Hi,
about posting the documentation for your Chinese THC unit. Who can say what inputs it requires or what outputs it offers?

Craig

IIRC anti-dive is an internal Mach routine, not a controllable input.

Hi weldsmith,
I profess no expertise on THC but I have, particularly through work, come across them.

The one I'm most familiar with is the PROMA unit. In essence you hook it to the plasma output and it produces a signal THC_UP or THC_DOWN in order that the plasma voltage
follow that you programmed into the PROMA. The UP-DOWN signal is presented to your controller and your controller sets the Z axis direction pin to up and then pulses the Z axis step pin
to cause the torch to lift. Should the torch go too high the plasma voltage will go above your setting and the PROMA unit will issue a DOWN signal and your controller will respond
by setting the Z axis direction pin to down and pulse the step pin. Sorry to bore you with what you probably already know.

Anti-dive is a software enacted modification of the basic up-down procedure above. To my knowledge the only way you can affect it is by those variables that the developer of the controller routine
makes available to Mach.

You could after a fashion write your own THC program but the delay between when your THC sense unit produced an UP/DOWN signal until the stepper started to respond would be about 100ms
for Mach3. This is probably too slow to be effective. It results from the fact that Mach3 or Mach4 come to that it is not a realtime program, it provides data in a buffer to a controller, be it a parallel port
or an external controller like an ESS, which is realtime. Hence the reason that the THC unit talks direct to the controller and bypasses Mach because Mach is too slow.
I have considered doing something similar in Mach4, as it turns out Mach4 is much quicker to respond, somewhere in the region of 10-20ms. Its still fairly lethargic but getting closer.
Apparently NFS are considering writing a module that does just that using a 'piggy back' axis, all very clever. It would still require controller support but a much simpler proposition than full bown
realtime THC.

The last possibility if you wish to write your own THC routine is to use one of the hardware controllers that allow/promote the ability to write code directly on the controller. The PoKeys 57 series,
Vital Systems HiCon unit, with appropriate activation and Gallil are all capable of such programming. Gallil is by far and away the most accomplished at this sort of board level programming
but a three axis Gallil is going to cost around $2000.

You might also have a look at the THC kit put together by Vital Systems. It is written for Mach4 but it specifically includes features related to anti-dive.

Craig

In a nutshell, unless you are running a slow machine, Proma sucks. I tried it and it worked ok for a tiny machine built with a hypertherm 30XP system but when i built my bigger table and got the 45 system, it was far too slow at processing to be any good at all. If the sheets warped etc it just failed miserably.

In the end i forked out some money on a CandCNC MP3000 THC which was much faster but still not fast enough for bad warps etc, I still run this box today though, just have to learn its problems.

The fastest systems take direct control of the Z axis and run the stepper themselves, its the only way to cut the loop process times down to a minimum I think.

I still suffer the odd crash and dive but thats life with a plasma i think.

Thanks Dave that is good information to hear. Even though I have a stepper motor on My Z axis I just purchased a floating head with a z lifter. hopefully this will be quick enough.

 I am curious if the new Proma SD will work faster. This unit does not talk to mach it is completely stand alone.

Could be better, the one i had was about 5 years ago now.