Brother TC-225 / TC-229 adventure!

well, its on its way - I ordered 4. now I can twiddle my thumbs for a few days.

actually, I guess I can try to read some of the inputs and see where they end up..

ihavenofish wrote: haha, $20 shipping. sigh

I have a tube of them, but that's no help for you.

You could have tried Amazon or eBay:
www.amazon.com/dp/B00CL7X2IM/

it was $8 in the end as i went to digikey. i prefer them and mouser because the parts arrive in 24-48 hours, as opposed to weeks usually from ebay/amazon etc.

so all good i think.

transistors delivered, have to wait til tonight to try em

transistors installed into cable and.... click-clock, works. drives are powered by the machine on button in linuxcnc.

on to the drives. it turns out that the wiring drawing for the drives was....

correct! x axis running. I did some tuning according to a guide I found on this forum, and its pretty decent actually. I set it to 787 ipm max speed (original speed) and 0.5G acceleration and it tracks kinda ok. I have some questions to see if I can make it better.

first, I am assuming the "1m" scope setting means 0.001 units, which is 0.001" on this machine.

if that's correct, then at all speeds my axis tracks within about 1-1.2 divisions. most of the following error is of course during acceleration. travel is very smooth at most speeds, its a little bumpy "sounding" at the lower end of things (1-2ipm), but following error isn't very big. at constant speed, it tracks very close to 0.

my pid settings turns out to be

P = 125
FF1 = 0.675
FF2 = 0.001
dead band = 0.00001

everything else is 0.

my encoders are 4096 count (1024 line), 5 pitch screws, 4000rpm.

so what would the suggestion be to tighten this up? or is it pretty good as it is?

I also should not, I think 0.5G might be higher than it was stock, OR, linuxcnc's lack of jerk control is making it feel way more aggressive. at low speeds you get a bump start to each move as its basically instant acceleration, whereas the original control seems much smoother.

thanks for all the help so far!

You should be able to pull the steady-state error down with a bit of I gain. Your results sound rather good for a P-only control.