Victor reversed wiring...

[Al Skierkiewicz]

Mark,
You are extremely lucky. If this was me, I would be ordering some Victors as replacements. There is no protection built in. What is working today has been stressed to the point of near failure and will likely take a hike in the future.

[fox46]

I expect there is back EMF protection built into the controllers which protects the controller when you cease applying power to a motor and the motor turns into a generator, backfeeding into the controller. If the controller was in brake mode it would have shorted the connection and could have damaged the brake mode functionality of the controller. Otherwise I would expect them to be equipped with some sort of supressors or diodes which (bearing in mind I have no clue how they are wired) could have fed back to the power input terminals? I would tend to think if the controllers work fine and they don't smell like magic smoke, you should be fine. I have seen many teams pushing their robots around while powered down and have watched various lights light up and fans spin from the current being generated and fed backwards through the controllers by the spinning motors.

[Al Skierkiewicz]

fox,
There is no protection. What you have observed (and what takes place during brake mode) is that the diodes that are a result of manufacturing in MOSFETs will conduct energy from the motors to the power supply or to a short with multiple FETs turned on. When power is supplied to the controller output, reverse currents may flow in the interface electronics and the microproc on board the controller. Many teams have reported controller failures as a result of this very same mistake just this year.

[EricVanWyk]

Originally Posted by fox46

I would tend to think if the controllers work fine and they don't smell like magic smoke, you should be fine. I have seen many teams pushing their robots around while powered down and have watched various lights light up and fans spin from the current being generated and fed backwards through the controllers by the spinning motors.

You may also notice that spontaneous electrical failures tend to happen in clusters. Some teams have had countless motor controller failures, where other teams haven't had any. The "unlucky" teams are abusing their electronics, whether it is by metal shavings, pushing their robot around unpowered, or other, and they don't realize it because the damage doesn't immediately show itself.

I don't want to sound alarmist, but you have to accept that you have caused some amount of damage to these parts. Have spares ready.

[fox46]

There is no protection

No disrespect to your expertise Al and I don't pretend to be an expert on any of this but there must be *some* sort of protection- (a "tolerance" would likely be more appropriate) for backfeeding power. Every time you cut power and the motors are spinning down they are applying power to those terminals so there must be some sort of suppression. How much "reverse" power the controllers can handle is anybody's guess but watching kit robots "skate" in coast mode and get pushed around even in brake mode would suggest to me it is at least enough to tolerate it for a little bit. Sure- extensive "pushing around" a robot would certainly damage something but the big question is how much before this happens?

In 2000 when I was in grade 10, our school used to participate in the now defunct "Canada First" robotics competition- brainchild of the late Bruce MacMillan, a spinoff of FIRST before the GTR regional came into existence. That particular year the competition organizers had designed a custom control board that when interfaced with an RC aircraft system provided control to the robots. Unfortunately the system was plagued with problems in the form of exploding MOSFETs/Transistors and the on-board microprocessor. The cause was a poor design where the designer had overlooked this kind of protection. Voltages which backfed to the controller when the power was turned off (so we were told) was causing the processor to cook itself- just like you describe. The motors we were supplied with were windshield wiper motors with very little "spindown" and were virtually impossible to backdrive which meant only a very small amount of power was needed to cause this condition. This is why I am convinced the Victors and Jags have at least some tolerance to reverse voltages otherwise I am confident they too would cook. The solution to these issues that was implemented AT the competition was replacing all the power switching hardware on the boards with relays resulting in only "ON/OFF/ON" control.

[Joe Ross]

Originally Posted by fox46

Sure- extensive "pushing around" a robot would certainly damage something but the big question is how much before this happens?

Hooking a battery up would be far worse then extensive "pushing around".

[fox46]

Hooking a battery up would be far worse then extensive "pushing around".

And yet....

they appear to still work fine

[kevin.li.rit]

Well if you're hooking up the battery where the motor goes then the current will flow through the motor (and possible backwards to the controllers) if one of the mosfets from the top and the opposite from the bottom is on. If both bottom or both top mosfets are on, such as in brake mode, then you'd be shorting out the battery.

[Al Skierkiewicz]

Fox,
Unfortunately the answer is NO, there is nothing in the way of backward protection. The FETs have a diode that is across the terminals that results from the manufacturing process. If you look at the spec sheet you can see the diode and the direction in which it conducts. The four legs of the "H" bridge do conduct as a full wave bridge and provide power of the correct polarity at a reduced voltage backwards to the power rails of the controller. These diodes do have a forward voltage drop though, so the conducted voltage isn't quite what you would think. Remember that when the controllers are powered correctly, and a 'coast' condition is commanded, all of the FETs are off and current only flows through the diodes (the same is true for pushing with power off). When a 'brake' is commanded both low side legs of the controller are turned 'on' resulting in a short across the motor but no power being delivered to the controller internals and no current flowing through the diodes. In the case of the power leads being attached to the output terminals, we don't know if they were wired in a normal polarity. In any case, considerable current may flow through the controller restricted only by the breaker feeding the device. Depending on the duration (or repeated attempts to determine the cause) significant damage to the device might be the result. Magic smoke is almost always the result.

[pfreivald]

Originally Posted by Al Skierkiewicz

Depending on the duration (or repeated attempts to determine the cause) significant damage to the device might be the result. Magic smoke is almost always the result.

The fun part about this is that robots get pushed around all the time... on the field, by other robots. I have found thus far that, while there is no good reason that I know of for this to be true, Victors used on the drive train tend to hold up well, while Jaguars used on the drive train need to be treated as consumables on a yearly basis (or so).

That's entirely anecdotal, mind.

[MrBasse]

If the wires were truley just connected and then powered up on the motor side they should actually be fine as long as no commands for driving were sent to the controller. Not saying there isn't any chance that damage occured but without activating the speed controller, there should be no real concern. Our lead mentor uses Victors frequently and his response to this situation was to shrug his shoulders and say "it's fine"

[fox46]

Al- I've never seen the schematic and doubt that I could fully interpret it if I did so I'll concede there is no specific engineered protection in the design. That being said, I still believe there is an inherent tolerance to this condition in the design of the controllers, whether intended by the designer or not.

I would run them during testing and so long as everything works fine, run them in competition. Just be weary that if you begin to have trouble at the competition that there is a possibility it is due to a damaged controller.

[Al Skierkiewicz]

Fox et al,
It is my firm belief that if you want to play on Einstein, you must have good engineering practice. That being said you believe what experience tells you and don't leave anything to chance. You don't wait for a failure to occur, you anticipate and replace when needed. It is for this reason we solder our crimp connections, keep spares handy and plan for easy replacement of everything in the field. Murphy is alive and well, why encourage him?

[Kevin Sevcik]

I'm going to second that the Victors probably aren't too badly hurt by this, depending on how long the were run like this. If the motor wasn't under load, it was mostly running at the free-speed current of a couple of amps. I don't have a schematic for the Vics, but everyone thinks Jags are wimpier, so I'll start there...

One of the FETs has a 4mOhms on resistance, so it'd be dissipating about 1.6W under normal conditions. The diode has a 1.0V drop, so it could handle 1.6A and generate the same thermal load. Which means you could run 3.2A backwards through the Jag and it'd be about the same as a 40A load normally. So I think the vics would be relatively fine if you hooked them up backwards and didn't try to run them, provided your motor wasn't under a significant load.

EDIT:That said, Eric and Al are right that it's no longer an entirely trustworthy part once you've taken it out of spec in some fashion. If I were a cash-strapped team I'd use it and try to be prepared for it failing on me. If I had the opportunity to replace it, I'd permanently brand it questionable and slap it on the practice bot.

[Shaun Coon]

i did this with 2 BLACK JAGS this year and they work on and off so we arent using them ny more. THEY are now very prone to act up so keep your eve on them .