esab ultrarex retro fit plasma / flame cutter

I guess that depends on if your torch is high-dev or just an air torch. I'm happy if my position is <0.010 from commanded and in my case that would be due to belt stretch and other items as mine is stepper driven.

axis.N.f-error OUT FLOAT The actual following error

Not much to go on in the manual for that pin, I assume your watching that pin and iirc it is in machine units. When I tuned my CHNC I triggered on velocity and watched commaned vs actual and or following error as I adjusted the PID.

John

the commanded vs actual velocity are exact up to around 400mm/s but for small high speed jogs it shakes the gantry so ive taken it down to 250mm/s i think

so in the image the text at the side of F(0.02400) = 0.26509 does this after the = sign mean 0.26509mm?

and for now the power sources a powermax 85 till we get it burning most of the day and can warrant a fancy new hydef one

also what needs a fine tune to get that spike out is it d or ff2? image missing http://www.linuxcnc.org/media/kunena/attachments/legacy/images/Screenshot_HAL_Oscilloscope-a87fd83e5e775aec86a5747a45fe3fd5.png

From memory I added P in small steps till the axis starts to become unstable then a tiny bit if I then repeat until the following error is small enough. Watch the following error with each change to see if your improved it or made it worse. I also added one of the FF's but forget which one... oh here it is.

Turns out I used a tiny bit of everything.

John

See this is what throws me every body else's parameters are like yours 40-60 on the p and such does it depend on the axis scaling ie encoder counts to axis movement as I cannot get my p above say 1 stably ATM it's around 0.9 with more d than I and I think there is still a bit much p as the scope is still quite course

Thanks for all your help john

afternoon lads just one quick question about this scope calibration method what kid of speeds do we want too be running at to tune the axis?

thanks

PID parameters do depend on scaling but this can be factored out if desired by making the PID input and output commensurate. Since the input to the PID loop is in machine units and velocities are in machine units per second, for velocity mode servos, making th e DAC scale or output scale equal to machine units per second will make the units commensurate (this has the side benefit of making the FF1 term known = ~1 for velocity mode drives)

My guess as to why you have such small P term is that you have mm machine units so the input to the PID loop is larger than a inch machine (and the PID output units are probably somthing that has not be scaled to mm/sec
like 10V)