PID on Jaguars is Illegal

[Stuart]

Second on Kevin's proposal.

[Alan Anderson]

Looks good. I suggest the following minor change:

Originally Posted by Kevin Sevcik

2. ...Reasonable interpretation would read this as the cRIO sending a position/velocity COMMAND to the Jaguar, and the Jaguar translating this into an the appropriate voltage command control to the motor...

[drakesword]

With saying that the jag in closed loop acts just like a servo wouldn't that make jags and victors illegal?

Both don't just change the voltage they switch it on and off very fast the uC and supporting circuitry fiddle with the transistors to switch the output on and off and they generate the commands themselves.

Does that mean we need to revert to spikes? I dont think they gave us enough with the kit!

[andreboos]

Originally Posted by Kevin Sevcik

... this would appear to be absolutely identical to the functioning of the perfectly legal RC servos, which implement closed loop position control of a DC motor based on potentiometer feedback and a cRIO position command.

A very good point. However, I think the distinction may be that the firmware for the Jaguars is to some degree mutable, while I believe servo control is implemented purely in hardware.

[Kevin Sevcik]

Drake,

My implication was that a jag in closed loop is highly equivalent to an RC servo. Just with external motor and sensor, and tunable PID parameters. So if a closed loop Jag is illegal, then..... The intention is to point out that we're already using independent position controllers on the robot in the form of servos.

All,

I'm informed that some rather more appropriate people are working on this issue now, so I'm going to hold off on any new posts to the Q&A for a little while. I'd like to see what transpires over the weekend before adding to the GDC's busy schedule. I'd ask that anyone else please hold off until Monday evening as well, so as to avoid complicating the situation.

[Kevin Sevcik]

Originally Posted by andreboos

A very good point. However, I think the distinction may be that the firmware for the Jaguars is to some degree mutable, while I believe servo control is implemented purely in hardware.

Except that only holds for analog servos. Digital servos implement their control in software, and some of them are even customizable/programmable in various manners.

[MikeE]

It's not unexpected that there is a review of this decision behind the scenes. As well as many volunteers, WPI, NI and particularly TI have invested in expanding the use of the Jaguars. And I thank them for that investment.

On a more general point, section 4.3 of the Game Manual includes the exhortation:

When reading these rules, please use technical common sense (engineering thinking) rather than “lawyering” the interpretation and splitting hairs over the precise wording in an attempt to find loopholes. Try to understand the reasoning behind a rule.

That really should be a two way street, and while the Blue Boxes have helped put context to many rules, this ruling appears to be a counter-example where over-judicious reading of the letter of a rule is elevated over the spirit.

[MikeE]

Originally Posted by Kevin Sevcik

Per GDC Q&A here:
http://forums.usfirst.org/showthread.php?t=16326



Plenty of time later for commentary on this lovely ruling, just wanted to get the word out before people got too invested in tuning and/or ditching PID on the cRIO.

There seems to be a Booth Review in progress. Or CD complaints caused a Challenge Flag to be thrown.

The original link is no longer viewable and the answer no longer appears in the Q&A Robot section http://forums.usfirst.org/forumdisplay.php?f=1481

[jhersh]

Woohoo... Closed loop control on CAN is legal!!

Update 4

[Alexander Meyer]

Originally Posted by jhersh

Woohoo... Closed loop control on CAN is legal!!

Update 4

{fistpump}

I can't tell you how relieved I am..

[klmx30302]

c/p from team update #4:
As long as the CAN bus is wired legally so that the heartbeat from the cRIO is maintained, the closed loop control features of the Jaguar motor controller may be used. (That is, commands originating from the cRIO to configure, enable, and specify an operating point for all Jaguar closed loop modes fit the intent of <R49>.)

(our software subteam breathes a sigh of relief.)

[zbanks]

I'm glad it's legal. I was looking forward to using it this year.

Of course, after reading this thread earlier today, I spent about an hour abstracting and reworking the code.

Although it's still improved, I wish I had known earlier...

[MikeE]

It's a relief to see the ruling reversed, but a pity the decision was published in the official Q&A in the first place.

Thanks to the positive contributors to this thread for raising awareness and being a part of encouraging a sensible resolution.