M68e3q0 (velocity 100%)

Is this from the post? I would imagine it is and is this setting the thc to 100% speed before thc gets active?

running a job with 2" X 4" pieces, the voltage went up and thc raised the Z and it didn't cut all the way through.

What causes the voltage to go up? and what do I adjust so thc comes on earlier before it gets up to speed?
 

I think the M68 has to do with arcs.
 

So I am always getting VAD led when cutting holes. Is this normal?

IS there docs on how the velocity anti-dive and void anti-dive works?

Indicates that the THC is locked due to a void being sensed. What is the void being sensed? I disabled it and it seems the voltage is stable.

M68 is described here linuxcnc.org/docs/devel/html/plasma/qtpl...t_velocity-reduction.
The intent is to reduce velocity on holes to improve cut quality. The accepted rule is to reduce to 60% on holes < 32mm 

The VAD (Velocity anti dive) simply monitors current velocity and if it falls below X% (Set by VAD Theshold) of the desired feed rate, it disables THC. By definition it will lock the thc on a hole if hole velocity reduction is used.

Its not wise to disable the THC before it gets to cut velocity becasue the manufacturer's cutting charts are based on a voltage and the torch height at a constant cut rate (feed rate). Change the velocity or the torch height and a different voltage will occur because a plasma cutter is a constant current device. The latest version of QTplasmac uses the THC delay from the point constant velocity is reached (feed rate/cutting speed is reached)

Void Anti dive (sometimes called kerf crossing) is designed to disable the THC when a kerf (void) is crossed. It does this by monitoring the torch voltage and look for a sudden leap in torch voltage that is much higher than that occurring from normal warpage. This increase leads to a violent attempt to adjust the height lower to reduce the voltage (and a  crash). An example where keerf crossing sensing should be used would be to cut a heap of squares or rectangles in a grid where parts exist on both sides of the kerf. This would improve cycle time and material yield. However, most post processors output code where the part is on one side of the kerf and scrap is on the other.

So for normal cutting (eg with sheetcam generated code), its best to disable  Void sensing as its never going to be needed.

The void sensing is also useful when cutting holes in thicker sheet when the slug drops away but if you have reduced the hole cut velocity, this should disable the THC. The other time I think its useful is if you wish to sever the sheet to prevent the torch diving at the end of the cut when the torch runs off the material (which is what you want).

QTplasmac's sensitivity for void crossing antidive is set by the Void threshold.  Void sensing algorithms are difficult to develop becasue it takes time, money and material to do the testing and behaviour alters based on material thickness so what works on 2mm is likely different at 6mm or 20mm. I know Snowgoer has done some work testing qtplasmac in this area. I've looked at it for a long time with a different algorithm to QTplasmac but as I don't have a use case for the feature (eg I leave it disabled),  I have not bothered refining the working version. This is in part due to the difficulty in integrating 3rd party code with Plasmac as is built to serve its own GUI. It requires a lot of pins unlinked, making it difficult to maintain given the constant (almost daily) upgrades.

This stuff is hard!

Thanks Rod,

that explains quite a big deal more.