Rods "Spaceship" Scratch built Plasma Cutter build

rodw wrote: So this got me thinking about redirecting the THCAD frequency output to an Arduino to do the signal processing before sending the processed minimum voltage to LCNC.

If you can write Arduino code to do the job then you can write a custom HAL component to do the same job.

andypugh wrote:

rodw wrote: So this got me thinking about redirecting the THCAD frequency output to an Arduino to do the signal processing before sending the processed minimum voltage to LCNC.

If you can write Arduino code to do the job then you can write a custom HAL component to do the same job.

Andy, I absolutely agree and that would be my preferred option. I thought the servo thread created in the line below ran 1000 times per second (or every millisecond)

I need a thread that runs at 40 usec and my base jitter according to latency-test on a 25 usec timing is sitting on about 140,000. Is the hardware good enough is really the first question.

Cool, latency-histogram lets you define a thread time period. First time I've used it.

image missing http://i210.photobucket.com/albums/bb270/rodweb/Lathe/Plasma/plasma-16Jan2017-447_zpsqyo4c1ok.png

Maybe that will help...

Note that you cannot use the THCAD for fast conversions (if you sample at 40 KHz you will get only about 5 bits of resolution)

Its conceivable to use a Arduinos A-D and capture the minimum value, though I do suspect you would need better torch power to make this of any use as the combination of ripple, hash from the plasmas switching power supply and probalems getting enough common mode rejection with the required input circuitry would mean you get a very high noise to signal ratio

(I suspect you have the fine sandpaper out when a chain saw is in order)

Do not take my word for it, but that does not look very good. Usable, most probably.
I gave my brother a Fujitsu Siemens Celsius workstation that makes some nice narrow spikes in the rangege of under 3000, now i have to find a way to take it back nicely. It is an 8 Xeon processor, 16GB ram, 2 RAID contoller monster, although several years old.
Did you get the laser cut parts?

PCW wrote: Note that you cannot use the THCAD for fast conversions (if you sample at 40 KHz you will get only about 5 bits of resolution)

Its conceivable to use a Arduinos A-D and capture the minimum value, though I do suspect you would need better torch power to make this of any use as the combination of ripple, hash from the plasmas switching power supply and probalems getting enough common mode rejection with the required input circuitry would mean you get a very high noise to signal ratio

(I suspect you have the fine sandpaper out when a chain saw is in order)

Peter, Thanks. The other thing I didn't understand was the recommendation run the THCAD at 1/32 for LCNC.

tommylight wrote: Do not take my word for it, but that does not look very good. Usable, most probably.
I gave my brother a Fujitsu Siemens Celsius workstation that makes some nice narrow spikes in the rangege of under 3000, now i have to find a way to take it back nicely. It is an 8 Xeon processor, 16GB ram, 2 RAID contoller monster, although several years old.
Did you get the laser cut parts?

Thats what I thought hence suggesting an Arduino frequency counter to read THCAD and then output a processed frequency signal using another interrupt service routine to THC so LCNC would not even know the Arduino was there... But 5 bits of resolution is a bit of a worry.

The laser is still coming.... Very impatient here...