Re: Speed Controller Design
 

Like ether said, it should be handled in software. However, if you have the opposite requirement, you need fine control at low throttle, the Jaguar is much better. See http://www.youtube.com/watch?v=_X0_aFMpm9I for a visual demonstration.

Re: Speed Controller Design
 

I saw this about a year ago. It's hard to tell for sure, but it doesn't look like there is any kind of sensor on either motor shaft to provide speed feedback. So I am assuming that this demo was done done with open-loop (voltage) control of speed ?

I wonder how the Vic and Jag would compare when using closed-loop control, and with the motor under at least some minimal load instead of free-running. Could be an entirely different story.

Re: Speed Controller Design
 

Anecdotal, but 330 has mentioned that their 2008 Arm performed better when under control by jaguars.

Re: Speed Controller Design
 

You are correct that video was unloaded with voltage control. Adam is correct that the difference was noticeable with our arm, perhaps even more so when loaded.

While we have never run our arm with velocity control, it was also noticeable with position control. It was much easier to get an over-damped response with the jaguar. Our drivers preferred it to be over-damped, they didn't like an extra oscillation even if it was faster. Because of the slop in our arm, being over-damped also gave a more accurate response, since we were always approaching from the same direction. It was also much faster to tune the PID to an acceptable response with the jaguar. Perhaps given the time for optimal tuning for both you might be able to get similar responses, but I don't think many teams have that much time in a FIRST season.

At some point, it would be interesting to repeat the same tests with a black jaguar, but I'm not sure we'd be able to do it. I do encourage other teams to perform similar tests and document the results.

Re: Speed Controller Design
 

Thanks for the detail Joe. That helps put things in perspective.

Re: Speed Controller Design
 

Ether,
My comment about the motor slowing down quickly when the speed is reduced was in reference to open-loop Voltage control.

I've opened up both Black and Tan Jaguars. (I've fried both). I noticed that the Black Jaguars don't have all 12 MOSFETs; they only have 8. I know the purpose of this was to allow room for the RS232 components, but I was wondering how this was possible (and still have the same performance)? Are they higher-quality MOSFETs with lower ON resistance or greater power dissipation?
If we take Ohm's law, we can say that the total power dissipated by a set of MOSFETs (for one leg of an H-bridge) is i*i*r/n.
i is current
n is the number of MOSFETs
r is the ON resistance of a single MOSFET

That means the power dissipatiion per MOSFET is (i*i*r/n)/n
OR r*(i/n)^2

If identical MOSFETs were used, that means the MOSFETs in the Black Jaguar dissipate 9/4 of what the ones in the Tan Jaguar do, given identical operating conditions. A more accurate way of saying that is if you took out one MOSFET from each leg of the H-bridge on a Tan Jaguar, each MOSFET would now dissipate 9/4 as much power as it did before. (If it doesn't, then it burns up)

However, it has been stated that the Tan Jaguars operate differently than the Black Jaguars. Does locked anti-phase reduce the current through the MOSFETs?

Re: Speed Controller Design
 

Tan Jag uses FDP8874

Black Jag uses FDP8441

Re: Speed Controller Design
 

OK, I can understand why a Black Jag would exhibit this behavior. Damping is an inherent characteristic of locked antiphase switching.

But what I can't make sense of (yet) is this:

Why would it draw a lot of current when you take the motor down from full speed to 10%, if you are using open-loop voltage control?

Re: Speed Controller Design
 

The reason it draws a lot of current is that instead of shorting the motor to itself, you're shorting the motor to the power supply.
I haven't actually measured the current while it's doing this. It might have something to do with the nature crowbar on the power supply. (Perhaps it's voltage controlled, not current-controlled?)

I will double-check this behavior.

Re: Speed Controller Design
 

Thanks for looking those up!
The FDP8441, the on-resistance is 0.0021 ohms and it has a power dissipation of 300W.
The FDP8874 has an on-resistance of 0.0036 ohms and a power dissipation of 110 watts.

So yes, they did use much higher quality transistors.

Re: Speed Controller Design
 

Never mind - I figured it out.

When a DC motor is free-spinning at high speed there is very little current. If you then apply a reverse voltage, the motor's back EMF adds to the applied reverse voltage and creates high current.

Re: Speed Controller Design
 

Okay, good.
I confirmed that it happens, but it only happens with Black Jaguars. On Tan Jaguars, the motor just runs down on its own.

Re: Speed Controller Design
 

That's because the Black Jags use locked antiphase switching (which applies reverse voltage during the "off" part of the duty cycle).

So, in your example, if you change the throttle command from 100% to 10%, that corresponds to a change in duty cycle from 100% to 55%. With a 55% duty cycle, you will have 45% reverse voltage with locked antiphase.