UK AC powered 6dof project

Did some more investigations and trouble shooting at the weekend to try and get to the bottom of the jitter issue,

Moved the amc out and away from the cabinet and set up 1 vfd to test and eliminate any issues with EMI from the VFD's, jitter the same so not an emi issue.

Also ordered 1 sample of the more expensive ball bearing pot a week or so back and hooked that up, jitter still the same so not a pot issue. (will be handy to keep as a spare)

I also concluded after trying to capture the problem frequencies better that the mini dso201 oscope i bought is basically rubbish inconsistent readings or no frequency readings at all.

So i decided to get a little more serious and order a Hantek DSO5102p like this one:



Still a hobby model but is mod-able to 200mhz via custom firmware and analysis tools look a hell of a lot more informative and easy to identify problems vs the little DSO.

Aside from that spent allot of time watching youtube videos and reading about LC filters, how capacitors and inductors work and how to use oscilloscopes. This project is becoming
a bit more involved than i though! but i can say im learning allot at least.

 

yes dont really have much choice Michael, with this jitter issue the platform runs horrible and ive checked about every other possibility.

The 0.1uf caps definitely help so this seems to be on the right track just need to find the problem frequencies.

Im also just reading about ground loops in case that is a possible cause.

On the other hand i could be completely off target and maybe it is a PID issue but since im getting jitter even with the pots zeroed by hand before any rotation i think not.
Ill be watching to see how you get on with yours and maybe you will have some advice

Been busy trawling through many webites at the moment to read about noise sources and management.

this is one of the better pages i came across regarding cable types, routing management and termination when using VFD's and motors. Might be of interest to some.

webcache.googleusercontent.com/search?q=cache:http://www.belden.com/docs/upload/vfd_choosing_wp.pdf&gws_rd=cr&ei=HkLMVryYBsmAywPO8K7gCQ
(not well laid out but good info)

one thing ive learnt, SY cable is not emi shielded..! and yep i used SY for cable to motor from vfd!! so looks like im ripping that all out and changing for CY.
Also didn't realise your supposed to terminate the shield to ground at both ends.
fortunately i used CY for the signal leads to amc.
. 
for the pots i used shielded audio cable and used the shield for ground. Only concern there is that is very thin gauge 28-30 so impedance will be higher.
although its only 5v so hopefully this is OK??.
might try a different cable during testing to be sure.

Then i put a 0.1uf ceramic capacitor across.



As you can see big reduction in the noise level. and this makes a significant difference to the jitter although does not eliminate it completely.
So i need some other capacitors?

Cuts  noise/frequency from about 1.15mhz to about 500khz

Im assuming this line should ideally be close to flat when there is no directional output/input from vfd.

That looks great with the 0.1uF cap. The remaining noise should be coming from the sensor side as they are in the same loop. What you can do is to add an 0.1uF cap on the sensor leads too, between Vout and GND. If not much difference, add another 0.1uF cap on the Amc1280usb side on the terminal connector.

Filtering is tricky business. If I only could filter all noise on the board I would, but the control loop extends to devices outside the board as well (wiring, PSU, VFD, sensors, etc).


It will never be flat line, but the smaller the noise Pk-Pk the smaller the impact it will have to the control loop.



Thanks
Thanos

Try add 0.1uF caps on the analog inputs like this: