18:16 by clunc.

Please Log in or Create an account to join the conversation.

More
14 Sep 2022 02:15 #251865 by PCW
If you are losing steps, its most likely a timing or electrical issue

As a first debugging step, I would double the step times.
You can often get by with shorter step-times with parallel port hardware
because the access time of the hardware stretches the pulse width for
short pulses, while the Mesa hardware gives you just what you programmed.

Also you can lose steps if your accelerations and velocities are too high

Latency will not cause lost steps with Mesa hardware

Please Log in or Create an account to join the conversation.

  • clunc
  • clunc's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
More
14 Sep 2022 09:10 #251879 by clunc

Latency will not cause lost steps with Mesa hardware.
 

That's probably worth tattooing backwards on my forehead because that was why I bought a 6i25. 

I will double the step times, and accelerations are a likely suspect; I dialed the max velocity down in AXIS until I could verify performance.

And I found a testing protocol from @tommylight to test. 

Thank you.

Please Log in or Create an account to join the conversation.

  • clunc
  • clunc's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
More
14 Sep 2022 09:42 #251880 by clunc
By the way, I'm embarrassed by my tortuosly slow pace, in following advice and reporting back, for which I feel the need to apologize.

I would like to share some of the other "projects" that I have had to tear myself away from LinuxCNC to address.

(If you live long enough you'll get to a point where even your old tools that you need to fix all your old stuff that is breaking are breaking.)

Thank you.
The following user(s) said Thank You: tommylight

Please Log in or Create an account to join the conversation.

More
14 Sep 2022 14:15 #251900 by PCW
No need to apologize, this is what the forum is for.

What type of jam/sauces are those?

When we get enough, we like making Loquat jam here.
The following user(s) said Thank You: tommylight

Please Log in or Create an account to join the conversation.

  • clunc
  • clunc's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
More
14 Sep 2022 22:25 - 15 Sep 2022 10:59 #251943 by clunc
Peter,
First things first.
My CNC machine is functioning "normally" again. I'm not sure which did it, for sure, because I simultaneously HALVED my STEPGEN_MAXACCEL to 75 and DOUBLED my STEPLEN to 10000. My, wooden carving, projects don't require feeds above, maybe, 180ipm which is to say, I'm probably not a candidate for tweaking performance.

After a test run of 50 laps around an 8x10.5 rectangle (fans of 8.5x11 paper will recognize the significance) found the origin spot-on, I once again rashly launched the "finishing" pass of the model I'd cut the rough pass on, roughly, 19 months ago.  It happens that it is the "450th" model I've made on this CNC machine.

Second things second: those quarts are "grape juice concentrate." We've long forgotten the varieties we planted, so "anything that looks green" goes into green concentrate, and similarly reds. We remain confused about the blush grapes... However before these, latest, products I'd picked and canned over 120 pints of wild black and red raspberries, blackberries and gooseberries (exactly one of the latter...). It has been a Banner Year in Wisconsin; I kept picking Precisely because I don't believe I'll see the like again in my remaining years.

Now, I'll say again...
I simply could not have done it (getting the CNC back on-line) without the rock-solid support of all of you here at forum.linuxcnc.org. God bless you.

Oh, and that reminds me to post the result of this successful last/first run (I decided to keep the mung-edness of the bottom to remind me of the Long Walk to get-back-up-to-speed.  The rest of you take note about Preserving Your Setup against Future Unknowns.)

Thank you, thank you.
--David/CluNC
Attachments:
Last edit: 15 Sep 2022 10:59 by clunc. Reason: clarify
The following user(s) said Thank You: tommylight

Please Log in or Create an account to join the conversation.

Time to create page: 0.061 seconds
Powered by Kunena Forum