Placing motor controllers on moving components?

[J_B_Wood]

Hello all,

This summer we are redoing our 2016 robot's electrical system. We are making great progress so far, and learning a lot. Our robot has a single large actuated arm, with several encoder-controlled actuators on it. In order to minimize the lengths of encoder signal wires running from our Talon SRX's to the encoders on the arm, we wanted to place some of our Talons on the arm itself. We had some reservations about how well the talons would hold up under heavy movement of the arm, and wanted to see if anyone knew how sturdy talons can be when shook around. Any advice?

Here's a video of our robot for context: (team 1389)

https://www.youtube.com/watch?v=P8pkLSnA-c4

[ollien]

I can't imagine it would be a good idea to have your CAN chain moving around like that. I would just use some cable carriers or ziptie the cables down to the arm.

[J_B_Wood]

Originally Posted by ollien

I can't imagine it would be a good idea to have your CAN chain moving around like that. I would just use some cable carriers or ziptie the cables down to the arm.

Thanks for the advice. But you couldn't see any possibility of the talons themselves breaking under heavy vibration?

[ASD20]

Originally Posted by doughy_goods

Thanks for the advice. But you couldn't see any possibility of the talons themselves breaking under heavy vibration?

We mounted a talon on our shooter turret this year in a high vibration area and it spontaneously died at Worlds. I can't say for sure what caused it to die, but it is possible it was due to the vibration.

[polytechnique]

Yeah, just to agree with what was said before ...

Placing sensitive electronics on moving parts not firmly attached to the chassis is a little sketchy. We've had some talons spontaneously die on us simply for being exposed to vibration or hard jolts.

A roll of encoder cable and some cutout plugs is cheaper than a new set of talons.

[Andrew Schreiber]

Paging 4143 or John Novak...

https://www.chiefdelphi.com/media/im...8de35014_l.jpg

There's not a good picture of it but 4143 put the talons for their drive in their drive pods and ran power and CAN to them.

[Max Boord]

On 4592 we ran a Talon SRX inside our flip down intake as it was the only way to get the versa encoder to reach the motor controller. Worked through 2 regionals and worlds with no problems on ether comp or practice bot.

[fergus180]

This year we mounted 3 Talon SRXs to our shooter in a reasonably high vibration scenario and they worked fine all season, but then again others have experienced issues with talons just dying during comp. I think that your intake is a lot more prone to vibration than the others listed. The only thing i can recommend is not soldering your can connections and use a nice connector you can lock like the Anderson ones (you can also buy colours other than red and black to match the can wire colours) we have had problems with the solder joints breaking. During a recent off-season comp we had three connections brake on us and with the way CAN works that's possibly all your motor controllers not getting a signal.


Coloured Anderson connectors
https://powerwerx.com/anderson-power...lored-housings

Our Robot:
https://www.chiefdelphi.com/media/im...dbab4eee_m.jpg

[Jewski]

[quote=Andrew Schreiber
There's not a good picture of it but 4143 put the talons for their drive in their drive pods and ran power and CAN to them.[/QUOTE]

This is a better shot of the CAN Tallons on our modules from last year.

[Steven Smith]

One of the suppliers might be able to chime in a little more definitively, but I would not anticipate vibration or g-forces associated with a moving arm to be excessive for the motor controller. You already experience quite a bit of this on a moving robot that slams into walls, other robots, defenses, etc.

As alluded to earlier, what I would be concerned about is wiring. The controller should be firmly affixed to whatever solid item makes up your arm, and your wiring should be strain relieved to that surface as well. You need to be aware of your connectors and their vibration resistance, screw terminals need to be torque'd to spec, etc.

If you have a section of wire that will be in repetitive movement, it needs to be of a high quality wire intended for repetitive movement (typically very high strand count with a more flexible jacket). If you imagine wiggling a paperclip back and forth until it breaks, the same concept applies to wire in a moving joint. Watch minimum radius bends (aka, don't lock down either side of the wire, forcing all the movement in a 1/4" area, but also can't leave too much wire flopping around such that it can catch on something that binds.

TL;DR - Probably not best practice for most folks, but not because the motor controllers can't take the movement, because you need to be really careful on the wiring side.

[Josh Tatum]

Originally Posted by Max Boord

On 4592 we ran a Talon SRX inside our flip down intake as it was the only way to get the versa encoder to reach the motor controller. Worked through 2 regionals and worlds with no problems on ether comp or practice bot.

If you want to extend the cables you can use more cables and these connectors. The cables will plug into each other, however, they will flip the pins so you need to use an odd number of cables. We used this setup and it worked great and allowed us to keep our motor controllers in a better position.

[RyanShoff]

We ran 12 srx talons on a bus with four 1' stubs. We did have some can bus errors. The errors didn't affect anything. We also tested with the bus running through 4 slip rings. It worked but we didn't compete like that. The can bus is quite robust.

Go ahead and try it. It worked for us. Just have a backup plan if you run into trouble.

[Michael Hill]

As for concerns of the motor controllers themselves, I think the big worry for some of the older motor controllers (Jags and victors 883-888s) would have been the TO-220 fets and the through-hole capacitor kind of flapping around in a high vibration environment. The jag took care of this (or at least helped) with a plastic ring that surrounded all the fets and the capacitor. It probably didn't eliminate the problem, but probably reduced the likelihood of failure to a good degree. The real worry in that situation are the solder joints becoming fatigued and fracturing (much like the paperclip analogy above). As far as the Talons, Victor SP and other motor controllers are concerned, I haven't taken any of them apart, but from my understanding, the fets are all surface mount and the capacitor is laid over on its side. Vibrations in surface mount components are less of an issue because there's no large weight and associated moment arm associated. The bent-over electrolytic capacitor is only slightly troubling, especially if some kind of silastic compound is used to keep it in place. It's not really going to "flap around" much in its enclosure. But if you look at the application of the capacitor (a power filter capacitor), it takes out some of the transients from the input power prior to the power converter, so taking it out of the circuit may not cause the unit to fail catastrophically (what will probably happen are intermittent power cycling when the power is unclean that are a pain to diagnose). Just my thought, I could be wrong, but I'd really like to see a motor controller designer chime in.