THC Servo Start

islander261 wrote:

Guys

As promised here is my current kerf crossing component.

John

John, nice simple easy to read code. One thing I picked up on. If the void is wider than your kerf, and you reenable the THC, you could still get an overcorrection. I think at that point, you should wait until the torch voltage is < the value at which the lock out was enabled or near the torch set point (eg somewhere within the normal hysteresis observed in the torch voltage around the cut voltage.. I'm not yet sure which one. I am going to log some data for a cut that crosses a kerf and then goes off the edge of the sheet when I get a chance. A halscope plot should give clarity around the torch behaviour when kerf crossing. But I wasted far to much time on this today, I did not get to it. Soon!

Rod

Thanks for the tip, I can certainly sample and hold the starting value as I do in my corner lock component. As I have mentioned before kerf crossing is a low priority for me. I only use full sheets and set my limits to prevent going off them. I also think that using a velocity driven corner lock (or VAD) with short leadouts will have the torch height locked when completing closed paths.

I have found a much bigger problem now and am struggling with how to communicate it to Les at SheetCAM. I found that the source of all the short spurious segments is when you select an outside offset for the cutting path (most likely the same for an inside offset cutting path). When I select no offset for the path I get a nice path in G code with no short spurious segments. You will only see this problem in pieces with lots of interconnected curves as the math for rectilinear pieces is simple.

John