Strategies to prevent overcurrent/breaker trip
 

Please describe in this thread your team's method of preventing motor overcurrent/overload (and subsequent breaker tripping) with an optimized drivetrain.

What are the different methods to prevent breaker trips, and what are the pros and cons of each? How did you implement it with a multimotor gearbox? Can your method adapt and still protect the breakers in case of sudden huge load changes like ramp inclines, goal towing under all load conditions, your robot being pinned, or tug-of-wars?

Suggestions I've seen so far, each with their own strengths and weaknesses:

1) Design the drivetrain to simply tire slip at any torque greater than 30A will provide. (Paul Copioli, Thunderchickens)

2) Software limiting PWM at "full stick" unless a "turbo" button is hit (Al Skierkiewicz, 111)

3) Current sense and throttle back (various teams), including Tim Skloss' "Brown Toast Breaker" method (team 930): Current sense & integrate, modeling the breaker's heating curve. Throttle back whenever the "toast starts to get brown"...

Current sensing is "currently" ;) being discussed at: http://www.chiefdelphi.com/forums/sh...?threadid=2205

Has anyone implemented any of the following?
- Breakaway/autoslip clutch of some kind
- Automatic transmissions (NOT operator controlled)
- *TORQUE* sensor feedback to the RC (i.e. strain gauge)
- How to throttle back multimotor drives CORRECTLY to prevent backdriving any of the motors
- Automatic load shedding / load prioritizing
If so, please describe the system.

Anyone have another system??

<edit>
If you've already discussed your system in another thread, please just cite the thread (add a link) and summarize it here. I'd love to see this thread become a nice summary / collection of all the methods in one place! :D Thanks!
</edit>

- Keith

Re: Strategies to prevent overcurrent/breaker trip
 

Our current limiting sounds similar to 930's, though maybe not as sophisticated as you describe. I've already described our system in this thread, but i'll summarize it here too:

We didn't do any integration in an attempt to predict and preempt a circuit breaker tripping. We simply sense the current being drawn by our drive motors, and programmatically limit the power requested of the drive system.

Each time through the control loop, if the current being drawn by any one of the drive motors is above a certain threshold, the program scales back the power being requested of all the drive motors. The scaling factor starts at 100%, and every time through the control loop that the current is above the threshold, the factor is cut back by 5 percentage points (but not below 10%.) Once the current gets back below the threshold, each time through the control loop the scaling factor is increased by 1 percentage point until it gets back up to 100%.

I was able to experimentally determine a threshold which would prevent the breakers from tripping even with the wheels stalled when pushing against an immovable object.
As I was tuning our system, it appeared to me that the only thing required to keep our drive from being back-driven was to not allow the scaled back pwm value to fall into the "dead zone" between 117 and 137.

Spray all the breakers with cold spray right before the match. Canned air will do the same thing if it's held up side down and sprayed.

Use lots of fans around your motors and breakers

Wrap the main breaker in some foam rubber to isolate it from shock when you get hit by another robot. A sudden jolt can cause it to trip.

You may also want to mount the breaker panel on something soft like a stack of O-rings where the mounting screws are.

If you use 4wd skid steer, design 2 or all of your tires to have traction only in the forward and reverse directions of the robot. Team 599 knows about this.:) Too much side traction will cause breakers to trip when turning.

Bingo.

We used that at the Midwest Regional last year after our several trips at the Buckeye Regional. After that we had minimal problems, compared to the problems we had before.

My team is working on a current-sensing circuit similar to Team #111's. Basically it'll display the current level of each motor and total current on a display mounted on controls. Not sure if it'll be programmed to affect the 'bots behavior yet. Thanks to team Wildstang for their help.


Matt

Great Idea!
 

Oh, that's brilliant... :D (Pouring some liquid N2 on them might be overkill though. ;) )

Anyone have a way to accelerate breaker cooling during the match? Perhaps a clamp-on vertical heat sink fin array (and fan?) to each drivetrain minibreaker with some heat sink grease? Or is the case's thermal insulation value too high for that to be effective?

Too bad Freeze Mist isn't an "electronic part", or I'd suggest buying a can from Future-Active, stick it in the "electronics box" with a kit servo, point its tube at the breaker panel, and have the RC "order a cool down" as needed. <chuckle>

Was there a ruling last year banning "remote sensors" outside of the Electronics Box (EB)? I don't see anyone using remote sensors (other than "wire resistors") outside of the EB. It would be nice to attach detectors to items, such as thermal, torque, or strain gauge sensors.

Is thermal sensing even useful in a two minute match, considering the sensor settling time?

- Keith

I know it's briliant. I'm just sorry I didn't come up with the idea. I read about it in some other thread around this place.

To best way prevent over curent in motors is to have a clutch in the system. this can be the clutch provided in the drill motor transmisions, tire slip, slip in a v belt or round drive drive, or a self made clutch. Stalled motors over heat very quickly, not just because of the high current, because the intergral fan is not moving also. It should be noted that breakers let a much higher curent pass than there rating for a few seconds (this info is out there some where)

For the drive system we fond that tie slip worked best, experment with diferent tires, gearing and the right weight load to find oout what works best. We found that ball pick up systems work best if they can slip also, it pervents excessive jaming as well as over heated motors.

As for other mechisims like lifting arms and such design them to do the normal duty at or less than 20% motor torque rating limit, if you want control with a speed control. Add limit swithces too.

Get a clamp on DC curent meter, learn what is going on. They are great for diagnosis alos, espically if you have two motors powering a mcechinism.

I have a couple of ideas. I was wondering if the following would be legal.

1) An onboard tank with some kind of cold liquid circulating through tubing wrapped around the motors. And/or a fan blowing across the tubes making a cold wind on the breakers.

2)An onboard can of cold spray actuated by a servo.

The old fashioned way...
 

Sanddrag, in the past FIRST did not allow any fluids that were not lubricants as far as I know. Not being able to regulate the safety factor on the compressed air tanks for FIRST robots would be quite scary to FIRST.

While I am typically a supporter of "drive-by-wire" systems in general, I feel that implementing a drive by wire system may not be as wholly benificial as one may think.

What do I mean by this?

To make a long story short, A FIRST robot is often asked within the course of a match to perform a task that "pushes the envelope of the robot" Placing an indicrimate software restriction on such an action could impede the maximum performance of the robot. Although such a system would certainly allow for consistant, reliable performance, I can think of many many times where a robot "pushed the envelope" and in performing the action could alter the results of a match.

There is still a great deal of credit to be given to an operator who can "feel" how the robot is behaving. While this idea may be a thing of the past and in texts like "The Right Stuff" I have found that to this point, it is difficult to gauge where a software program should draw the line between safety and pushing the envelope.

Sorry... Not on AHL. How about expanding the air supply???
 

Sorry, BOTH are illegal due to using materials not on the Additional Hardware List (AHL). Remember, All materials on the robot must be either from the kit or the Additional Hardware List (AHL). Cold spray is not on the AHL, and you have no AHL legal refrigerant to circulate. You can use cold spray as a TOOL, but not as a PART on the robot. BTW, I already joked about that before. See above. :)

Hey, how about clamping a thermal mass (made out of AHL stuff) onto each breaker and chilling THEM before each round?

--- Air supply expansion as a coolant? ---

Depending on the pneumatics manual rules, you MIGHT be able to create an expansion valve for 120psi air and a heat fin assy to create a pretty lousy cooler out of the compressed air supply... Gad, that "feels" AWFULLY expensive in terms of compressor power usage for a TINY bit of cooling.

Are there any Air Conditioning experts out there that can run the numbers for that one and comment on it? How much cooling can you get by expanding the 120 PSI air supply to atmosphere for two minutes?? Assume you have NO other cylinder loads, have precharged ALL the tanks in this setup before the round, and clamped the heatsink onto ONE breaker:

(compressor)--> (ALL tanks)--> (control valve)--> (expansion_valve & heatsink assy)--> open air

At round start you open the control valve, and run the compressor full tilt. How much cooling could we get, in terms of degrees dropped on that one breaker over the two minutes?


That aside, I'm more interested in seeing DESIGN METHODS to prevent breaker trips listed here, versus simply finding ways to FOOL the breakers with chilling. Anyone have any OTHER ways to prevent breaker trips other than chilling, and the design methods already listed?

- Keith

Re: The old fashioned way...
 

The compressor has a safety valve attached, that pops at 120 PSI. Not a problem.

BTW... As a reminder before we get any further into pressures and cooling, this thread is for people to list their general robot design methods to create machines that INHERENTLY prevent breakers trips.

Please list how you design your robots to normally prevent breaker trips, yet still allow you to push the power demand situation whenever needed.

Thanks!
- Keith

How about gearing your robot reasonably, or if you choose to go Extremely fast, step the voltage that is put into your motor so you do not put as large of a strain on the motor.

If you are geared for, say 15 feet per second, you will most likely pop the 30A breakers as you try to accelerate, if you do it with full voltage. But if you start by applying 1V, then 2, then 3... etc, you will not draw as much current. Simple physics, V = I*R.....

Either that or that oboard can of freeze spray :)


the 60 amp breaker will handle current draws of MUCH more than 60 amps..... these stats were taken from Woody's speach during kickoff in 2002 -

@200 Amps - Trips in 8 seconds

@100 Amps - Trips in 1 minute, 48 seconds

Just know that you cannot pull too much current for too long. If you design within the limits of current draw, your robot will be fine.

Tom

Re: The old fashioned way...
 

that is why you have a big red button, and when u push it the program overrides all the little software restriction, kinda like when u overload a cisco router and it defaults to a basic security mode, but also kinda not like a cisco router at all......muhahahahaa

Legality
 

Im not sure as to the legality of this, although it is quite tricky, and I am rather impressed by whoever created it. Although, In my opinion, it violates the "spirit of the game" and "gracious profesionalism" in that it gives an unfair and unjudgeable and unregulatable advantage to the team who uses it. Although, in any other instance in life, i would say "more power to you, you beat the system." So take my word for what you will, whatever that is.

Here is something i posted about "cold spray" last season sometime. The figures were done by an engineer that wanted to figure out if all the circuit breakers were the same, and if cold spray/freeze spray, etc. had an effect on them.

--------------------------------------------------------------------------------

Ok....I am not an electrical person and this is basically all copied and pasted from an email i recieved about our breaker problems...this may help some people...


------------------------------------------------------------------------------------

I've picked two of the five breakers as being much better performers, these two
>go about 45% longer at 120 amps, compared to the worst breakers.
>
>COLD SPRAY WORKS
>Precooling the breaker provides 60% more duration on the first power cycle.
>This rises to 120% more duration on repeated power cycles at 5 minute intervals
>(to simulate the high cycle rate found in finals competition). Bottom-line: a
>10 second blast of cold spray will restore the breaker to its original
>performance even if it has just been overloaded. Where do we get more of this
>stuff. I've used my only can during this test?

------------------------------------------------------------------------------------

1. Suspend our best circuit breaker by the 6 ga. wire and 2-4 strips of latex
tubing, directly between the battery and the power panel
2. Possibly move the power panel closer to the battery (or the battery closer
to the power panel) so we can cut down the length of 6 ga. wire.

------------------------------------------------------------------------------------

Ok....it's me again.....i know our final solution to our problem was that we did some electrical work...mounted the breaker onto a flexible/bouncy kind of V shaped lexan strip...and we really use the Cold Spray....

Re: Legality
 

Unless I missed something, there is no rule that says you cannot do this. Power to you people. Now you know, so use it.

Cooling off the circuit breaker defeats the purpose of a breaker in the first place, by cooling them you are allowing more current to pass through, the reason its tripping is that there is too much current, it has to be thought of in that respect, otherwise you might as well just circumvent the breaker all together

2 exceptions:

1- the breaker is tripping due to mechanical forces
2- you cool the breaker between matches

Freeze spray never allows significantly more current to pass through the breaker. It should be noted that every time the breaker trips it is easier to trip the next time. Freeze spray can help to get the breaker back to normal, in other words it will trip as if it was new or close to it. Last year, the breaker problems became more apparent than in past years. We had a breaker that took only a slight tap with a screw driver to trip it because it had tripped so many times. While I know this thread was designated for design strategies, there are several tricks you should use around the breaker that we learned last year.
1. The freeze spray does work. We used it a lot last year.
2. Mount it loosely and orient it in a way that the switch is not in line with the most likely collision direction.
3. If a breaker has tripped several times, replace it. Especially do this if you have many matches left to play.
4. Don't use the breaker as a switch. These breakers aren't switch rated, and by using as a switch it subsequently is easier to trip. Turn the robot off by unplugging the battery cable.

Are you talking about the main 60 A breaker or the mini-30A breakers?

BTW, if people are worrying about shock to their breakers, why not use the vibration isolation mounts provided by Small parts to mount them? (Or equivalent product available from this year's legitimate supplier)

Last year, we routinely tripped our 30A breakers on our main drive motor. However, that occurred due to a design flaw in our gear box. The tolerances were too tight and we were machining grooves in our rotating parts. Once diagnosed and fixed (in the off season), we have beaten the heck out of our robot with no breaker trips.

On our first robot (2000), we used a lag algorithm on the joystick to avoid the maximum acceleration at zero speed issue.

In 2001, we picked a gear ratio which seemed to keep us in the "sweet spot" on the torque-current curve. That robot runs cool under all circumstances. Although it is not as fast as we would like.

For future reference, we bought a bunch of 30A breakers from innovation first. Our plan is, when they start popping, replace them. Now that we know about freeze spray, we'll probably incorporate both forced cooling of the breakers (we already do this with our drive motors) and a blast of freeze spray at the start of a match.

As for driver instructions and software over-rides, I think that is a function of your drive team. In general, we prefer to give the driver the ability to push the robot as far out on the performance envelope as he feels necessary. However, we have had operators who needed some software limits to keep from breaking things.

Usually during the course of the competition, we try to customize the controls around the driver's/operator's preferences.

During the off season, we have all kinds of limits which we impose, since we use a lot of untrained drivers.

A question for the breaker gurus...
In 2002, we were running both Chalupas at full blast for the full two minutes + both drill motors at close to their limits + a globe motor + two seat motors + one fischer price motor intermittantly. We never tripped the 60 A breaker. We did mount the breaker switch perpendicular to our direction of travel.

For teams that tripped the 60A breaker, what kind of loads were you putting on your system? Was the breaker trip more shock related than electrical?

Andrew
Team 356

The year when we had to balance our robot kept tripping the 60a breaker. After several trips over the bridge, our robot wouldnt work when the bridge came down. We found out that the circuit breakers would trip when something hard smashed into the robot. SO what we did was mount the breaker with surgical tubing!

OTHER than "cooling breakers", what do you do?
 

Let's set the "cooling the breaker" thing aside for a bit...

What ELSE to you do DESIGN wise to keep from overloading the electrical system?

Examples:

1) Do you have a AHL legal slip/breakaway clutch design? Please describe it!

2) Ever done anything funny in drivetrain or software design to avoid stall currents, especially when expecting Tug of War situations? (Ex's: Enable/disable [PWM-like] algorithims to "jerk" your machine? Deliberately adding backlash or an elastic element in the drivetrain so the motor gets a "running start" when reversing?)

3) Four wheel drive users WITHOUT wheel turning: How are you making "holonomic wheels"? (A tire that slips sideways easily to keep turn scrubbing forces from blowing the breakers, while maintaining good forward/backward traction.)

Please, let's hear some DESIGN methods to avoid breaker tripping, in addition to "cooling tricks"...

Thanks!

- Keith

I'm not sure if it does a whole lot for current draw but we put a deadband in our drive joysticks. For the first bit of motion in the positive or negative Y axis directions, no signal would be sent to the PWMs. This was more to prevent strain on the drive componenets but it may have done something for current draw.

Other teams I think had a sort of "soft-power" system where power to the motors came on slowly rather than all at once. However, I am not a programmer so I don't know how that works.

Filtering Algorithms - Do they work for you??
 

That's a "low pass filter" algorithm, an excellent way to prevent sudden spikes in the power. However, the driver deep down must ACCEPT that filtering.

Last year we tried filtering, but it became an unconscious "war" between our programmer and our driver! <chuckle> Each upped their effects, until the robot was sluggish and the driver started slamming the sticks like buttons. They were BOTH bound and determined to get a specific behavior out of the robot, and they didn't agree on what that was! I even started fearing for the sticks' survival, and we didn't have any spares! :eek:

I finally asked for ALL the filters to be ripped out of the code, and pleaded with the driver to "try not to break the robot too often" with the controls... :D The robot worked fine, AND the code was simpler.

Anyone have better luck with filters?

Anyone have any OTHER good subroutines for breaker protection that work well DESPITE an "insistent driver"?

Has anyone found software "load shedding" to be useful, i.e. locking out the compressor or other "lower priority loads" whenever the drivetrain demands a panic boost? ("Scotty, we need more power for the warp drives!!!" ;) )

- Keith

Re: OTHER than "cooling breakers", what do you do?
 

I believe that you would be talking about Omni-Wheels. There are numerous threads about these. I am not sure exactly what they were called, but they should be fairly easy to find. There are also wuite a few pics in the 2002 robots gallery of teams Omni-Wheels

For various reasons, current draw being one of them, our '00 bot used a 6 wheel drive. In this setup, all 6 wheels are powered, where the middle two wheels are set about a quarter of an inch lower then the 4 corner wheels. This allowed us good climibing on the ramp (the primary reason for the 6 wheels) but also reduced scrubbing drasticly when turning.

Of course, having 6 wheels is more expensive (wieght, build time, repairs) then 4, in our case it was a fabulous solution. If you plan on lifiting loads while acending inclines, it's perfect. Also, if you can get away with a two wheel drive, Why not? The bot turns faster and uses less current, the drive line is far more simple to make and repair, and probably most important, is lighter! Naturaly, using 6 or 2 wheel drives creates all sorts of problems, but its all a trade off.


-Andy A.

Weight balance?
 

An excellent idea...

OOC, Were the wheels evenly spaced, or was the center pair nearer to one end? How was weight distributed across the wheel pairs? Did your robot rock back and forth between the outer sets of wheels when you reversed, and then run on the "back" (outer) pair each way?

- Keith

Re: Re: OTHER than "cooling breakers", what do you do?
 

I'll go check those out. Holonomic drives are pretty cool... Thanks!

BTW... Omni-Wheels are only ONE way of doing a sideslip wheel. It doesn't have to be VERY slippery side to side, only significantly MORE slippery, enough to manage the turn scrubbing forces without sacrificing too much forward and back drive traction.

Another simpler example than homebrew Omni-Wheels might be treads made from repeated slices of inverted U channel bolted to your wheel or tread. If the inverted U can be seen looking at the side of the robot (align from side to side), AND the side to bottom edges of the vertical pieces have a radius on them (rounded, so it looks like a "U" from the front), it makes for good claw traction forward and back, yet acts like a sled runner if shoved sideways. Any slide push causes it to "climb the carpet nap" and slide easily.

Any other SIMPLE methods of making "side slip wheels"?

Remember, this thread is about ANY technique to reduce breaker blowing. Not just wheels, chilling, etc...

- Keith