Spurious engagement of Plasmac's Anti-dive

Look inside the inverter for a dip switch.. I hope you have luck today !

No problem. In your attached files, i cannot find this hal file : esab-plasmac.hal
It could be the error lives in that file. So that's why i mentioned the missing file...
It's loaded in your hal config. When it's not loaded, you get a message btw.

It looks like the kerf detection disables the THC for some period of time (kerf width?) so once you got the first kerf detection event, the THC is never enabled again or the control duty cycle is is low enough that the control loop gain is insufficient to stay within the threshold and therefore does not track the slow change in voltage.

Have you tried simply increasing the kerf crossing threshold to say 3V?

I wonder if a threshold that's more independent of feedback loop bandwidth
would work better (like DVDT of filtered voltage > some threshold) plus some
smarter recovery scheme like detecting the other edge of the kerf
(with DVDT < some other threshold)

It would be nice to see a plot of plasma voltage at an actual kerf crossing...

if(arc_voltage < target_volts + (kerf_threshold * 0.5)){
                                    kerfcross_is_locked = FALSE;
                                    last_arc_voltage = arc_voltage;
}

thefabricator03 wrote:

This might be a silly question but what is the purpose of kerf crossing?

I only run simple parts mostly but have not really had the need to cross over a kerf I have just cut,

Could I get a example of when it would be used.

rodw wrote:

I felt the same but I have found that if you are cutting a hole in thicker material, you can get a dive at the end of the cut when te slug falls away. It would also let you do an overburn so the arc on the bottom side can finish the hole cut. I think it would also be useful when severing a sheet.

I have heard some people actually cut rectangular parts without wasting any material between them. Eg cut them out as if you were cutting out paper pieces with a ruler and knife.

thefabricator03 wrote:

The way I overcame the first problem is turn the THC control off before the end of the cut, That way it maintains the same height it was at just before the end and will continue on without diving as it reaches the keft.