Motor Controllers

[FrankJ]

Talon SRX are my favorite. Unless you need the advanced features of the Talon the Victor SP makes more economic sense. I will wait to pass judgement on the new 2016 controllers when they prove their durability.

Other than size the jaguars don't deserve their bad reputation. We found them reliable when connected correct and not filled with metal shavings.

[aldaeron]

Originally Posted by JPBlacksmiths

So this is a question that I have had for a few years now, and never really gotten a good answer to it:

Which motor controller is the best for which situation? Meaning is there specific applications were one controller is better then the other, or is there just one motor controller which is the best, period?

This question has resurfaced for me especially with the creation of the new motor controllers such as the spark.

Thanks,
JP
Co-Captain Team 806

Getting back on topic to answer OP's questions...

As with most things in engineering design, the answer here is "it depends". In my mind there are three key features to look at: cost, capability and reliability. I am only going to focus on the controllers you can buy going forward (Talon SRX, Victor SP, REV Spark, MS SD540). Since you can't get spares for other controllers easily I would stay away from or use them sparingly (and with LOTS of spares ready just in case).

Talon SRX at $90 is the most capable, but also the most expensive.

Capability: Talon SRX allow use of a Controller Area Network (CAN) bus. The other option is Pulse Width Modulation (PWM). With a bus you get two way communication between controllers and the RoboRio.

Also, the Talon SRX has an input data port that allows you to do some routine control calculations directly on the controller instead of using the RoboRio. You can hookup small breakout boards to the data port and do a number of control loops. Some examples are that you can hookup an encoder to a shaft and thru the data port and tell the motor controller to drive forward until the encoder has reached 1000 ticks and then stop. Another example is that you can tell the motor controller to drive forward until the upper limit switch on your elevator is tripped and then stop. With PWM the RoboRio has to check all these sensors and issue the commands to stop. Since the RoboRio is doing a lot of calculations it may be slower to respond (by some number of milliseconds) to when the switch is tripped on your elevator. The Talon SRX offloads these mundane control calculations from the RoboRio.

Cost: You pay for all these nice features (CAN bus chip, extra connector for data port inputs). Over a whole bot this extra cost can add up.

Reliability: Last season the Talon SRX performed well with few (if any) failures. Since there was no defense and therefore few long duration high current pushing matches, there could be an argument made that some more testing is needed under these extreme conditions. Most people would agree it is a solid and FRC tested design. It also has a very small foot print and a sealed aluminum enclosure. This prevents debris from getting in and shorting out electrical components (a problem with older controllers). Aluminum is a great heat conductor and allows the heat generated by the controller to be moved away from the electronic chips without the need for a separate fan.



Victor SP at $60 is a robust, but basic controller.

Capability: Talon SP is a simple Pulse Width Modulation (PWM) driven controller. No fancy data inputs, just the ability to set the speed between 0 and 100%.

Cost: You pay for a smaller form factor and an aluminum case. More expensive than other PWM controllers and does not have data inputs.

Reliability: Same as Talon SRX. Most would agree it is a proven, dependable controller for FRC.


REV Spark at $45 is a new low cost controller with some nice looking features.

Capability: REV Spark is a simple Pulse Width Modulation (PWM) driven controller. You can set the speed between 0 and 100%. It also has a basic input data port that will allow you to directly wire in a set of limit switches to stop it. These limit switches are great for an elevator or a rotary arm with a top/bottom limit switch. No fancy encoder or other control capabilities. Pretty impressive for such a low cost device.

Cost: Cheapest one on the market for FRC right now.

Reliability: This is brand new this year. Preliminary test data looks positive. It has a plastic case, which will not transfer heat as well as aluminum. It claims not to need fan cooling, which is being tested in lab conditions (I believe results release soon). Obviously the real test is to see it during the season. Be cautious about switching to a new controller with limited testing unless you can afford to replace the controllers. It is a bit larger than the Victor SP and Talon SRX. For the price it looks promising.


MS SD540 at $50 is a new low cost controller with some nice looking features.

The SD540 implements many of the same features that the REV Spark does, albeit in a different layout. There are some concerns about the design choices made for the SD540 that are detailed in another thread. Same overall story - low cost plastic body controller with PWM input and limit switch control build in. Untested.



Hope that helps. PM me for more detail (I glossed some of the nitty gritty details)

-matto-

[GreyingJay]

Originally Posted by GeeTwo

However, this exception was not included for the withholding allowance; if you wanted a pre-terminated controller as a spare part, it needed to be included in the withholding allowance.

I wonder why they made that distinction. Is it too much of an advantage to have a drop-in replacement motor or controller ready to go?

I could see the first task for a student in the pit being to crimp some connectors onto a set of spare motors and controllers, ready to go.

In fact... *writes that down*

As for the poll, I voted for Talon SRX since it seems to offer quite a lot of potential advantages for the fairly minimal extra cost. Virtually every one of our robot failures last year, both on our practice field and unfortunately during our first regional, was some combination of our software addressing the incorrect PWM or DIO, the encoder wired incorrectly, read incorrectly, software values set/reset incorrectly, motors trying to run the forklift past the edge (e.g. we didn't have time to wire in a hard limit switch), etc. If we thought to use the limit switch and encoder functionality in the Talon SRX, and communicate over CAN instead of PWMs and DIOs, our code (and failure modes) could have been much simpler.

[JPBlacksmiths]

Originally Posted by Michael Blake

Wondering why Victor 884/888 was left out of voting yet Jaguar was included?

--Michael Blake

simply because this is all of the current motor controllers, nothing as far as I know has been made to replace the jaguar.

[Ari423]

Originally Posted by JPBlacksmiths

simply because this is all of the current motor controllers, nothing as far as I know has been made to replace the jaguar.

Though they don't have the same name (because they're made by different companies IIRC) the Talon SRX effectively replaces the Jaguar. It does the same PID control, limit switches, etc with better CAN support. I wouldn't buy Jaguars new even if they still were for sale.

[GeeTwo]

Originally Posted by SoftwareBug2.0

There's no reason for a Jaguar to be anyone's favorite

Jaguar is my least favorite, but I couldn't let that challenge pass. I came up with three reasons for Jag to be someone's favorite (I left out nostalgia and thinks like being the biggest, heaviest, and having the smallest continuous current capability, and having a nonstandard PWM response):

• Supply Voltage: works down to 5.5V and up to 30V, Tied for first best at both ends with SPARK.
• Aesthetics: The jaguar case is a work of art compared to the boxes with fans (Victor 884 and 888) and the boxes without fans (the others).
• You can control a Jaguar with RS-232. I can contrive of a few applications where this would be useful, like controlling from an old PC.

Edit:

Originally Posted by MrBasse

Don't forget the fact that the flames that poured out of Jags were very entertaining for the kids!

Now I feel cheated; I've smelled and seen plenty of blue smoke, but never seen flames .

[techhelpbb]

Originally Posted by GeeTwo

• You can control a Jaguar with RS-232. I can contrive of a few applications where this would be useful, like controlling from an old PC.

PWM from parallel port and they all can be controlled by an old PC

I am waiting for the single IC speed control module. It is coming. Just a chip with bolts on the top like an IGBT and an electrically isolated metal plate on the bottom.

[GeeTwo]

Originally Posted by techhelpbb

PWM from parallel port and they all can be controlled by an old PC

I'd never heard of this, and a bit of web searches about the old centronics port leads me to think that the PC-side hardware was not made to support PWM; whether it would be possible to hack it with a tight loop that shut down the PC for any other purposes is probably hardware dependent. Has someone actually done this (using passive components, which was my meaning)?

Originally Posted by techhelpbb

I am waiting for the single IC speed control module. It is coming. Just a chip with bolts on the top like an IGBT and an electrically isolated metal plate on the bottom.

Heat dissipation is a big issue here, unless you're talking about replacing the spike rather than a speed controller. The amount of electrical isolation demanded by FIRST is also a steep challenge. I expect brush-less motors and controllers to be FRC legal before single-chip brushed motor speed controllers. I'm not holding my breath on either one.

[MrBasse]

Originally Posted by GeeTwo

Jaguar is my least favorite, but I couldn't let that challenge pass. I came up with three reasons for Jag to be someone's favorite (I left out nostalgia and thinks like being the biggest, heaviest, and having the smallest continuous current capability, and having a nonstandard PWM response):

• Supply Voltage: works down to 5.5V and up to 30V, Tied for first best at both ends with SPARK.
• Aesthetics: The jaguar case is a work of art compared to the boxes with fans (Victor 884 and 888) and the boxes without fans (the others).
• You can control a Jaguar with RS-232. I can contrive of a few applications where this would be useful, like controlling from an old PC.

Don't forget the fact that the flames that poured out of Jags were very entertaining for the kids!

[techhelpbb]

Originally Posted by GeeTwo

I'd never heard of this, and a bit of web searches about the old centronics port leads me to think that the PC-side hardware was not made to support PWM; whether it would be possible to hack it with a tight loop that shut down the PC for any other purposes is probably hardware dependent. Has someone actually done this (using passive components, which was my meaning)?

Ether did PWM 'bit banging' of the handshake wires on a serial port:

http://www.chiefdelphi.com/forums/sh...52&postcount=1

I used to play .mod MIDI files on a parallel port justing just a resistor ladder and a capacitor to couple still have that stuff somewhere and that was on a Intel 80386 DX 16MHz. That's complex analog audio at up to 15kHz. So yes I am sure if I sat there with some C code I could control several FRC PWM speed control on FreeDOS. I would not attempt this in on any Windows NT kernel because you need a kernel mode driver. It is likely possible on BSD/Linux.

DOS runs in real mode. You will not block the timer interrupts or other interrupts doing this. I used to run this silly trick even with EMM386 running and Windows 3.11.

HighTreason take it away old school...
https://youtu.be/4KUdBgg8Oe4

Then again....I still maintain Commodore 64 code I wrote....so I am pretty dusty.

Originally Posted by GeeTwo

Heat dissipation is a big issue here, unless you're talking about replacing the spike rather than a speed controller. The amount of electrical isolation demanded by FIRST is also a steep challenge. I expect brush-less motors and controllers to be FRC legal before single-chip brushed motor speed controllers. I'm not holding my breath on either one.

The heat dissipation of any single part in an FRC robot is no where near the heat dissipation of the big IGBT we used to charge Navy submarine batteries with this big charger we built. The base of the integrated package is metal and you torque the package to the heatsink plate with either a coupling patch or compound. One of my cohorts left a databook on a board during a test...let's just say we will not read that databook again .

The issue is merely distributing the heat away from the parts that get hot. Either make them more efficient or make the removal of heat more efficient. Same reason we could fill power supplies up with canola oil....as long as you do not let the pressure build or reach the flash point it will not conduct but it will make the heat distribute around (do not try at home kids). That silly trick allowed almost double the wattage from a power supply. By the way - TDI in Hackettstown, NJ was selling a design like this actively before AstroDyne bought them.

There was a time when people argued no FRC speed control would last without a fan either. Power MOSFETs keep getting better and better. Lower the switching losses and on state resistance enough and the heat dissipation will follow. Just please do not do it with silver in the semiconductor die (let's just say it makes the power semiconductors go bad eventually and leave it there).

[GeeTwo]

Originally Posted by techhelpbb

Ether did PWM 'bit banging' of the handshake wires on a serial port.....

I'll accept your word and Ether's.

Originally Posted by techhelpbb

The heat dissipation of any single part in an FRC robot...

I think you made my point better than I did.

[techhelpbb]

Originally Posted by GeeTwo

I think you made my point better than I did.

The actual problem is not actually the heat:
It is the whether doing that tight integration is financially workable given the market.

A massive power transistor has a much wider market than FRC and a much higher price point.

However with all the interest in drones and places like SuperDroidRobotics selling kits that are FRC size like for plows and snowblowers...the market is growing.

I suspect back in the day when audio receivers were individual power transistors many people also thought they would not be surpassed by integrated modules. Odds are if you tear down a retail home reciever you find modules in plenty of them.

[BenGuy]

Originally Posted by JPBlacksmiths

So this is a question that I have had for a few years now, and never really gotten a good answer to it:

Which motor controller is the best for which situation? Meaning is there specific applications were one controller is better then the other, or is there just one motor controller which is the best, period?

This question has resurfaced for me especially with the creation of the new motor controllers such as the spark.

Thanks,
JP
Co-Captain Team 806

Well the new Talon SRX is best if you can take advantage of what it offers over the Victor SP which is control by CAN and direct sensor feedback (encoders). This is why it costs more. If you can't take advantage of that then the Victor SP is better because it is cheaper and offers the same size advantages as the SRX. Hope that helps.

Also, Jag - ewww - don't use that...

[electroken]

Originally Posted by aldaeron

REV Spark at $45 is a new low cost controller with some nice looking features.

Reliability: This is brand new this year. Preliminary test data looks positive. It has a plastic case, which will not transfer heat as well as aluminum. It claims not to need fan cooling, which is being tested in lab conditions (I believe results release soon). Obviously the real test is to see it during the season. Be cautious about switching to a new controller with limited testing unless you can afford to replace the controllers. It is a bit larger than the Victor SP and Talon SRX. For the price it looks promising.

Part of my own evaluation of the SPARK was having a look inside. While much of the case is plastic, the SPARK label in the center is covering a heavy c-channel shaped slug of aluminum with heat transfer pads to encourage cooling of the 4 MOSFETs and 3 electrolytic capacitors under it. My gut feel (with nothing but experience to back it up) tells me this device will survive a 2-minute pushing match just fine.

The footprint is close, but not identical to that of the Talon SR.

[evanperryg]

Originally Posted by DaveL

Would like to here from others on how they connect the Talon green and yellow CAN wires together. Vex sells a nice device but it's expensive and it takes up a lot of room.

Check these out.
We used these last season. They're toolless, and we've never had a problem with them falling off or breaking. They're designed for larger wires than the CAN wires, but they still work fine. They're a little bulky, but not big enough to cause problems.