PDA

View Full Version : MX 28 Servo Questions



Deimos
09-28-2013, 11:36 PM
Hi all,

I've been considering using an MX 28 on my mech for a variety of reasons (easy to mount, precise encoder, easy to interface with, etc...) and I had a few questions.

First of all, can I get absolute position feedback while the servo is in wheel mode? I've seen nothing in the documentation to confirm or deny this. The idea would be to have some flavor of microcontroller babysitting the MX 28 so i can control the direction, speed, and goal / stop position. (a gear motor with a separate encoder would probably be less expensive, but also would introduce more points of failure)

Also, I've seen a large number of people experiencing problems with these (MX line) servos. Would I have a good chance of avoiding issues if I closely monitor the servo temperature (I seem to recall a large amount of problems involving overheating)? The servo would mainly be doing low torque, large & moderately fast movements.

btw: this is for a continuous rotation capable turret (complete with slip ring). If this were not the case I would just use an ax-12.

jwatte
09-29-2013, 12:52 AM
I believe that just reading the position register will give you continuous position. I can't swear 100% that this works, as I haven't tried it myself, but I know of no reason why it wouldn't just work.

There is no goal/stop position in wheel mode, though.

The failures we're seen on MX servos have to do with the servos destroying their own communications circuits when run with heavy load, and when run on 4S LiPo batteries. There are also some motor failures when the servos go into stall. With proper wheels, this might be less likely to happen, and with 12V voltage sources (3S LiPo or such) you may also have better luck.

Pocar19
09-29-2013, 10:48 AM
Indeed you can read the absolute position of the encoder ( I tested only MX28) while the servo is spinning in wheel mode.

Pocar19
09-29-2013, 10:51 AM
I did have experience few communication issue with MX-series. For my project I don't use heavy load and I powered my daisy chain of servos at 12V using a power supply.
Do you know if there is a way to protect again the failure of communication circuits for MX servos?
Thanks!

jwatte
09-29-2013, 12:09 PM
Do you know if there is a way to protect again the failure of communication circuits for MX servos?


You will need some electronics at each servo. A small resistor (10-20 Ohms) and a 5.1V Zener diode at each servo should protect all other servos. I had that set-up at the head end on my bot, and it protected by microcontroller and buffer, but it didn't manage to protect all the servos from each other.

Btw: I'm going to try 3S LiPo with my robot. If it still doesn't work robustly, I will give up on these servos.

Pocar19
09-29-2013, 01:39 PM
Thanks for your feedback!
In my case I put 500uF capacitors into 12V line and for the time being I haven't got anymore communication issue... I'll keep my fingers crossed!

Deimos
09-29-2013, 10:54 PM
Hmmm, the way i plan to have it set up the MX 28 communication line will be completely isolated from all other servos. As i'm using a 3s lipo, I'd assume i wouldn't need to worry about sticking any electronics onto the 12v line.


There is no goal/stop position in wheel mode, though.
I probably should have explained myself a bit better, the microcontroller I mentioned would be separate from my main controller, (a BBB w/ usb2ax) in theory it would poll the position and correct/move to new goal positions just like the servo does normally, except with control over cw/ccw and ability to cross the 0 degree mark.

Also, thanks for the help, much appreciated!

jwatte
09-30-2013, 12:04 AM
In my case I put 500uF capacitors into 12V line

The problem I'm talking about is not on the power line, it's on the TTL line. The protection circuitry needed is to clamp the TTL line to 5.1V, and to limit spike current with a small resistor. These servos, when mis-behaving, have destroyed a protective 500 mW Zener I've put on the TTL bus. I just kept one centrally; this let the servos destroy each other. To combat that, you'll want one resistor/zener combination per servo (split off from your strand.) Note that the Zeners will in turn introduce capacitance on the TTL communication bus, so too many of those may impede communications.

Separately, the power bus does get 3V-4V spikes while the motor is moving. I'm clamping these with a 16V 5 kW TVS diode and a 470 uF capacitor on the servo side, and a 33 uH 14A inductor between the servo power and any other electronics. It's still not perfect, but better than nothing. For really effective clamping, you probably want one of these at each strand of servos.

escott76
09-30-2013, 05:10 PM
As a data point I have been running a quad walker with MX-28's all around. Relatively decent loading, the NUC on top weighs a fair bit, and not any trouble so far with any of the servos. I am running 3 cells.

jwatte
09-30-2013, 05:29 PM
not any trouble so far with any of the servos. I am running 3 cells.

This gives me hope!