Suicidal Victor 884's...
 

We had a Victor start to smoke on our prototype bot after only around 5 minutes of running around on the ramp...

We took off the fan and saw that one of the SCR's or (whatever they are) had cracked IN HALF and shorted out the two next to it, causing the plastic to melt and smoke, etc...

We were using the Bosch motors with a 30 amp fuse (the 40's were not here yet) and it was a very random event... Does anyone have any ideas as to what went wrong? Are these things still under warranty?

The power it draws was probably too much for ur 30 amp fuse box to handle...there IS a reason why you need 40 amps for motors...

But that shouldn't cause any problems. It would actually make them less likely to fail due to current.

The 30a would just trip if it was drawing too much power...

If I were using a 60a, for example, then I could see why this could have happened.

Had the victor been dropped or something? Like, was it already damaged, and it just took those 5 mins for it to heat up enough for it to smoke?

I would contact FIRST (and/or Innovation First) about this - you shouldn't have to pay for it if it came like that.

(also, running the 30A fuse w/ the 884 should ensure it NEVER gets too much current [we have 884's this year because we now have 40A fuses])

The events you describe are unusual for speed controllers unless they are hooked up backwards or shorted. The devices you saw when you removed the fan are the FET (field effect transistors) that do the switching to the motor. There are four sets of three FET in parallel. When the motor is running two banks of FETs are turned on at a time.
It is possible you had a early failure (we call it "infant death" since it occurs very early in it's life) due to defective parts but I suspect shorted wiring, metal flakes in the controller or damaged motor. Since you were using the new drill motors and there are known problems with these when handled wrong, I would check the motor and make sure the brush assy is still in good shape.
"We took off the fan and saw that one of the SCR's or (whatever they are) had cracked IN HALF and shorted out the two next to it, causing the plastic to melt and smoke, etc..." This is a normal failure mode for these devices in that at high currents, there is a lot of internal heat generated. However, one dead device probably did not kill the other two, they all went together as a team. Inserting another engineering slogan here..."They gave their life to save a fuse!"

The wiring couldn't have been backwards because it was exactly the same as the other side.

Could I assume that a short would trip the fuse before it started SMOKING?

Is it possible that it was a random defect?

Check this out

http://forums.delphiforums.com/n/mb/...jpg&frames=yes

Specifically related to Victors tearing up the FET's, one guy uses a zip tie through the hole in the FET's on the corners to insulate them from their neigbors and there is a nice pic.

Caleb,
Like I said in an earlier post, infant mortality is always a possibility. There may even be a defect in the controller. (solder splash, incorrect device, etc.)
Yes, the breaker should have been tripping, RC should have been resetting, etc. if there was a wiring error. Be sure that the controller was the cause and not the effect before powering up a new controller. A picture would be a nice thing if you could post a closeup of the damaged transistors. Most people never get a chance to see solid state destruction at it's finest.

Another possibility is this:

When the Victor tells a motor to stop, the motor, in its obstinate way complains, "I can't, I got this stupid robot on my back, but I'll try." The motor continues to generate current (Lenz' law), sending it through the Victor's FETs in a loop. This loop is independent of the upstream fusing, occuring between motor and Victor only.

IFI used to have a paper that 'bragged' about a spike of 300 A being handled by a Victor (883), when subjected to full forward, , accelerate to speed, arrrgh...no!!!!, full reverse.

Perhaps you were in brake mode, which would make it more likely.

As Al says, this could be normal mortality. I suggest you talk to IFI about it. They may replace it.

-=-=-=-=

After reading AL's last again, I must caution against drilling Aluminum over Victors. Once the controls are mounted, all machining must be done in another room, even another country, as Aluminum has an affinity for the circuit board material in Victors, and will easily swerve 20 feet to the side in falling 3 feet to the floor.

To quote a Rudyard Kipling phrase, it is due to "The Heartbreaking Perversity of Inanimate Things".

I heard a story that when IFI did the first victor speed controllers, they had everyone send the failed ones back. They received 6. 5 were dead because of metal shavings inside (team's fault). The 6th was so badly damaged that what cause the damage could not be determined.

Whether that story is true or not, I don't know.

Thanks for the help, everyone...I'll try to post a pic soon :)

The first Victors, a big improvemant on the tiny Tekins, did not have the cover of plastic in the area inside the FET quadrangle. They probably cover the circuit board better now than they used to. We turned in one in Florida, that year.

Same thing happened to our first victor, ran for a few minutes and then sparked and started smoking (also running off the 30 amp) it burst into flames before we hit the off swtich.

We ended up not neating that 4th victor, but should we try to get a replacement anyway?

the motors shouldn't be drawing more than 20amp...the bosch under normal driving only draws about 16amp. the reason they have 40amp breakers this year was so people would start designing thier bot to be strong, they want to see crazy transmitions

The reason for the 40 A breakers is to delay the onset of "Puff". When those engines start (with a Drive-it-like-you-stole-it driver), or when some Idiot tries to push the end wall and your wheels have absolute traction on the carpet, the motors aren't turning very fast, the back emf they generate is zero, and they draw the "Stalled" or "Locked rotor" current in the specifications. For how long depends on the fuse. One sheet said that when one of these motors (I forget which) was stalled, it acted like a .0999 ohm resistor, or, using Georg Simon's wisdom, it draws blessed 120 A ! ! !

Once they are moving the 'bot somewhere, and their back emf approaches the applied voltage, the current drawn becomes a lot less (like "20 A"). This is the "under normal driving conditions" state, where the fuses don't blow anyway.

And, when the load is primarily braking (sudden shift to speed = 0 on a 'bot going all ahead full), the current in the Victors is in the loop between the Victor and its motor. As stated above, this circulating current can be very high, and it doesn't flow through the breakers.

These are not TOYs, but they get the respect Rodney Dangerfield gets ! Pfui !:mad: :D

I would believe that the move to 40 amp breakers is in response to many teams complaining that the 30 amp breakers were tripping too often to allow reliable operation during a match. In robots during the past two years that were tripping 30 amp breakers, most often it was due to mechanical designs that demanded that motor be operated too far out on the power curve and robot designs that had extremely high frictions when the robot turned. In these case, even the 40 amp circuit breaker would not have changed things very much. The current draw on those robots still exceeded 80 amps per motor with peaks above that and many robots were tripping the 60 amp breaker.
All teams should be aware of the high currents that these motors can draw. Even with the 120 amp main and 40 amp motor circuit breaker allowed this year, some teams will still be tripping breakers. Each of these breakers can handle 200% of rated current continuous for only 15-20 seconds before trip. The main breaker will trip off at 200 amps after 10 seconds minimum to 30 seconds maximum. On the resettable breakers, once trip temperature has been achieved, lower currents will trip the breaker.
Having a design that calls for a gear ratio that gives extremely high speed on a flat surface will most assuredly run the drill motors to 90-100 amps on the ramp. A drive system that uses two drill motors with a low gear reduction, when in a pushing match will trip the main breaker well before the end of a match.

My guess for the reason that they included 40 amp breakers this year is that in the 2001 Season, teams constantly had issues w/ motors tripping while going up and over the teetertotter. Last year, they removed all non-horizontal obstacles, but since they are adding in another one this year, they decided to give you a little more power :)

Was this a 883 or a 884? The 884 can take the 40A, but the 883 can only take 30A

my team had a similar problem with our main breaker....
our conectors were not properly installed and it arced causing it to spark and smoke after about 30 seconds of running and drawing major current without our breaker popping. you might want to check all your connections.

884 failure
 

We just had an 884 fail. I was just trying to reset the cal to the defaults. I was holding down the cal button, we powered up the electronics and in a couple of seconds it was tossing sparks. Now it controls in one direction but not the other. I have not taken the fan off yet to survey the damage. Can these things be repaired or is that it?

It is best to check in with Innovation First on repairs on these devices. So far the arcing and sparking complaints could be metal shavings in the controllers or could be a bad run of FETs. If the later is the case the guys and Innovation First will know and will tell you what to do. Do not attempt repairs without checking with Innovation First. If you try, I would virtually guarantee no warranty coverage.

Since a bunch of people are having similar problems, I'm not sure that our problem from last year relates to this year's 884 deaths.

we had two victors (one on each drive motor) connected. We had a short through case between the M+ on one victor and the M+ on the second victor. This created a current loop which blasted one of the victors (the weakest).

I always demand that we do a continuity test between M+, M- and case before turning on any power. We slipped last year.

Anyway, that's something else to try.
Andrew
Team 356

Andrew,
When you say "short through the case" what do you mean? Did you have motor wiring accidentally shorted to the robot frame? If that is the case and you had the motors running in opposite directions I can see how that resulted in the failure. Did the Victor that failed have shorter power wiring feeding it than the one that survived?

884
 

Our team just had a Victor 884 fail. We heard a pop and then saw and smelled smoke. I'm fairly sure it wasn't a short because I couldn't hear any clicking from the 30 Amp breakers and the wires were well insulated and not close to each other. We were using a van door motor hooked up to a 30 Amp breaker. I'm going to have to take a closer look under the fan, but I don't know what could have caused this.

What should we do?

We just tested all of our Victor 884's (7 in all, 4 from the kit, 3 ordered from IFI later) on our simulator to make sure they are OK before putting them on the robot. We used a van door motor and fused it with a 20 A breaker.

All of our victors performed fine under these conditions.

Since so many people are experiencing difficulties, I would suggest doing a functional test before installing and checking all wiring on the robot for shorts to case before turning them on.

We will probably also do a 40 A breaker + drill motor functional test prior to full install, just in case.

Andrew
Team 356

This was a flukey short, but here's what it was.

One drill motor had a short between its motor terminal and our aluminum heat sink shroud which surrounded it. This shroud was mechanically (and electrically of course) connected to our aluminum frame.

On the second drill motor, we had a short between the ring connector and the mounting screw for the speed controller. The mounting screw was mechanically and electrically connected to to the same case, although on the opposite side of the robot. (Note: we now use plastic 6-32s to mount our speed controllers to the frame).

I'm not sure whether we had M+ on one motor connected to M- on the other (in which case forward would have been bad) or whether we had M+ connected to M+ (in which case opposite would have been bad).

I'm sure the length of wiring was different, but not by much. I just assume that the weakest component started to flame out first.

Andrew, Team 356

Andrew et al,
This is one of the very reasons I think inspectors should check (with an ohmmeter) for power wiring to chassis shorts when performing inspections on the robot. All teams should perform this test before "power on" when in construction phase and repeat if any wiring passes through chassis parts. (tubing, enclosed spaces, etc.) This first bit us in 1996 when wiring that was passing from one side of the robot to the other through chassis tubing was shorted after a drill and screw operation was made to the tubing. (This played havoc with the controllers at the time due to just increased noise in the ground lead.) That problem is what brought me to the team. My son, a first year student, dragged me out of the stands to help out and I have been here ever since.

We also had one blow up on our first robot test, However it was not connected to anything. We had mounted all the electronics and wired input power to them. Only our drive motors were hooked up. The speed controller had no motor connected to it at all. Power and pwm signal where there no motor and it blew all by itself. All the others were fine. We removed the circuit breakers for all the others just in case for the rest of our testing.

From this it looks like the controller suffered a shoot through failure - Where both the high side and low side FETs are on at the same time causing a dead short in the device itself. The circuit breakers will not protect for this type of problem, they take too long to respond to protect the FETs. This problem is usually caused by not allowing enough dead time between switching on and switching off a design issue in the speed controller.

Ron,
Was there a software adjustment of the deadband by any chance? (the issue with 883 vs. 884) Did you have any particular indication on the Victor. (no light, all orange, etc.) Did you try to reset to factory calibration? All of this is assuming you didn't have a flame out, of course.
I am just looking to get a track record on the failures for future reference. If anyone returns 884s to Innovation First and gets any info on the failure, please post.
Thanks,