Question on Jaguar Coasting Specifics

I just thought of an interesting way we can spice up our drive system, but I’m not completely sure of how coasting works in Jaguars. Does the Jaguar allow the motor to rotate at a speed higher than the input voltage would cause at ALL times, or does the Jaguar still lock down rotational speed when power is applied? More specifically, if I’m giving a Jaguar an input of 0 (domain is -1.0 to 1.0), does the Jaguar allow the wheel to spin freely?

The reason why I’m concerned with this is because we’re running an omnidrive this year. It works fine now without coasting, but I’m wondering if we have coasting on, can we run around the field without an adverse effect to our pushing resistance?

On another note, is there a way to engage coasting on the fly? As in, there’s a three-pin connector that controls it, so is there any way to hook up a connector that by default jumps one pin to ground, but when it receives voltage (through a relay I guess) it jumps the other pin? I have a manual switch like that for my computer to reset CMOS when I’m trying some overclocks, heh, just don’t know if there’s something out there that can actuate via the sidecar or maybe the cRIO breakouts.

datasheet





Your first question isn’t very clear, but the answer to the second is a resounding yes.

The distinction between coast and brake lies in the switching of the H-bridge when signal is neutral; there is no effect when power is being applied to the motor. When the speed controller is set to coast, there is no connection between the motor terminals, which means that the only force reducing the motor’s momentum is friction. When the speed controller is set to brake, it shorts across the motor terminals, turning the motor into a generator that powers the motor against itself; that is, the power generated by a motor turning in one direction will make it turn in the other direction. The brake setting makes use of this power to “brake” the motor.

Coasting just means you’ll have less control when decelerating. Omniwheels inherently have a low CoF compared to most other drivetrains, so any pushing resistance you’ll have is negligible.

This has been confirmed to be legal by the GDC:

Q. On the victor and jaguar speed controls can we dynamically during game play change the brake/coast setting by use of a logic/mechanical switch controlled by the drivers?
A. There are no Rules prohibiting this.

That being said, there are many ways you could do this. One is through the use of one of the KoP provided limit switches. You just need an SPDT switch of some sort, where the common is tied to the center pin, and the switch to the coast/brake pins.

So I can definitely find a switch, I’m clueless on how to control it remotely though. I don’t have too much experience with electronics (okay, I’ll admit it’s embarassing to me that I’ve tinkered with computers but don’t own a soldering gun :yikes: ehh…), and the only type of switch I’m really aware of are physical toggle switches. I know I can possibly use a spike or maybe the pneumatic breakout to control something by sending power, though I’m not sure what I’m looking for :confused:

edit :: would I just be using these? I still have no idea how to hook up those, though.

edit :: ah wait, I think I might now. so there’s going to be one pin ID’d as the center pin that I’ll connect to the jaguar, and two pins that’ll go to the side ones that engage either brake or coast. and there’ll be two other pins that engage the electromagnet or some other engagement structure that throws the switch. direct output from a sidecar relay output won’t do it though, or a pneumatic breakout either, right?

If you are using CAN, you can tell the Jaguar to ignore the jumper and set it to coast or brake on the fly.

If you are using CAN networking to control the Jag you can send coast/brake commands that override the jumper settings (see “datasheet” link).

Has anyone successfully used DIO to switch the coast/brake? See Table 7 attached.

*





Sorry, I was replying from a mobile device and didn’t open the link.

For the record, “Coasting” isn’t exactly electrically identical to disconnecting the leads of the motor, though it’s functionally equivalent under most operating conditions. The FETs that make up the H-bridge have body diodes that make the H-bridge a rectifier when the FETs are unpowered. A rambling explanation follows.

If the robot is turned off, the voltage generated by a turning motor feeds back through these diodes to the Jag’s V+/- terminals and can actually power up your robot. Which means there’s current flowing through the motor, which means resistance to turning.

If the robot’s turned on, there’s +12V across the Jag’s V+/- terminals. Which means the voltage generated by the motor would have to exceed (12+1.4)V before any current would flow and cause resistance. That’s highly unlikely, but theoretically possible. For instance, if you spin a shooter up to full speed then put it in coast, the motor’s generating near to 12V. If you then get in a shoving match that drops your battery voltage to 9V, the voltage generated by the shooter motor will be greater than the input voltage on the Jag. The difference in voltage means current will start flowing, slowing the motor more rapidly, and actually powering the rest of your robot slightly.

We use this method all the time when we need to control coast/brake. We even used it with the IFI controllers. There is a huge caveat for those who have not chosen this method for drive trains. If you robot has a high COG and you are moving fairly fast, engaging brake mode might cause you to tip over.
One last item, the brake mode only has an effect when moving. While stopped it will not hold you in place. On an arm or other moving robot appendage, it will not hold the device in place. Only true brakes can perform this function.

Originally Posted by Ether
Has anyone successfully used DIO to switch the coast/brake? See Table 7 attached.

I am thinking about doing something like this for our drive motors; is it as simple as wiring the digital output to the central signal pin? The jaguar has a 200kohm pull down resistance, according to the table 7 referenced by Ether. Does this work like that? I am not familiar with whether the voltage goes high when opened or closed on that signal pin. (Is it even a current sink?) How would you roughly connect that to the I/O?

It appears to be (but I’ve never tried it). See pinout. Coast is when the signal is 2 to 5 volts. Brake is when the signal is -0.3 to 1.3 volts. See Table 7.

*

brake_coast pinout.png


brake_coast pinout.png

It appears to be (but I’ve never tried it). See pinout. Coast is when the signal is 2 to 5 volts. Brake is when the signal is -0.3 to 1.3 volts. See Table 7.

Based on the Digital I/O on the sidecar, and the schematic of what is behind the coast brake pins, would you connect the brake signal pin to the + or the ground on the sidecar? It appears that the signal pin on the Jaguar likes to be grounded (correct me if I’m wrong?). I would think that the + on the digital IO sidecar is the signal that turns on and off (does the ground on the digital side car also behave like this?) If you connected the + on the sidecar and the signal pin, would the net-voltage still be “pulled high or low”

www.usfirst.orgsitesdefaultfilesuploadedFilesRobotics_ProgramsFRCGame_and_2012-03-05_08-31-31.png


www.usfirst.orgsitesdefaultfilesuploadedFilesRobotics_ProgramsFRCGame_and_2012-03-05_08-31-31.png

It appears that the signal pin on the Jaguar likes to be grounded (correct me if I’m wrong?).

Pin2 on the Jag coast/brake “likes” to be switched: Lo (-0.3 to +1.3) is brake, Hi (+2 to +5) is coast.

The GPIO signal pin goes to Pin2 on the Jag coast/brake, to switch it from Lo to Hi.

The ground pin at the GPIO goes to Pin3 (ground) on the Jag coast/brake.

That would be my guess. Is that the way your team does it, Al?

Yes that is correct. No power connection is used.
I shy away from the ‘ground’ reference though. Being an old RF guy, there is no ground on the robot, only power return.

Thanks Al.

Old habits die hard. I’ve been calling this thing “ground” ever since I started designing and building my own tube-based audio equipment in the late 60’s.

*

brake_coast pinout 'ground'.png


brake_coast pinout 'ground'.png

I know old habits die hard and those amps needed a ground. I have several Heathkit products sitting here in my room and all of them have a stud on the back.

I had to chuckle, I guess I’m in good company*: ** Just saw this and this.
*
*


sidecar GPIO.png



sidecar GPIO.png

Can you guess how hard I have been trying to get things changed?

Originally posted by Ether
The GPIO signal pin goes to Pin2 on the Jag coast/brake, to switch it from Lo to Hi.

The ground pin at the GPIO goes to Pin3 (ground) on the Jag coast/brake.

Thanks! I will test this out sometime between this week (on our old robot) and next competition in 2 weeks.

Yeah but don’t forget: all the new control system stuff was introduced back in 2009. That year the robots really needed a ground even if they didn’t actually have one. Maybe it was a bit of prescience by the designers.