Today, whilst doing drive train testing I noticed one wheel was not spinning. Went and checked the jaguars and found the jag controlling that wheel to be off. No light, no smoke. Upon closer examination, the jaguar was emitting a distinct “tick”, and each “tick”, the fan would twitch. The jag “ticked” about once per second. I also noticed that the positive motor terminal had melted the plastic surrounding it. The robot was only under a moderate load.
Language: Java
Controller: Black Jaguar (PWM) 2012 KOP model.
DriveTrain: Direct Drive
It’s highly unlikely, the control board is wood, and they had covers up until we started testing them. I also did not see anything in it when i had it open, or any scorching on the pcb.
Although I am not for sure, the ticking you may be hearing is the Jags built-in overload thermal reset. Check your supplied current and make sure it falls within the specs of the jag!
The jag was on a 40 amp snap action breaker, and none of the wires felt hot. We let the jag sit for about 15 minutes and tested it again, same thing, tick and twitch.
What do you mean by Direct Drive; motor directly connected to the wheel?
Did you notice it going into the Current Fault mode (Blinking amber-red with at couple second cutout) when you were driving around?
Please post more details like what motor you were using, gear ratios, etc.
Carefully unscrew the case and open up the Jaguar and look for any more visible damage. Pop out the PCB gently. The fan cable is connected via a connector and you risk breaking the fan if you aren’t careful.
You may also want to take a look at the Jaguar FAQ. It lists some common symptoms and what their common issue/fix is. I’m not saying that the FAQ has your exact failure mode, but you may recognize some symptoms and that can help pinpoint the cause.
The FAQ had nothing… As for gear ratio, I don’t know. It is a single CIM into a toughbox circa ?2010-2011? The wheel spins easily without the motor, no problem there, and with the motor it feels just like the rest of the wheels. The jag never went into thermal/over-current protection, it just turned off. I’ve had the jaguar apart, no aluminum chaff in it, no bridged traces, no scorch marks on the pcb. All of the MOSFETs look normal. I unplugged the fan to find if it was the source of the clicking, which it was not.
I almost think it’s trying to boot, but cant, since the fan twitches.
The periodic twitch could be the Power Distribution Board’s circuit breaker resetting and immediately cutting off again. I remember seeing symptoms something like that when the motor controller was wired with power directly to its motor outputs; might that be the case here?
Originally that’s what i though too, but no red light on PD, or breaker resetting. Also, don’t hate on me, but i did use a Anderson plug straight from the battery to test it, same symptoms…
What happens if you remove the load from the Jaguar?
If you have access to a oscilloscope, if not try a multimeter and put one lead on the 12V input, do you see anything abnormal?
Verify that the PWM cable works and is plugged in properly.
I recall an issue slightly similar to this on a tan jaguar while using CAN. What would happen was the jaguar would turn on and would flash yellow once and turn off for around 2 seconds then flash again. If I tried to drive it at all the status light would stay off indefinitely. I reflashed the firmware but the issue remained. I reflashed the bootloader ad the firmware (through JTAG in the back) and it worked fine after that. I have a full firmware dump on my other computer somewhere.
jp,
The fan is controlled by the micro proc and it in turn is powered by a series of regulators. I would remove the motor leads from the jag and then inspect for damage to the screw terminal. If the plastic is melted, it is likely that the two output terminals are shorted together. Since the Jag uses a bootstrap power supply, the output is always jumping to full output for a short period of time. I bet the symptoms are: power up, regulators come up and fan starts to spin, short on output shorts out during the bootstrap and the Jag senses the loss as a power loss and starts the sequence all over again.
It took me a while to remember, but I have seen these symptoms before.
When there is a short on the 5V rail inside the Jaguar the 5V DC/DC converter can’t come up. It tries to turn on, but it quickly shuts off due to its over-current shutoff. This shutoff is inside the DC/DC converter and has nothing to do with the software current limit in the Jaguar firmware. After the shutoff time is over, the converter tries to power on again, repeating the whole process.
The clicking you hear is the inductor changing shape due to magnetostriction* every time the converter tries to turn on.
The fan is powered by the 5V rail so it twitches at the same period as the shutdown time.
Many things can cause a short on the 5V rail, but one in particular stays even after you disconnect all the wires and clean out the debris from your Jaguar.
There are several ESD suppression diodes connected to the sensor inputs on the front of the Jaguar (added after Grey Jag and Lunacy :)). If one of the pins experiences a static shock, it shunts the discharge to ground and hopefully protects the microcontroller. The diodes can handle high voltage (several kV) for a short amount of time, but if they experience a much smaller voltage for a longer amount of time (6V for these diodes in particular), they can fail in the worst way possible; short-circuit. There is one diode in particular that is connected to the 5V rail and if it fails, you now have a short from 5V to ground.
So, what can cause this diode to see more than 6V? Debris is one way. It is definitely possible that a piece of debris bridged 5V and some higher voltage rail.
I’m not sure where the melted motor terminal plastic fits in to the story. It may not even be related to the failure.
You mention this Jaguar was from the 2012 KOP. You seem to have debris under control this year, but can you guarantee that debris wasn’t introduced last year? Was this Jaguar tested and guaranteed working right before adding it to your 2013 chassis and noticing the failure?
The Jag worked all of last year, and was pulled off and immediately put on this years chassis.Worked flawlessly Saturday, and died halfway through Sunday.
The mention of ESD makes me remember another student talking about static, was was complaining of it when he touched the chassis. Same student was reinstalling pwm cables after putting them in wire harnesses.
Might be onto something?
And I may have been mistaken but i think i recall a bit of water, might have just been a shadow or w/e (i doubt it was water, i’m ocd about liquid near the robot) around the brake coast,encoder, limit switch pins. Could that be a low enough resistance short to pop the diodes?
Today the same thing happened to our team and actually the real weird part is it was the same black jaguar with that same tick-tick noise.
So what happened was that we turned our robot on and we hear that tick-tick noise, so we immediately turn the robot off. Upon farther inspection we thought it might the result of a faulty wire/ PWM cable. So we remove the PWM cable and the noise goes away. Therefore we got an other PWM cable, plugged her in, and the next thing we notice, the black jaguar does up in smoke, causing the other jaguar beside it to do the same thing.
So we really dont know whats the actually problem, but we were hinting its either the PWM cables (hardly likely) or the wiring of the robot.
Even with the ESD diodes there is always a chance you can damage something. Also, water is never good. You mention it being near the encoder input… the only external tap for the 5V rail is on the encoder input.
I don’t think that a temporary short between 5V and something else would cause this issue. However, I have never tested that exact scenario.
I would do three things. First, with the Jaguar unpowered, use a multimeter (in either resistance mode or continuity mode) to look for a short between 5V and ground. Connect probes between the “+” and “-” pins of the encoder input.
Second, I would look at the 5V rail on an oscilloscope with the Jaguar powered and clicking. If you can, take a screen shot of the waveform and post it here.
Third, take a look at the inside of the bottom case plate. There is a hole that is covered by a sticker. That sticker acts as a great debris evidence collector. Any shiny specs, no matter how tiny, indicate that there is/was debris in the Jaguar. You can never predict when debris will cause damage. It may work an entire season and fail after you’ve moved it around and put it on another robot.
So, today i metered the 5v and 0v pins on the encoder out, and got 180 ohms, not good. A mentor also had their O-scope there, and here is the snapshot of the 5v pin. 1 Volt/division, 3.2 volts total.