Spreading the power over more motors should reduce the temperature of all the motors. I could see the benefit of 6x MiniCIMs over 4x CIMs for that single reason. This also holds true for running 8x 775Pros. I have to believe that when cruising long distances at speed, such as the game required this year, the 775Pro’s being fan/air cooled helped keep them at a lower average temperature verses the sealed motors. A game that requires a lot of moving short distances quickly might not be ideal for the 775Pro setup.
We have never had auto shifting drive trains.
The driver shifts into low only when needed.
Sometimes our driver uses the low gear to decelerate from speed quicker.
Seriously though, I we need like a bullet point list of all the things you have to do to make sure your 775Pro drive train doesn’t go up in smoke. A poorly implemented CIM drivetrain overheats, trips breakers, and generally drives badly. A poorly implemented 775Pro drivetrain would do all those things and then not drive at all.
The Battery Beak internal resistance measurement will give you a decent top-end to available power. That’s usually in the 0.010-0.015 ohm range. Probably more like 0.015-0.020 after wire drops. So maximum available current at the PDP is going to be around 600A. That assumes a dead short on the other end, which stalled motors aren’t quite. 8x 775 Pros would be around a 0.011 ohm load, assuming no wire losses. That would be about 400A, which the 120 breaker would put up with for a surprisingly long 5-12 seconds. At 50A per motor, the 40A snap actions wouldn’t ever actually trip. The rub is that this knocks your apparent voltage down to 4V, and you theoretically blacked out at 4.5V.
So, backing up, let’s assume we limit ourselves to avoiding brownout protection, and that our effective battery resistance is 0.02 ohms. Brownout protection kicks in at 6.8V. So if we assume our battery voltage is 12V, that’s a 5.2V drop over 0.02 ohms. Which means you actually only have 260A to spread amongst your 8x 775Pros, so about 30A apiece. Maybe 40-50A if my resistance assumptions are off.
Having to limit battery voltage to this drive system didn’t raise a giant red flag? It seems like big enough a deal to me that I would think twice before evangelizing this design to the community. I believe there are benefits to this design with proper software implementation, but your average team isn’t going to nail that and will spend a bunch of time dealing with the sadness of sitting with a blown main breaker.
Don’t forget they had to put in a voltage ramp function to avoid browning/blacking out under load reversals.
Really, you’ll generally have to think a LOT more carefully about your controls with this kind of setup. You’re almost better off with dumb PWM control, honestly. If you’re using a Talon over CAN, don’t use voltage compensation mode or it’ll happily compensate up to 100% output as your battery voltage drops. And be sure you set your peak output limits if you’re using any of the closed loop modes. And all the voltage and command ramping rates.
Just want to say I do agree that your average FRC team should probably not be running a 775 drive, as not putting the correct limits in with code will definitely result in dead robots, which no one wants to see.
You know, that’s an interesting point. A Banebots 775 - 18V run at 12V is actually very close to a 775Pro limited to 9.6V. The BB is slower with more torque, the max power at 12V would be 267W instead of 223W. You could almost go with an 8x BB 775-18V and forget about the voltage limiting and such.
If you know the internal resistance(s) of your power system, you can model the voltage drop under load for more realistic sprint calculations and to predict and prevent brownouts. For example the MK-12 ES-17-12 datasheet lists the internal resistance as ~12 milliohms.
In 2014, team 687 measured the internal resistances across the cRIO-era power system, and integrated it into our custom drive train calculator, which accounts for voltage drop for sprint calculations. We’ve yet to get this data for the new control system. Our calculator does not take into account voltage drop across wires, which can also greatly affect performance.
It would be interesting to see a refined calculation with wire weight and extra gearing included, and some use-case calculations involving voltage drops, traction limiting, and the relative heat de-rating of 775 Pro vs CIM.
Maybe some teams will do some off-season testing so we can have data instead of calculations
I will admit that I have lots of interest in a 8x 775pro drive, mainly due to potential weight/space savings. Current limiting is easy to do with the Talon SRX. We will likely experiment with this in the offseason.
The potential issues with gear wear do concern me, though, and I hope that our 3d printing is accurate enough to make a set of those “inverted CIM sims” whose design was posted a month or two ago.
Maybe CTRE will come up with some kind of “smart” current limiting that allows the user to specify a current limit as a function of fan speed. Or even as the time integral of some function of fan speed and current.
Or maybe the Talon can figure out what the coil resistance is, based on the speed, voltage, and current, and use that as an indicator of temperature and make the current limit some parameterized function of that.
This would be nice, but I suspect simply picking a reasonably conservative hard current limit (it takes about 300A to brown out the control system with a fully-charged battery, so you’re never going to be able to really use more than ~35A per motor in an 8-motor drive anyway) would probably do okay. Will report back if we do indeed try it.
All of our testing to date have been concentrated on comparing motors with a constant voltage power supply in order to compare apples to apples.
The problem with testing under battery conditions are the variances in the batteries themselves.
I think this is something I want my team to do for the 6 CIM vs 4 CIM vs 6 miniCIM vs 8 775pro drivetrain situation.
I am really interested to see the results. Getting proper test conditions is going to be an interesting challenge.
Anecdotally, our 6 miniCIM drive was the best, low maintenance drive train we have ever done. The key was practicing for hours on end never made the motors warm, even with the crazy Robowrangler defense bots beating on us.