MicroControllers at competitions

[RyanCahoon]

Originally Posted by Ether

Depends on how you do it I would think. What rule did you have in mind?

Thanks for catching my error (again ). I've edited my original post. Contrary to the wording of my original post, I was only thinking of the general "custom circuits can't affect power pathways" rule and forgot the exception for low-impedance current monitoring. Presumably (0.075 volts) / (200 amperes) = 0.000375 ohms is low impedance...

Originally Posted by Ether

Yours is 200 amps at 0.075 volts (shunt)
Al's is 125 amps at 12 volts (load resistor)

Oh that makes sense. I was reading the 12V as a safety rating or something along those lines.

Originally Posted by magnets

Your multimeter (should) have a fuse to protect itself from current, so you should be all set.

Well it would be good to check the rating on the multimeter first. Not worth blowing unnecessary fuses. Also make sure that the probe leads are large enough to handle the current as well.

[aldaeron]

Originally Posted by magnets

Here's a 400 amp dc one for $50.

http://www.sears.com/craftsman-digit...p-03482369000P

I own one of these and used it last season for measuring shooter motor current. Works great and is easy to attach to various motors on your robot

-matto-

[yash101]

Sorry. I just got home from school. I don't have one yet, but I was thinking of a very small shunt. The main point of this would be to find shorts. Also, I don't think it is illegal. You can have things between motors, as long as they don't change the power. In this case, the power difference would be so minuscule, that it would just be ignored. I wanted to use Ohm's law to get a precise measurement of a high amperage. For that, I wanted to use the MCP3204 ADC to measure the voltage drop, and plug that into the formula, to get the exact amp draw. Last year, that would have been useful because we had no idea how to check what was shorted, so we had to use trial and error. It would have told that our motor controller was shorted (though we could tell because of hot the wires were getting!)

[EricH]

Originally Posted by yash101

Sorry. I just got home from school. I don't have one yet, but I was thinking of a very small shunt. The main point of this would be to find shorts. Also, I don't think it is illegal. You can have things between motors, as long as they don't change the power. In this case, the power difference would be so minuscule, that it would just be ignored.

Ahem. Please Read the Rules Carefully.

You're almost correct, but you're missing one minor detail: It's power pathways that can't be changed (R54, 2013 rules)--if changing the power was all, there's a number of things that could probably be really bad that could happen. Except, as noted, for low-impedance current monitoring, such that the effect on outputs is inconsequential. Hmmm... how hard would it be to get a current sensor, wire it into the robot's own control system, and have a way to store and extract data from it? Hmmm... Must think about that one for a few minutes.

Yep. Totally possible, and plausible. You're the programmer, you get to figure out how to pop up the current info for any motor at any point during a match and/or log it for future reference. (Hint: I know it CAN be done.)

[yash101]

It can be done. Quite easily, actually. However, using an MCU is a cheap and dirty way to get it done quickly. You can code an MCU much faster than you can code the cRIO!
Also, the current pathway isn't being changed. It is just being passively monitored using sigma-delta circuitry!

[MikeE]

Originally Posted by RyanCahoon

Al,
What's the difference between the device you linked and this one: http://www.amazon.com/Amico-current-.../dp/B008SOXFZC ?

Off topic but I can't help but chuckle at Amazon's shipping info tag "Ordering for Christmas?"

I can just imagine the little 'uns eagerly tearing open the wrapping paper under the tree, hoping after hope that they were on Santa's nice list...

[MikeE]

Originally Posted by yash101

It can be done. Quite easily, actually. However, using an MCU is a cheap and dirty way to get it done quickly. You can code an MCU much faster than you can code the cRIO!
Also, the current pathway isn't being changed. It is just being passively monitored using sigma-delta circuitry!

Dev - there is a substantial hint (by Eric two posts above mine) that there's already a microcontroller-based current monitoring solution widely used by FRC teams.

By all means think through how you would solve the problem yourself - in fact it's encouraged - but also check into existing solutions.
The first step in any research project is to survey the literature, i.e. read around first.

[Al Skierkiewicz]

As Ether has pointed out above, wire is a known resistance for length and can be easily monitored in much the same way as you propose for your shunt. A 1 foot piece of #10 wire will drop 0.1 volts at 100 amps. Inserting a shunt resistor in the pathway (wiring) feeding a speed controller does violate the rules as Eric pointed out above. The Jaguar has a resistor built into the input side of the controller to generate the current reading available on the CAN buss.
The clamp on ammeters may give you a false reading do to the switching load of an operating controller. Look at frequency response for the device you are considering or using. For most applications the switching frequency of the Victor is within spec while the Jaguar is not. Your mileage will vary.
Magnets, it would help if we knew what your team number is when you refer to team history.
A good method for diagnosing shorted device problems is simply to pull all the breakers except for the the DSC and then add them back one at a time until you find the offending device.

[yash101]

I was thinking of using this method because it seems simple, and easy to calculate. E=I/R; I=E/R
That means that the voltage reading you get will be equal to the amperage/the shunt resistance. I want a low vRef to help me get a much better resolution! This should give me the exact value.
Anyways, you can measure the voltage drop and you know the resistance, so I=E/R

If you are measuring a 1 volt drop, and your resistance is .25 ohms, you will have: 1/.4, or 4 amps going through the circuit. Say you have a .025 ohm resistor (much more likely in FIRST), and you have a 1.024 volt drop, you have 1.024/.025, or 40.96 amps going through the circuit.
Now, say:
R(shunt)=.025ohm
V(drop)=4v096 (4.096 volts)
4.096/.025=163.84 Amps running through the circuit.
If you see that constantly, you know that there is a fault, and you know to fix the fault. I'd say, have code ready to put on the cRIO that cycles through your robot's motors at 100%. You can write down each current value, and if it seems excessive, you know to change the motor AND the motor controller!

MAKE SURE YOU CHANGE BOTH THE MOTOR AND THE MOTOR CONTROLLER. IF YOUR MOTOR IS BY CHANCE BAD, YOU CONTROLLER MAY DIE. VICE VERSA! (I think that the caps were required

[Ether]

Originally Posted by yash101

Now, say:
R(shunt)=.025ohm

The impact of an added 0.025 ohms series resistance on the motor power is not insignificant.

If you don't know how to do the calculation for that, ask and people will be glad to help you.

[yash101]

Yeah. That's the main point: To get a reading without having a significant impact. I think something like this is called a "sense resistor"?

[Dad1279]

Instead of a shunt resistor, Honeywell makes sensors that don't interrupt the circuit (senses current through wire like a clamp-on ammeter)

one example is: http://www.digikey.com/product-searc...=0&pageSize=25

[Ether]

Originally Posted by yash101

Yeah. That's the main point

Maybe it's just me, but that sure didn't seem like the main point of your post. It sounded like you thought 0.025 ohms would be a good value to use.

[Al Skierkiewicz]

Dev,
To put this in perspective, 0.025 ohms is equivalent to 25 feet of #10 wire or 12 feet of #12. It is also twice the internal resistance of the battery. That steals a lot from the motor(s) you are monitoring. So once you are drawing enough current to drop 4 volts at the resistor you will be dropping 2 volts across the internal resistance of the battery. At 5.5 volts, the regulator in the DSC gives out and stops producing PWM to the speed controllers.

[yash101]

.025 ohms was just a practical example. In real time, I would use a 25mOhm resistor, and a 1v Vref on the MCP3204. That should reduce the drop significantly, without getting rid of the precision!
What do you guys think of this


Also, about the inductance-based current sensors, don't you need a circuit to constantly monitor it if you are using DC? Won't only the changes in amperage be shown? This still seems to be quite a simple way of getting a very accurate read of the current consumption! I think that this was of induction-based current sensing is relative, not absolute. This is what I learned when I went job shadowing at SRP. They use donuts, secondary coils for the wire, allowing them to read how much current is passing. Maybe I could do an experiment on a compass near a weak magnet, and this and see if the deflection changes as more power goes through!