Quote:
Originally Posted by Tom Line
I don't think that anyone can deny that FRC in general is in a bit of drive train war right now. The latitude we've been given in the last few years in motor choice and availability have made 6 cim drive trains increasingly common, with some few teams even moving to an 8 motor drive train.
That isn't a bad thing from an engineering standpoint.
Drive trains area already limited by the 120 amp breaker. Assuming the new CrossTheRoad PDB and current monitoring does what we hope it will, teams should be able to ride 'upper limit' of power while not popping their breaker.
I have to wonder, though, if we would be better served in FRC by limiting the total power output of the drivetrain. There is an argument to be made about the increasing price of multi-cim multi-speed gearboxes. A reasonable argument can also be made that the kitbot drive train has been rendered obsolete. This year, we saw an extremely brutal game. This made partially possible by the wide open field, but the high acceleration and high top-end speed that our drivetrains delivered were also responsible. A lot of robots left the field in pieces: even those of multiple-time world champions.
I think we've reached the point where it's time for FRC to consider reigning in the drive train power. I wouldn't be adverse to a max 6 cim, or even 4 cim and 2 mini-cim 'power' limit. I don't have my motor sheet in front of me to lay down the power numbers of those motors right now. What do other folks think?
|
I personally would like to see the rules more relaxed. I would rather the GDC do more of "I wonder what teams will come up with if we allow more of this..." rather than preventing what teams can create.
But playing devils advocate: Assuming the rule you described was implemented, how would one begin to regulate this rule? Drivetrain output power is determined by gear ratio, motor configuration, motor load, etc. All of these variables are engineering parameters that teams can choose based on their desires. To ensure robots are under the maximum power usage, the inspectors would need to measure or calculate power draw while the robot was under some worst case driving load to show it never pulls x much power. Is this even possible in the pits? Wouldn't this also mean that you would need a fully functioning/driving robot before you were able to pass inspection? Furthermore, unless the inspectors had their own current or power meters that they could connect to your robot while you were doing this demo, you would also be forcing every team to use the CAN interface. The inability to regulate this rule, my just be a reason why it can not be a rule.
In addition, the PDP we are currently using for Alpha Testing provides current monitoring for each channel on the PDP (1-16), not for the entire system. I suppose one could measure all channels, and then sum them up to provide an estimated system current draw.
However, I think I understand why you are thinking of this approach - to be a uniform solution to prevent main breaker trips, correct?
The 120Amp main breaker is not limiting at all. It takes a long time to trip and can handle 500+% of 120 amps for seconds before tripping under the right condition. The main breaker is a thermal device, not a current device. It trips after passing a certain temperature. Large current surges is just one way to raise the temperature of the device, however it is completely true that if you took a heat gun to the main breaker it would trip even under no current load. The proper way to prevent the main breaker from tripping is to design motors with the proper gearing ratios to prevent such cases.
However, If I were to design an automated control system to help prevent tripping the main breaker in 2015, this is how I would do it:
Since I know the main breaker trips based on temperature, that is the variable I want to measure. In control systems, when you have a system that is controlled by one variable (heat in this case), but you measure something else (current in this case) and use that to estimate the other, that is called an indirect measurement. This can leads to all sorts of trouble like making the wrong decision. When I am designing any new control systems, I want to directly measure all of the control variables, only if it is impossible for some reason (i.e. against the rules) then will I venture down the indirect measurement path.
I would use a thermocouple mounted to the power terminal of the main breaker to monitor the terminals temperature. When the temperature of the terminal goes above some threshold (based on spec sheet, or team experimental data) I would then warn the driver (via driverstation/lights on robot etc.) that the main breaker is approaching the tripping point. I could then automatically enter some "reduced power mode" of the robot where I would start scaling back the max power to the most power hungry systems determined by the max current draws measured on the PDP. For example, I could reduce the max output of the drivetrain to 75% while in reduced power mode, and stop the motor all together when I detect a stall condition (large current surge for some duration of time on some channel). I would keep doing this, until the temperature drops down below another threshold where I could then re-enter "normal power mode" and let the robot rip.
For added effect, I would mount the main breaker on top of a fan, and kick the fan on during reduced power mode to help it cool down faster. Or leave it on at all times to help prevent going into reduced power mode in the first place.
The system would also have the ability to be disabled by any of the drivers by pressing a button in real time.
Why would I do this over just measuring current? Well the main breaker is a thermal device. If I were measuring current, all I could determine is that the the current spiked over some threshold. This is not an indication that the breaker is approaching its trip point. The breaker can handle large loads for many seconds and even minutes depending on the conditions. The problem is that the time to trip is not deterministic. It can be milliseconds, seconds, or even minutes. In a match, I want to stay competitive as long as possible, for example staying in a pushing match for as long as possible. Prematurely reducing/disabling power to robot systems because I detected a current surge is very overly cautions and can reduce my competitiveness. The last thing I want to do is lose an elimination match because I disabled/reduced power to certain robot components because I thought the main breaker might blow, especially when it was no where near the tripping point. Measuring temp gives me the best information to make the best decision at that moment in time, I can let the robot rip it up for as long as I can until the temp of the main breaker reaches a worrisome point. I don't have to worry about prematurely reducing my systems. I have more useful information measuring temp and so I can make smarter decisions. I can also allow the robot recover back to full potential and rip it up again, once the temp drops down keeping my competitiveness, and the full advantage of my robot.
Remember these systems don't guarantee that the breaker will trip, they just inform you that the probability of the breaker tripping is higher. In a regional/district winning match, I would probably let the robot approach the trip rather than prematurely disabling/reducing power to critical systems. This is just me, I am more of a gambler and I trust the systems we design as a team. This is especially true in the last 30 seconds of a match, where I probably need my robot to push its limits to the end. I'd rather take the chance and hope it doesn't trip during a critical match, rather than reduce the capability of my robot where I can no longer be as effective/competitive. Its a trade off, its a gamble.
One could argue "couldn't you just use the PDP and average the current surge over some amount of time to help increase the probability that the main breaker is approaching the trip point.?" Yes you can, however this is still an indirect measurement. If measuring the temp of the main breaker is not feasible for some reason in 2015, this is the next best approach, and the one I would employ. It still has the problem that I do not know if the breaker is warm or not when the current surge is detected, but it is way better than nothing.
All of this doesn't replace good drive train design. These are just warning systems and can provide the drivers useful information in a match. The best defense against main breaker trips is to design appropriate drives in the first place. 2 speed gearboxes, ramping up your drive commands, and knowing when to switch to low gear can get you the 99% solution.
Main Breaker manufacture Spec PDF is attached
Hope this helps,
Kevin