gokhan
Junior Member
Posts: 89
|
Post by gokhan on Nov 9, 2016 19:35:14 GMT
I am having oscillation from the backlash of the reductors. It is so much that I am not even able test Fabi' s sw.
I think I will need to replace the reductors with low backlash ones.
I am also considering a top frame to fix reductors to eachother, in order to make the reductor installation more rigid.
Please feel free to comment, advice if you have some similiar experience?
|
|
|
Post by tronicgr on Nov 10, 2016 11:21:12 GMT
Yes, you can hear the backlash!
You can try a few things:
1. Lower the MAX motor speed setting (if the ratio of the gearboxes is low it might have trouble to stabilize).
2. Use low PID values like P=20 I=0 D=0 ( increase P value only later)
3. Use audio shielded wiring for the position sensors to minimize electrical noise from the motors.
Thanks Thanos
|
|
|
Post by clyevo on Nov 11, 2016 10:12:02 GMT
more or less the same thing that happened to me a year ago Some help with setting up the 1) Ki and Kp of the inverter, i should be doing this with full load on platform, but i need to be sure before setting up the platform 2) PID settings of thanos board I am sure there is relation between both this settings I cant get fabi software tuning to work so i need to tune it manually. Can someone suggest PID settings and way to tune it manually. This is Ki and Kp settings from manual 1 Relaxed the PID settings like thanos said 2 Increase acceleration and deceleration time 3 Auto-tuning with rotation is important 4 Kp Ki settings of vector control must be tailored according to the load and gear inertia 5 Reduce the cockpit weight 6 Solve the EMI problem : shielded metal chasis on critical electronic component, shielded signal cable, etc In the end i was able to minimize this oscillation but at the cost of lagging in the movement of the simulator. So i scrapped it and in process of building a new one probably with better EMI suppression.
|
|
|
Post by riton39 on Nov 11, 2016 14:16:38 GMT
|
|
gokhan
Junior Member
Posts: 89
|
Post by gokhan on Nov 11, 2016 20:17:55 GMT
Thanks for all responses.
I did lower the max speed setting. Also lowered Kp to 10. And put Ki to 1. I was also experimenting with low speed torque compensarion parameter and it was set to 150%. I set it back to 100%.
Now it seems allot more smooth.
But the reductors have very large backlash, 5 of them 3-5 degrees, one of them is near 10degrees. Their ratio is also quite low 1:30.
I ordered 1:50 ratio low backlash reductors and will be able to return the old ones back. I also get help to weld the bolts which fixes the heim joints on both high and low ends. I hope with the new reductor these will all help reduce the backlash. Then I will be able start working on tuning.
I will share the result with all.
Thanks
|
|
|
Post by clyevo on Nov 12, 2016 3:11:21 GMT
In that video it is potentiometer terminal without shielded cable right? What improvement did you see after you use cat7 cable (on oscilloscope and real life performance)? And how you wire the twisted pair. I suppose signal and return path are suppose to be twisted together in pair right?
|
|
gokhan
Junior Member
Posts: 89
|
Post by gokhan on Nov 12, 2016 9:04:46 GMT
Amc is installed in an aluminum solid box. Inverters are on a seperate box away from Amc box. All slgnalling cables are standart shielded cnc signalling cables. They are connected wlth metal cover connectors. Dc ground and ac ground are seperately tested. At the moment, i don' t think that signal and return cables are twisted together. The cable is ordinary shielded cable with 3-5 cores in it. I will buy double shielded cable, which has an outer shield over all pairs, also every signal pair is shielded itself. In that, the signal and return cables will be twisted to form 120 ohm balanced signal path. I think it will improve, or at least it will be safer.
I experienced heavy interference on 5 m. USB cables that i use to connect sim controls to sim computer. I had to shorten them and route them away from the power cables in order to eliminate the interference. I will also look for shielded power cables if possible.
I believe a big part of my problem comes from the backlash and the low ratio of the reductors? But I will try to get rid of the other potentiel risks.
|
|
|
Post by paulg100 on Nov 12, 2016 10:22:13 GMT
"I experienced heavy interference on 5 m. USB cables that i use to connect sim controls to sim computer. I had to shorten them and route them away from the power cables in order to eliminate the interference. I will also look for shielded power cables if possible."
Im currently using usb cable for my pots, when i tested a shorter run against sheilded CY cable on oscilliscope there was no significant different in signal quality, maybe with longer runs it becomes a problem.
|
|
|
Post by riton39 on Nov 12, 2016 10:24:21 GMT
In this video I use a shielded USB cable for potentiometers. I have not yet received the hardware, I would give you the result when everything will be wired
It must be ensured that the voltage of the potentiometer is perfectly smooth, without any interference.
The 5v USB power supply is not perfectly clean.
I am considering a hub with linear 5V power supply
|
|
|
Post by riton39 on Nov 12, 2016 10:32:21 GMT
Amc is installed in an aluminum solid box. Inverters are on a seperate box away from Amc box. All slgnalling cables are standart shielded cnc signalling cables. They are connected wlth metal cover connectors. Dc ground and ac ground are seperately tested. At the moment, i don' t think that signal and return cables are twisted together. The cable is ordinary shielded cable with 3-5 cores in it. I will buy double shielded cable, which has an outer shield over all pairs, also every signal pair is shielded itself. In that, the signal and return cables will be twisted to form 120 ohm balanced signal path. I think it will improve, or at least it will be safer. With all these good things, you always have parasites? Seen with an oscilloscope? The motor cables are shielded?
|
|
gokhan
Junior Member
Posts: 89
|
Post by gokhan on Nov 16, 2016 21:42:25 GMT
A friend convinced me that the shielded power cables are overkill. So i will postpone that???
I tested sensor signal with oscilloscope and very little noice so I am very happy on that.
Welded parts reduced backlash allot. Increased ramp up and down times from 50 ms to 500 ms and re do the automated motor tuning. Now it is almost usable except one reductor with huge backlash. The new reductors will be delivered next week. But at least i can work(on PID tuning) with this for the moment.
I will put a new video after the reductor replacement.
|
|
|
Post by paulg100 on Nov 17, 2016 6:05:20 GMT
"I tested sensor signal with oscilloscope and very little noice so I am very happy on that."
would be very interested to see the oscilloscope reading of your sensors to compare to mine are you able to post a picture?
thanks.
|
|
|
Post by riton39 on Nov 17, 2016 6:22:02 GMT
I think it is very important to shield the motor cables, they spread over a long length. I have just received mine. Some mVolts can make a difference and movement on engines, no? You can compare with my oscilloscope video of noise. www.youtube.com/watch?v=yaUJlDW7I5kfor me 0.5v noise.... !!
|
|
|
Post by clyevo on Nov 18, 2016 3:28:52 GMT
A friend convinced me that the shielded power cables are overkill. So i will postpone that??? I tested sensor signal with oscilloscope and very little noice so I am very happy on that. Welded parts reduced backlash allot. Increased ramp up and down times from 50 ms to 500 ms and re do the automated motor tuning. Now it is almost usable except one reductor with huge backlash. The new reductors will be delivered next week. But at least i can work(on PID tuning) with this for the moment. I will put a new video after the reductor replacement. Using bolted aluminum profiles is the bolts doesnt hold against intense vibration of the motion simulator. It tends to slide over the bolted part. I did not use any thread locking glue but you can try it.
|
|
gokhan
Junior Member
Posts: 89
|
Post by gokhan on Nov 18, 2016 15:32:48 GMT
"I tested sensor signal with oscilloscope and very little noice so I am very happy on that." would be very interested to see the oscilloscope reading of your sensors to compare to mine are you able to post a picture? thanks. Next week I will bring my osc. back again. Then I will take a pictur/video and share it. Would you please share yours too? including your osc settings?
|
|