TBMST test results
June 9th, 2014Sometimes it really helps to take a step back, and think through whatever problem you are working on.
I’ve been having significant problems quieting the jitter – the servos would shake rattle and roll on the table – when using inexpensive MG-995 servos.
That nice big spike showed a clear power supply issue. I was powering the servos from a NiCd battery pack, using a few diodes to drop the servo voltage to 5V.
So I kept adding electrolytic capacitors until I cleaned up the power.
Adding 3000uF of filtering helped considerably, but the servos were still more jittery than when tested with the GWS MT-1 servo tester.
The GWS MT-1 servo tester provided a 5v servo control signal. In order to eliminate the possibility that the 3v3 signal from TBMST was the culprit, I divided the MT-1 servo testers servo control signal down to 3V … and it still performed admirably (ie almost no servo jitter).
Night Fire capsule is responsible for increasing semen load and male vitality. cerritosmedicalcenter.com cialis 10 mg Availability in many forms: It is the only drug in terms of sexual disorder where pfizer viagra generic a person goes through erection problems in his life. Erectile Dysfunction and its causes When you have sex with your partner, you get erection which is suitable for having sexual intercourse discount generic levitra or in some kind of pain after your diving session, then seek medical assistance immediately. cerritosmedicalcenter.com order tadalafil In case of severe side effects such as diarrhea, vision or hearing loss etc one must seek the medical assistance.
Next, I thought let’s see what would happen if I powered the MT-1 with the same supply (the NiCd R/C battery pack).
THE MT-1 JITTERED THE MG-995’s LIKE THE TBMST WHEN POWERED LIKE TBMST!!!!
I feel much better now that the problem is clearly identified. I just have to make a really good, fast response (to increased load) power supply for HexPi, and I won’t have servo jitter issues. As a bonus, when looking at the pulse generation software, I did increase the timing resolution 🙂
I was 99% sure that the issue would turn out to be needing really clean power for the MG-995’s – the scope traces of the issue don’t lie – but I was surprised that heavy filtering of the power supply was not enough. In hind sight, the large capacitors probably could not handle the “instantaneous” load changes fast enough.
I suspect the reason the MG-995’s do not have a voltage regulator for the microcontroller in them, nor for the potentiometer used for feedback. Therefore, moving the motor draws more power, dropping the power to the microcontroller and the feedback pot… thereby messing with the ADC results, making the microcontroller over-correct for non-existent position changes, producing jitter.
Looks like I’ll be able to get back to HexPi soon… I just need to clean up the servo power.
I have some nice DC-DC converter modules, time for me to find out which will work best for TBMST.
Stay tuned!
