There was significant discussion for and against 775 Pro drives preseason, especially when AndyMark announced the Redline gearbox.
Did your team attempt and/or successfully use a 775 drive in 2018? If so, what were the results and what lessons can you share with the FRC community.
My team ran an 8 775 Pro drivetrain successfully this year. We use AndyMark Decimates going into a WCP SS. We had it geared 27.9:1 if I recall correctly. One thing we wish we would have done is change the motors out after our second event at least because performance started becoming noticeably different by the end of the season.
We developed and tested a two-speed 775pro gearbox in the 2017 offseason, we ran the same gearbox in competition this year. Three robots, five competition events, countless hours of driving, and the only motor failure we’ve had was from a rookie dropping it after pressing on a pinion. :rolleyes:
We did use voltage ramping and peak / continuous current limits, but CTRE makes these easy for any FRC team to use now. We used one Talon SRX with three Victor SPXs per gearbox, SPXs have made an 8x 775pro drive appreciably cheaper.
We’re discussing building a single speed 775pro drive this offseason, we used low gear only a handful of times this year.
In what way was your drive performance different? Is there any noticable wear on the motor brushes? What, if any, limiting did you implement?
By my calculations, only ~12fps with 4" wheels and 18fps with 6". What wheels did you use?
6". I believe adjusted FPS came down to about 14~15 FPS.
We noticed that we started pulling higher amps, and drifting was noticeable between each side (more than at the start). We didn’t ever inspect the brushes, but after talking to 2767, we know they replace theirs every so often (cant remember how often right now). We actually had an issue at Worlds where our motor was slightly shorting with the case, took awhile to realize the motor was the issue. It was over a million ohms so the inspector let us pass but it was something that concerned us and we want to prevent for next season in between events.
For current limiting, we did use some with the SRX, I believe it was when it spiked up above 45 amps for 4 seconds, it would limit it to 30 amps? Something along the lines of that since half our motors were on 30 amp breakers. We didnt do the current limiting the way I would have liked thinking back (more limiting would have been better), but it got us through the season with 0 motor failures.
For Steamworks, 2767 started the season with higher current limits set (initially 80A?) and found that we had to replace motors regularly. We lowered the limits gradually during the season. We also started doing motor “health checks” between matches as part of the normal pit turn process. Basically we run each actuator on the robot at low, medium and high voltage and monitor resultant speed and current draw. These values are compared against benchmark/new values. If we saw significant change/degradation, we looked closely at the driven mechanism. If the mechanism wasn’t damaged/worn in a way that explained the change, the motors got swapped out. This allowed us to get motors replaced before they failed on the field.
That said, we did some off season investigation and found 40A would work indefinitely. For Power Up, we had current limits set at 40A and no longer had the issues. We continued to run the “health checks” because they are valuable in proactively finding issues.
I’m curious, if the health checks weren’t finding bad motors this year, what were they finding? (I’m not doubting that they’re effective, just curious.)
This year, health checks told us we weren’t damaging motors. Last year, typically, we would see a step change in current and/or speed and could relate it to an event (usually a pushing match).
It’s like checking the oil in your car every day just to find that the oil level isn’t changing. Why check your oil every day? So you know that your not burning oil.
The health check caught a bad elevator motor at States. It also showed us a Talon that was not operating properly. It also could pick up a damaged mechanism, over tighten belt etc.
Last year, with the higher current limits, we did see changes in the 775Pros’ voltage constants which led us to proactively change them.
With a 40A limit, we didn’t find a lot of things this year: I believe we had one motor changed out as a result of the health checks and it was a 9015 on an intake. We also thrashed through a Dual motor 775Pro master/slave issue on our elevator drive, but that turned out to not be the motors. The “health checks,” along with some other instrumentation software we wrote, helped in the troubleshooting.
In general, this kind of test can also help identify mechanical issues like worn gearboxes, bent/bound mechanisms, over-tight belts, things needing lube, etc… Closed loop control compensates for most of these (to a point), but fixing the root cause is preferable, otherwise the “plant” has changed since the tuning was optimized.
Jinx, you owe me a coke.
So, one of your motors was current limited and the remaining 3 on each side didn’t have a problem? Cool.
(The SPX doesn’t do current limits. We ran 3x miniCIM in a ~21fps drive, and ended up going from 1xSRX/2xSPX to 3xSRX in order to have current limiting working on all three motors)
The SPXs were slaved to the Talons, so their outputs were the same. We assumed that current draw per motor should be similar since all of the motors were the same age and all driving the same gear. We did check current draw in driver station logs and found slight variation in current draw per motor but not significantly enough to affect anything.
We’re you slaving the SPXs to the SRXs?
Assuming all the motors are in the the same gearbox and one motor has a SRX with a current limit set while the rest of the motors are on SPXs set to follow that SRX then you are current limiting all the motors in that gearbox.
The current limiting just adjusts the voltage being fed to the motors based off readings from the current sensor in the SRX. As long as the voltages being fed to all the motors are identical, the motors will share the load equally and each will see the same amount of current whether they are on the SRX or the SPX.
As a practical example we ran 4 775 pros per side with 1 SRX and 3 SPX following the SRX running a current limit. All our motors were current limited perfectly.
There is no reason to have a SRX on every motor in a gearbox. The PDB also provides slower current monitoring anyway if you want to do fault checking.
I’m curious about your current limiting. On an 8 motor drivetrain, current limiting to 40A per motor will result in momentarily pulling 40x8=320 amps on a full drive reversal or momentary stall / pushing match. In my experience that is more than enough to brown out the Rio on a full battery. Was there anything else you did to prevent that? Aggressive voltage ramping, etc?
We found that with low voltage ramping, we ended up current limiting to around 22 amps per motor to prevent brownouts this year on full drive reversals / pushing matches. Under 180 amps continuous seems to be our brownout sweet spot on a full battery. Going higher on the voltage ramping resulted in a very floaty feeling in the drivetrain because it didn’t ramp down quickly enough.
TJ² (team 88) have competed with 8x 775pro drive trains for two years now and we’ve really loved it. The TL;DR is it’s incredibly durable, and very simple to code.
Our 2017 robot used a standard Vexpro 3 CIM Ball Shifter with a custom conversion stage on the input side. The custom box was a 3D printed housing that acted like a CIM-ile, except allowed us to use 4 motors. (It’s not easy to describe, so I attached CAD of it below.)
Going into 2018 we realized that we were relying on a lot of modification to the ball shifter and decided we should just make a custom gearbox designed specifically with 775s in mind. Using a design from Chief Delphi as a launching point, we made a custom box that baked in a lot of the lessons from 2017. The 2018 gearbox was single speed for both mechanical and driving simplicity. It also let us use an SRX mag encoder, which solved basically all of our drive issues that stemmed from the ball shifter’s grayhill encoder. We haven’t released our full CAD yet, but here is our custom gearbox and last years gearbox.
One of the biggest lessons we learned from the past two years was that 775pros are incredibly durable, considering their reputation for being one short stall away from death. This year we used ramp rates to smooth out our current draw and went most of the season without any current limiting. We added 35amp current limiting after our first big pushing match, but only to prevent brownout. The motors survived without current limiting for nearly 4 events. We have never smoked a drive motor in either year. All of our code was done with basic TalonSRX functions. I kind of understand how we do it, so that’s how I know it’s really simple.
I know every parent thinks their child is the best, but I really can’t express how much I loved our power train this year. 8x 775 saved us a lot of weight (1 gearbox w/ motors weighs the same as 2 CIMs alone) and the space savings allowed us to add a second set of wheels on the inside of our drive rail that were crucial for us being able to traverse the corner of death. We are always looking to improve (5406 did some smart stuff with their gearboxes 
) but making the 775 switch has definitely been worth it.
I thought so. I think I remember we sent a single command and had all three motors spin. IANAP. *
I thought that would happen (which is why I bought the SPXs), but that was not my experience.
.Bold claim, how did you verify that?
I do respect both of your opinions, so this issue just went back onto my “build and test it this summer” list…*
They probably checked the driver station logs to see current draw, the same way we did.
Yes, obviously.
Mmman, to clarify - I’m interested in whether you induced locked/stalled conditions or just drove and checked. When we locked out our DT (put a bolt through it then put power on the motors), we found that the SPX was not limiting current, and made the switch to 3x SRX. When we were just driving, it was hard to replicate.