PDA

View Full Version : Weird issue with the power distribution block


X-Istence
02-16-2009, 09:57 PM
Hey guys,

I am mentoring for a team in Chandler, AZ and we are seeing a weird issue. Even when no fuse is inserted into the power distribution block we are able to measure battery voltage across the two terminals.

We tested with a different teams robot and they were having the same issue, so it is not limited to just our team. Has anyone else seen this? Can anyone else confirm that this is the case?

Thanks,
Bert

DonRotolo
02-16-2009, 10:02 PM
I don't have access to our robot to check, but try putting a small load on the terminals and see if the voltage doesn't go away. Even a 1k resistor should be enough. I think there are small pull-up resistors (or similar) in there causing that effect, but there is no energy behind it.

Don

samir13k
02-16-2009, 10:44 PM
i cant confirm this until tomorrow, but there is a chance it might be possible that we have the same problem. we got a faulty globe motor that burned up to the point where we could not touch it for 2 hrs. It turns out that it also burned up the victor without the distinct sound of the fuse resetting.

MrForbes
02-16-2009, 10:47 PM
What is the symptom with the robot that makes you suspect that the PD board is causing a problem?

Alan Anderson
02-17-2009, 12:58 AM
Even when no fuse is inserted into the power distribution block we are able to measure battery voltage across the two terminals.

The circuitry that was intended to detect circuit breakers tripping does have a tiny amount of leakage. While a good multimeter will detect the voltage, there isn't enough current to have a significant effect on anything.

Ziaholic
02-17-2009, 11:17 AM
Speaking as a Mentor on the "other" bot that the OP verified this with ...

The observation that started it all was that the indicator LED on a SPIKE was dimly glowing ... even without the breaker plugged in.

We weren't really reporting a "problem" per se ... we just noticed it and wanted to see how many other teams saw the same thing.

X-Istence
02-17-2009, 01:45 PM
There is enough power flowing through to also have the LED lit up on the digital side car without a fuse in, and a spike where the LED was dimly lit as was mentioned in the post above this one.

MrForbes
02-17-2009, 03:27 PM
That is probably NOT enough current to allow the wires connected to the panel to melt before the breakers open.....so the breakers can still do their job.

It is an interesting phenomenon, but I don't see that it's a problem.

X-Istence
02-17-2009, 06:33 PM
That is probably NOT enough current to allow the wires connected to the panel to melt before the breakers open.....so the breakers can still do their job.

It is an interesting phenomenon, but I don't see that it's a problem.

It is not about wires melting, but certain electronics generally don't like being undervoltaged. It was just a surprise to us to see that there was some power running through the system when we thought there was none at all.

Russ Beavis
02-17-2009, 07:00 PM
If you measure the resistance from the battery + input to the PD and any of the + WAGO outputs (with no breaker installed), you should measure 100kOhms. With a "dead short" between the + WAGO output and ground, you'll only get 12/100000 = 120 microamps. That's only enough current to start lighting an LED. Any "real" current drawn by a Victor, Spike, Jaguar or Digital Sidecar will VERY quickly "turn the LED off". You can't do anything very interesting (at least in FRC, that is) with 120 microamps other than blink an LED.

Russ

Jim E
02-17-2009, 07:13 PM
I noticed something interesting last night pushing the robot around the warehouse between the packing crate and the scale. The motors generated a voltage that caused the wireless bridge LEDs to illuminate. Other LEDs on the sidecars, and PD also illuminated.

I don't think anything was damaged, but we will find out in 3 weeks at Orlando.

Trent B
02-17-2009, 09:26 PM
Jim, probably a week and some change ago we did the same thing and even had the compact rio booting, so far nothing has acted funny with our control system so I am doubtful it would have damaged anything for you guys.

For future reference though (provided you dont use crab drive, just standard 4 wheels in linear arrangement

| |<-
| |<-

It is very easy to push the robot from the side as opposed to front or backwards, infact it is almost easier than pushing with the motors

Al Skierkiewicz
02-18-2009, 07:32 AM
When you push a robot, the motors begin to act like a generator. The speed controllers (both Victors and Jaguars) use power FETs that have a diode internal to the devices. As you generate current in the motors, that current forward biases the diodes in the FETs and since they are in a bridge configuration, power is back fed to the electrical system. There is no way you can generate enough current to back feed a full 12 volts to the system, but you can generate enough to light some LEDs and make it look like you have an electrical problem. All of this is normal.

EricVanWyk
02-18-2009, 09:14 AM
When you push a robot, the motors begin to act like a generator. The speed controllers (both Victors and Jaguars) use power FETs that have a diode internal to the devices. As you generate current in the motors, that current forward biases the diodes in the FETs and since they are in a bridge configuration, power is back fed to the electrical system. There is no way you can generate enough current to back feed a full 12 volts to the system, but you can generate enough to light some LEDs and make it look like you have an electrical problem. All of this is normal.

Actually, it is possible to generate more than 12V - you just have to back drive the motors faster than the 12V speed. It is hard to do this with drive motors, but possible if you have a manipulator that is heavily geared down or a bunch of students.

Another thing to consider is that when the speed controllers are being back driven through their diodes, their ability to handle current goes way down. These diodes are usually about a volt (they are worse than intentional diodes), where as when they are turned on they are usually a couple milliohms. Consider that at 10A you are dissipating 10 W instead of .1W in those FETs, and that the fan isn't on.

I have seen components destroyed this way, but it took a lot of physical effort - several students pushing a robot geared for torque (not speed) at running speed to see what happens. They got sweaty, the electronics got unhappy.

In conclusion, it is normal but not recommended.

Russ Beavis
02-18-2009, 09:39 AM
Simple physics - where does the energy go? It's certainly not recharging the battery since the 120A main breaker is open. So... something is going to get warm and/or generate light (kinda like a different type of "warm" actually). At some power level (energy per unit time), that extra heat will cause a failure.

Try pushing a robot that's powered on (with actuators disabled for obvious safety reasons). Then try pushing a robot with the main breaker open. Any difference? There probably will be since more of the energy will be used to recharge the battery.

Good lessons to learn. But, to Eric's point, be gentle to avoid harming your robot. And never play around with enabled robots.

Russ

Al Skierkiewicz
02-18-2009, 11:03 AM
Just so that we don't confuse anyone, the diodes are speced at 1V@20 amps and 1.25V@40 amps each when forward biased. So even if you could drive the motors at full RPM (highly doubtful) then the diodes would drop at least 2 volts in the speed controller. Remember that the bridge would have two diodes in series with the power supply input pins plus the 0.0005 ohm resistor in the Jaguar negative lead plus the wiring resistance. It is unlikely that you could provide enough power to light up the power convertors for the Crio but you might. Just not long enough for any of the system to boot.

ZInventor
02-18-2009, 11:51 AM
When you push a robot, the motors begin to act like a generator. The speed controllers (both Victors and Jaguars) use power FETs that have a diode internal to the devices. As you generate current in the motors, that current forward biases the diodes in the FETs and since they are in a bridge configuration, power is back fed to the electrical system. There is no way you can generate enough current to back feed a full 12 volts to the system, but you can generate enough to light some LEDs and make it look like you have an electrical problem. All of this is normal.

odd, we've gotten ALL SYSTEMS to boot up while pushing the bot!

by that i mean, we had connection between the bot and DS without the battery being plugged in! the station read 12.24 volts at peak (which lasted about a minute)

needless to say, i was exhausted after pushing the robot that long.

-Z

sparkscarz
03-08-2009, 06:32 PM
nice
!