blog; Motor Controller Options for 2015

[DampRobot]

Originally Posted by Ty Tremblay

Let's hope this doesn't result in a massive shortage of motor controllers in 2015.

This. Given the price point, size, and the availability of other options, these things will be in truely massive demand once the season starts. I wouldn't be suprised if they sold 10,000 in the first month they're on sale. I also wouldn't be surprised if they sold out within the first day. So, basically, there will be massive problems if there aren't enough of these controllers to go around.

I trust Vex and CTRE to learn from their mistakes and do this launch right, but then again, I've been wrong before about trusting companies to have enough stock in time for the season. I hope we won't see any problems with inventpy come January (or even better, come November).

On the other hand, these things look amazing. I'm so jealous of the students that will get to use these things. I remember that it seemed that motor controllers just got 2x better when the origional Talon came out, and somehow they've don't it again. This is a huge leap forward in FRC motor controller technology.

[Jon Stratis]

Originally Posted by donkehote

What rule disallows thermal paste? Sticking chemicals on the outside of electronics is modification? Would the same rule disallow Velcro attachment, or labels being applied to speed controllers? Both of those use chemical compounds applied to the heat sink causing different operating conditions.
Your post reminds me of the bumper pool noodle tape issue from last year.

Some things don't need regulating IMHO.

Personally, I doubt any inspector would even notice thermal paste between a speed controller and an aluminum mount - you would have to take the controller off to see it! And as an LRI, it's something I've never actually considered before. R64 from 2014 does state that they "shall not be tampered with, modified, or adjusted in any way"... so the question comes down to whether thermal paste would be considered a modification or adjustment to the speed controller. I think most people would say no, but someone wanting to be strict may draw a parallel between R64 and R76, which states "All pneumatic COMPONENTS must be used in their original, unaltered condition", and has a blue box that specifically mentions items like painting...

You could, if you wanted, draw a parallel between R64 and R76 as both say you can't modify a part, and then a parallel between paint and thermal paste, as they are both topical coatings for a part. It's a bit of a stretch (and not anything I personally would call at an event unless otherwise instructed by the GDC), but it seems like a good question for the Q&A to me.

[cjl2625]

"Onboard closed-loop PID control" for the Talon SRX? What exactly does this mean? Can you tell it to move to a certain position like a servo, or something like that?

[theawesome1730]

I personally think the integrated pigtails are great. No more soldering crimp connectors (even though you shouldn't need to, better safe than sorry) to 2 sets of wires. I don't foresee reuse to be an issue if you invest in anderson powerpoles or similar connectors. We look forward to being able to beta test these awesome devices

and I'm liking the more standard measurements of the new victor

[Tristan Lall]

Originally Posted by donkehote

What rule disallows thermal paste? Sticking chemicals on the outside of electronics is modification? Would the same rule disallow Velcro attachment, or labels being applied to speed controllers? Both of those use chemical compounds applied to the heat sink causing different operating conditions.

That's exactly the dilemma: FIRST presumably doesn't want you to douse it in toluene (because you might dissolve something important), but they probably don't care if you write on it using a marker. And yet both chemicals' effects are governed only by the rather unspecific rule against modifications.

The typical compromise is to adopt an ad hoc interpretation of the term modification that is as loose as the circumstances will allow, in order to approximate a just outcome. It's a mess and an inefficient use of everyone's time to have to design with this nebulous constraint in mind.

A clear rule predicated on the functional effects of the modification would be more equitable, but it would also be harder to enforce with consistency and rigour, because of differences in officials' ability to identify and evaluate failure modes. A narrower rule that permits more freedom in design choices, but which also exposes some additional failure modes would also be more equitable, but less safe.

In this case, I vote for the latter: the catch-all safety rule is so strong that the additional safety afforded by the no modifications rule is sometimes negligible. Where that's the case, allow modifications, and focus attention on the teams that do something unsafe, rather than dividing attention among the teams that make mundane modifications safely.

[BitTwiddler]

Originally Posted by cjl2625

"Onboard closed-loop PID control" for the Talon SRX? What exactly does this mean? Can you tell it to move to a certain position like a servo, or something like that?

PID stands for proportional-integral-derivative. This three dollar phrase means that for a speed control loop you can tell the speed controller to go to 2500 RPM and expect it to command the motor to go to 2500 RPM quickly, smoothly and precisely (assuming you have a rotary encoder to measure the speed at the shaft). I may not have the details quite right in this case but that's the general idea of PID for speed control.

For more on PID enter PID Controller in Wikipedia.

[cjl2625]

Originally Posted by BitTwiddler

PID stands for proportional-integral-derivative. This three dollar phrase means that for a speed control loop you can tell the speed controller to go to 2500 RPM and expect it to command the motor to go to 2500 RPM quickly, smoothly and precisely (assuming you have a rotary encoder to measure the speed at the shaft). I may not have the details quite right in this case but that's the general idea of PID for speed control.

For more on PID enter PID Controller in Wikipedia.

I know what PID is, but I'm wondering what specifically an onboard PID on the motor controller entails and how you could use it. My last post was pretty unclear, sorry.
For example, could you just send an error, output range, and gains to the motor controller and would it calculate the PID itself?

[BitTwiddler]

Originally Posted by cjl2625

I know what PID is, but I'm wondering what specifically an onboard PID on the motor controller entails and how you could use it. My last post was pretty unclear, sorry.
For example, could you just send an error, output range, and gains to the motor controller and would it calculate the PID itself?

Oops sorry if I sounded condescending. It would be kinda cool if it could automagically compute the PID coefficients but I suspect life wouldn't be that easy.

[Monochron]

Originally Posted by theawesome1730

I don't foresee reuse to be an issue if you invest in anderson powerpoles or similar connectors. We look forward to being able to beta test these awesome devices

Could someone post some examples of the Anderson powerpoles they have used with success in the past?

The only types that I have used (excluding the battery connectors) have been very shoddy; wires falling out, plastic crumbling, etc. Admittedly, this was years ago so I am hoping that newer versions are much better.

[Ether]

Originally Posted by FrankJ

The second controller, the Talon SRX, is a CAN enabled motor controller with additional communication protocols and PID control.

Has anyone been able to find any details about the algorithm used for the built-in PID in the Talon SRX? Is it strict PID, or does it have, for example, user-selectable feedforward etc.

[timytamy]

Originally Posted by Monochron

Could someone post some examples of the Anderson powerpoles they have used with success in the past?

The only types that I have used (excluding the battery connectors) have been very shoddy; wires falling out, plastic crumbling, etc. Admittedly, this was years ago so I am hoping that newer versions are much better.

We use these as well as the 15A/45A versions when appropriate. The trick to them is that you really need the correct crimper, we've used that one for the past three years, and I can't recall anything really going wrong. We've found that generally the contact gets pulled out of the housing before the crimp fails.

[Jared]

Originally Posted by cjl2625

I know what PID is, but I'm wondering what specifically an onboard PID on the motor controller entails and how you could use it. My last post was pretty unclear, sorry.
For example, could you just send an error, output range, and gains to the motor controller and would it calculate the PID itself?

It means you can plug a sensor like an encoder or potentiometer into the speed controller, then, with CAN, you tell the speed controller you want to enable PID control and what your PID gains are, and it will do the PID control logic on the speed controller itself.

The advantage is that you don't have to waste cRIO processing power on PID loops and you can run them much faster. The disadvantage used to be that the only controller you could use for it, the Jaguar, was big, expensive, and not as reliable as the Victor.

[Andrew Schreiber]

Originally Posted by Jared

It means you can plug a sensor like an encoder or potentiometer into the speed controller, then, with CAN, you tell the speed controller you want to enable PID control and what your PID gains are, and it will do the PID control logic on the speed controller itself.

The advantage is that you don't have to waste cRIO processing power on PID loops and you can run them much faster. The disadvantage used to be that the only controller you could use for it, the Jaguar, was big, expensive, and not as reliable as the Victor.

And the CAN bus was a ticking time bomb on your robot. If one device dropped out the whole bus was screwed due to CAN Timeout exceptions.

Sincerely hope that's been fixed.

[FrankJ]

Originally Posted by Ether

Has anyone been able to find any details about the algorithm used for the built-in PID in the Talon SRX? Is it strict PID, or does it have, for example, user-selectable feedforward etc.

The original blog says they will be in hands of the Beta testers soon. Hopefully they will post their experiences.

[Jared Russell]

Originally Posted by Ether

Has anyone been able to find any details about the algorithm used for the built-in PID in the Talon SRX? Is it strict PID, or does it have, for example, user-selectable feedforward etc.

I would love to see a velocity PID loop with:

Selectable Frequency (though a fixed rate would be fine if it is fast enough)
User tunable P, I, D gains
User tunable Feedforward gain
User tunable Forward/reverse stiction compensation
User tunable Maximum output cap
User tunable Maximum integrator output
VERY IMPORTANT: Slave mode so that a single sensor can be used to control a multi-motor, multi-Talon mechanism (like the drivetrain). It was never really clear how to do this with Jaguars.

I do not think I would ever run a position control loop on the speed controller, but I can envision making heavy use of velocity control.