Bumpy Motion in SMC3

The only thing holding me back right now is the motion of my motors are "rough" or "bumpy" feeling when the motion should be smooth. This happens whether I'm sitting on the chair or not. I've been tweaking settings while running sine motion in smc3 and I've been able to improve it, but going too far creates other problems which makes me think I'm not looking in the right place.

Decreasing Kp improves it, but if I go below around 250, then the large scale motion starts to fail.
Decreasing PWMmin improves it, but same thing starts to happen.... large scale motion falls apart below around 15.
Kd/Ks don't seem to have an effect.

I'm pretty confident my pots are rock solid. If anything there might be some play at the bottom and top, but there shouldn't be an variation mid-stroke. I'm not sure what else to try. The SMC3 screenshot is where I'm at now and it's the best I can get it. Hopefully someone will have some ideas of things to try or check.

24v, 250w, 75rpm wheelchair motors
IBT-2's
360° hall pots
2x 24v, 600W PSUs

Thanks in advance.

set pwm min to zero
set pwmmax and maxrev to 255

reduce clip and limits to much smaller numbers

show us a video of the problem

PWMmin to zero does smooth out the motion on the way up and down significantly, but then the levers get delayed when switching direction. They stall a bit at the bottom and top. Notice the flat spots after the peaks and troughs when pwmmin is 0. 30 smooths those out completely. I think 10 to 15 is a good balance between the smoothness of the straight line motion and the stop-to-start motion. In flight with the butt kicker and all the axes active, it's imperceptible.

This makes it worse. I just found that if I lower Kp to 150, it's quite smooth, but then it has trouble staying on target. (attached)

I'm wondering if it's a problem with the PSU or the IBT-2's? Both axes are behaving with the same sensation... I wonder if the PSU's are putting out enough power? If I put SMC3 output to square mode or shake the rig quickly with manual mode, the PSU's actually sound like they lose power during the violent motions.. the fan sounds like it turns off for a moment during the motion.

Hey Tim, how did you manage with the noise problem? I'm in the same boat for months and nothing I'm trying to solve the problem worked.

Thanks!

I have not. I haven't been flying much actually since I completed the chair. It's just a huge pain to set everything up and get going, it isn't really worth it anymore. I hate to say it but it kinda killed my interest in flight simming for the time being at least. And I can't fly without the chair because all of my peripherals are attached.

I would love to have a solution, but at this point I'm not too interested in working on a solution. Hopefully the interest will return. I'm still glad I built the thing, I learned a ton and it feels like quite an accomplishment for me.

What Arduino code are you using? Try to reduce the communication speed between the PC and the Arduino. I'm using the SPS code for the Sabertooth controllers and the key is put a jumper from the 8 pin to GND and left the jumper connected. If you are using the "regular" code I think you can configure it manually from the Arduino IDE and then in adjust it in Simtools/FlyPT interface.

Reduce the speed from 500000 to 115000 reduce that behavior and the motors temperatura significantly. Before that solution my motors got very very hot after a few minutes of work, now the motors are cold after more than an hour of work.

I know how frustrating it is and I hope you find a solution man, really.

Thanks for the input. I'm using SMC3. When I get a chance, I'll look into it. I have no idea how to reduce the speed, but I'll look into it for sure when I get a chance.

What is your lever length ? Try to reduce it. I thought your motors are too weak (
Or PSU

I don't think so, I built a 2DOF rig similar to the OP and with more or less the same motors and PSU specs and it was powerfull enough for a 2 DOF if it is well balanced.

Can you move manually the hall sensor while it is connected to the motor? I installed an elastic belt (instead of a toothed belt) that makes me able to move the hall and make the motor moving without any command from the PC. By this way the arduino will correct the position of the motor by they own. Try it (be careful manipulating the motors and sensors manually) and notice if the movements are smooth or what. If the movements are good, I'm sure the problem is in the communications between the PC and the Arduino. That way is the way I use to isloate the problem (then I discovered by chance about the jumper I mentioned)

I would think that as well except it behaves this way whether I'm sitting on it or not. I can put my hand on the lifter rods and feel it when I'm not on the chair. It doesn't seem to change at all as I apply load.

CTC is 65mm.

I have toothed belts, so no. That would be a good way to take the sensor out of the equation (not to mention it would be much easier to level the levers). I could just disconnect the lifter arms and remove the belts and let them spin freely while I try to assess whether or not it is still misbehaving. It might be tricky though with such a short lever to try to assess motion smoothness, but I think it could be done.

Fwiw, I don't think it is a malfunctioning component as both motors exhibit this "lumpy" motion. But maybe the halls don't have a high enough resolution or something? They are the 360° kind as I didn't want to take a chance with breaking them. IIRC, they're geared 3:1 so I'm getting almost 270° of travel from top to bottom.

One thing I have noticed that may or may not be related... if I use the SMC3 utility to make an abrupt motion change using manual mode, or if I set the output pattern to square wave to get hard/violent direction changes, the PSU's sound like they power down for a moment during that violent direction change. The fans spin down, etc... like they're not getting enough juice. This type of violent action doesn't occur during a normal flight, even during a wreck, as it is extremely jarring and could break something, but it is something to note.

Your Hall effect sensors are perfect, I used that kind of sensors with my 2DOF rig and now with my 3DOF and they just work perfect.

The problem with the PS when violent movements is totally normal, keep in mind the controllers are surging lot of current from the PS and when the motors stops a regenerative current come back to the controllers and PS and they will clip in the best cases, I burnt a few (cheap) PS because of that.

I had exactly the same lumpy motion with my motors till the PC and Arduino speeds match at 115000.

PD: I see in the SMC code at the first of the Setup() function (line 360 in the SMC v1.7a) the serial port starts at 500000, try to change it to 115200. Change the interface (if in Simtools) or the (serial) Output (if in FlyPT) to 115200 too to match speeds and try again.

Hope it helps

try setting 20khz and Fpid/9 in smc3

Hi Josh, interesting stuff! I wanted to give it a try. I changed the bit rate to 115200 in Arduino and successfully uploaded the code. However, SMC3Utils does not seem to work with this setting. I guess it expects SMC3 to run at 500000 bits per second. I also see no option to change that in SMC3Utils.ini. Did you also encounter this problem and if so, do you have a solution for this issue?

Oh yes, in fact with the SPS code (a SMC3 modified code) if I put the jumper that configure it to 115200 SMC3utils does not work so the workaround is disconnect the jumper to acces SMC3utils, I don't mind that it can happen with the regular SMC3 arduino code. My though in your case the way is maybe configure the PID manually in the code and upload it to arduino directly (it maybe add a complexity layer of course)

Ooof. I don't want to without a way to check the pot signal and to manually control the motors. So losing SMC3utils doesnt sound attractive to me.