Recent PlasmaC Update and No Motion

A NAN error or having to reboot to recover from a fault suggests a
hardware problem to me, perhaps on the host PC

Phillc54, at this point the only reason I've changed them is to just try something to get the machine running again. I messed with low pass initially because it helped with losing arc when I used mode 0. I use longer lead outs and it would dive. I never messed with debounce before this.

As for "nan", this was never an issue before the update. I had never even seen it before and didn't know that could read anything other than a number to be honest.

Again, this was never an issue before updating to a new plasmac version. Admittedly it had been a while. But everything worked great before.

Tesremos also appears to have the same nan issue on his machine several time zones over.

PurduePete wrote:

Phillc54, at this point the only reason I've changed them is to just try something to get the machine running again. I messed with low pass initially because it helped with losing arc when I used mode 0. I use longer lead outs and it would dive. I never messed with debounce before this.

Setting lowpass to 2 would hide a lot of issues. Are you sure you are cutting at the correct height, have you done several Probe Tests to confirm the probing is set up correctly. Void Sense should limit diving at the end.

As for "nan", this was never an issue before the update. I had never even seen it before and didn't know that could read anything other than a number to be honest.

There has been nothing changed to the reading of the arc voltage section of the component or GUI. As I said, I have never heard of it before.

I'll admit noise could be a problem but it does it in dry run too. Typically the plasma is the main issue when it comes to noise. I've been running this machine since March with little to no issues other than not being able to get hypersense to work exactly right. It would probe until wet and no amount of number tweaking seemed to do it. That was actually the reason for the update. Ohmic3 and newest update. Nothing changed physically either. I am baffled and not real sure this is on the update still. But have no idea what to do.

What process did you follow to update your machine?

I have been reading through the thread again to try to see if I missed anything and I am a little confused by the following two posts, I know it is a bit frustrating but detailos are important:

PurduePete wrote:

Phillc54, I actually thought it was all good. If I run a program it runs fine with the torch off. No issues. If stopped in the middle there are no problems. If it completes no issues starting another. But when a cut program is run with the torch on only that program will complete. Runs great the whole time for the most part, I'll circle back to this, but after that program it will only do the first cut and then will do what it has been doing not moving after pierce.

PurduePete wrote:

I'll admit noise could be a problem but it does it in dry run too.

OK, a bit (lot) more testing does show up the error in multicut jobs, before I was only testing on a single cut.

It seems plasmac has issues with certain Z axis setups. I seem to recall Tom mentioning it quite while ago but I could never emulate it. With your config I happens quite regularly.

Can you try changing the foolowing in the ini file:
In your Config Panel change Probe Height to 1.0

All the distances will still be the same as far as the machine is concerned.

After you home the machine for the first time jog Z down to the bottom and do a touchoff then rehome Z and see how it goes.

In the meantime I will work on a proper solution.

Yep, I can see the confusion there. And I cant quite figure it out myself. It's made it a challenge to communicate what I'm seeing. I ran a 50 pierce cut program and it was fine when I posted that, but havent done it since (havent tried). Ran a few parts after that in dry cycle and then posted that because I was so stoked to have it running. But then I needed some parts for a bench I'm making and couldnt get it to go past the second pierce again. So more dry cycle testing using conversational programming. A single part can be run multiple times by itself (torch off so no noise). But if I array it, it fails second pierce (still dry cycle). Couldnt find anything in the gcode to cause this and havent tested any drawn files in the last day. All the parts I've needed are tube caps, concrete anchor feet, and supports.

Phillc54, I am spending time today writing out exactly what it does. So I will document it more clearly.

- Purged linuxcnc 2.9 and any ohmic.comp file last night and loaded in 2.8. No change to what has happened in the past and is detailed below more clearly.
- Added z homing switches just in case the z axis offset issue was still part of the problem. No change to what has happened in the past.
- Started the machine for the first time from a zero power state. Jogged machine around to make sure full motion and checked homing of switches. Shut linuxcnc interface off and restarted only the interface. Ran a program for 6 bench supports in dry cycle 3 times with no issues. 6 pierce points. Ran it with no issues with torch on to completion. Loaded a new cut program for tube caps and ran in dry cycle. Stopped on second pierce.
- Restarted only Linuxcnc interface, not the computer. Went to cut 4 thicker tube caps, only a total of 4 pierce points. Stopped after the second pierce in the program. No jog function after killing the program and in manual job.
- Restarted the entire computer. Loaded the same program. Ran with torch disabled for the path of the first part (since it had already cut). When it get to the probe point of the second part I renabled the torch, and it attempted the next cycle. However it could not establish arc. So I paused the program, let it retract and then killed the program. Found a piece of 16 ga steel to set on top of the 3/8" plate at the pierce point, used start from line, selected the m5 command of the first part and started the cycle again. It ran the whole program to completion. I then restarted the program with the torch off and it ran to the 2nd part just fine, so I killed it in the middle of the fake cut. Restarted and let it run to completion and it ran just fine with the torch off. Loaded the first program in a folder full of old previously run programs and it stopped on the second pierce. No jog function after killing the program

I have done a poor job of explaining but this is exactly what I have been experiencing since the beginning AFTER you made the new update to the offsets. Before v.166(?) it didn't go past the second pierce when I updated from v.142(? close at minimum). That is probably because it took some time to figure out just how to make it do something different. So initially when it seemed fixed I ran a cut file that I'd been trying to run and it ran past the 2nd pierce, oh happy days, but then the next file didn't cut. Very confusing road for me to figure out the entire computer needed to be shut down.