Random arc signal lost

snowgoer540 wrote:

Oh, I thought you could see them as a Moderator.

Possibly but I wouldn't know how.

phillc54 wrote:

snowgoer540 wrote:

Oh, I thought you could see them as a Moderator.

Possibly but I wouldn't know how.

No worries, incoming email.

AgentWD40 wrote:

So I'm running a 7i96 and a hypertherm powermax 45xp and am having arc lost issues. The torch fires each time but sometimes linuxcnc throws the error. I'm using machine's arcok.

There's been a lot of back and forth in this thread. Do I software denounce or use resistors???

I'm using an older build of plasmaC, the computer doesn't have internet so updating is difficult. I can follow up with my version if needed.

If I should try software denounce can someone specify exactly the hal changes required? I tried clicking those prelim docs a few pages back but the links were no good.

First, I would encourage you to update... a lot has changed, and a lot of important bugs got squashed. QtPlasmaC is pretty stable. Something to consider. If nothing else, plasmac v.191 (assuming you are on 2.9) had a lot of the aforementioned bug fixes. If you're on 2.8 it might be time to move up.

On to debounce. Personally I would start with the software debounce. I think that for everyone who has had an issue and has a powermax, the software debounce was enough to cure it. That said, if there was still issue afterward, I would use the ceramic capacitor method as opposed to the resistor method. But more than one way to skin the cat.

I am just going to list the lines you need to add/change, you will have to figure out how to get them in there. Otherwise you could post your <machine_name>_connections.hal file and I could help that way.

Obviously change your input to suite. That should get you there.

I've seen where one user reported that 5mS wasnt enough and bumped it to 10mS and hasnt had issue since. So you may have to experiment a bit. 5mS works fine for me. I'm seriously doubt the 5 extra mS makes any noticeable difference in cut quality.

I should also mention that the delay is actually measured in servo-thread counts. But assuming you use the standard 1000000nS period, that's 1mS. Make sense?

If debounce doesn't fix it I would use a capacitor across the contacts as suggested by PCW.

forum.linuxcnc.org/plasmac/40819-arc-ok-...fire?start=50#191678

AgentWD40 wrote:

Do these last two lines need to be placed in any particular order within the hal file?

net plasmac:arc-ok-in hm2_5i25.0.7i76.0.0.input-06 => debounce.1.0.in
net plasmac:debounced-arc-ok debounce.1.0.out => plasmac.arc-ok-in

From a machine standpoint, I dont think it matters.

I just do it that way because it makes it easier for me to read later and understand WTF I was doing.