Jaguars, what can we use in FIRST 2010?

[Al Skierkiewicz]

Erik,
So as not to confuse, the Victors do not have any sensor inputs or outputs. You can use whatever you would like to monitor motor current, position, speed with whatever sensor you can think of that fits all the other rules, as long as it is external to the Victor.
Joe, another Joe, Joe Ross, is a good reference for the new control system.
As a reminder, the power wiring does contain some resistance that you should add to your calculations. #10 is about .001 ohms/ft., both leads please. (red and black) #12 is .0016/ft. To put it in perspective, a one foot length of #10 wire at 100 amps drops 0.1 volt. This is my definition of the "wire foot" of WF. Four good connections equals one WF, a Jaguar series resistance is 3.3 WF including the current sense resistor (which is in the return lead of the low side FETs). If you use the voltage sense in the Jaguar, you will be more accruate and can forget the resistance between the PD and Jaguar. You will need to add in the resistance between the Jag and motor. Some motors have internal thermal limits that may add some resistance as well as changing resistance with brush wear in. I suggest you mount the Jag close enough to the motor you are controlling that you can use the wiring that is attached to connect directly to the Jag (CIM motor) or as small a connection as possible for the other motors.

[Joe Johnson]

Originally Posted by Al Skierkiewicz

<snip>
As a reminder, the power wiring does contain some resistance that you should add to your calculations.
<snip>

Yes you are right, the voltage on the motor will not be as simple to calculate as it might be (e.g. battery voltage X PWM duty cycle) because there is some wiring resistance to factor in. But even with this, given the current, it is still pretty straight forward as compared to the pain and suffering associated with getting a sensor out on the end of an arm or some other inconvenient spot.

As to the particulars, it is a pretty simple thing to calibrate: command a known duty cycle, read the current from the Jaguar, read the battery voltage, read the motor terminal voltage, calc a wiring voltage drop, use V=IR to estimate the resistance of the wiring.

For extra credit, repeat this at several duty cycles & loads while approximating (or measuring if you can) the actual speed of the motor. A simple curve fit later and you can know motor speed for that motor in real time without having any other sensor than the Jaguar. Pretty nifty.

For double bonus you can even try to model motor temperature...

...but I don't suppose it's worth it. After all, I am just trying to keep my ball possessor rollers going at a constant speed. Let's not get crazy here... ;-)

Cheers,
Joe J.

[Al Skierkiewicz]

Joe,
Your response triggered something else I had forgotten. The Jaguars have a much higher switching frequency than the Victors (15kHz vs 150 Hz for the Victors). As such the inductance of the motors play into the rise times of the pulses supplied to the motor. This effect is particularly nasty at low throttle values but the effect is a more linear response of throttle command to motor speed than in the Victors. Due to the effects of the inductance, small motor pulses cannot reach full battery voltage during the pulse width.

[Joe Johnson]

Originally Posted by Al Skierkiewicz

Joe,
Your response triggered something else I had forgotten. The Jaguars have a much higher switching frequency than the Victors (15kHz vs 15o Hz for the Victors). As such the inductance of the motors play into the rise times of the pulses supplied to the motor. This effect is particularly nasty at low throttle values but the effect is a more linear response of throttle command to motor speed than in the Victors. Due to the effects of the inductance, small motor pulses cannot reach full battery voltage during the pulse width.

I've seen this a number of times in my automotive experience. We get a door working at 100 or 400Hz PWM frequency, then in the middle of a program, someone decides that that hum has just got to go... ...so we switch the PWM frequency to 8kHz or 10kHz or in one case 20kHz, but the system inevitably has some flaky behavior. The culprit is usually the effect you discuss.

Essentially, changing to the higher PWM frequency screws up all the motor calibrations for low duty cycles. A 10% duty cycle at 100Hz is a 10msec pulse which is FOREVER for these small PM motors so they get up to "full speed" in less a single pulse. Not true for a 10% duty cycle at 10kHz (= a .01msec pulse).

On the ups side, velocity control at low loads usually improves (for the same reasons only in reverse).

It's all good...

Joe J.