[EVDL List]

When a controller fails so it acts as a short between the batteries
and the motor, what is actually happening? I was discussing this with
a fellow EVer, but I'm at a loss to understand the actual process. In
the case of a Curtis, there are many paralleled MOSFETs - is the
failure of one of these enough to act as a strong short?

Intuitively to me (whose practical experience with high power
electronics ends at the basics of how my EV works) a short across a
single mosfet, even if rated at 30A would blow the MOSFET before much
happened. Is it a cascade failure of some sort or can a single MOSFET
conduct well enough to cause trouble?

Thanks,

Erik

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev

 

[EVDL List]

[No message]

 

[EVDL List]

[No message]

 

[EVDL List]

Dewey, Jody R ATC COMNAVAIRLANT, N422G5G <[email protected]> wrote:
> Depends on the "Type" of MOSFET. A Depletion Mode mosfet is already
> shorted. The source and drain sections are connected for current flow.
> You must block the current by putting an opposite current flow in on the
> gate and push the electrons out of the channel.
>
> An enhancement mode mosfet can be shorted, in fact, to get it to work,
> you must create a channel by applying the bias on the gate TO get it
> shorted so current can flow.

That's how they act during normal operation, not when the device is
damaged or destroyed.

All power MOSFETs used in EVs will probably be enhancement mode MOSFETs.

However, even enhancement mode MOSFETs can fail as a short circuit. In
some failures, the resulting device acts like all three pins (gate,
source, and drain) shorted together, which causes lots of problems for
the rest of the circuit. This happens when you get too much voltage,
current, or temperature, breaking the MOSFET and making it
electronically act like a short circuit, not a MOSFET.

-Morgan LaMoore

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev

 

[EVDL List]

>fellow EVer, but I'm at a loss to understand the actual process. In
>the case of a Curtis, there are many paralleled MOSFETs - is the

You can get several failure mechanisms within the MOSFET. A punctured
gate oxide layer can do this. Simple avalanche (overvoltage)
can do this as well.

An overvoltage at the gate will puncture the gate oxide layer -
this can occur because of ESD, or parasitic inductance (typically,
Miller capacitance and lead inductance). Normally Miller tends to
slow the MOSFET down, but in the case of multiple paralleled MOSFETs,
things are not as they always appear.

An overcurrent can cause the MOSFET to fail because of (very fast)
thermal buildup, and also in the case of paralleled MOSFETs, if they
don't switch at exactly the same time, a surge current violation can
occur, and it'll blow up, too.

Thermal cycling is an issue, but this usually causes a bond wire
to fatigue first. Unfortunately, a bad connection here will overheat
the die and it'll short, yet again. Common theme here.

Oscillation, which can happen fairly easily in paralleled MOSFETs,
will cause the die to overheat. In all cases,

Avalanche ratings of MOSFET's are specified, and many older ones do
not have particularly good numbers. Some of the newer ones are quite
impressive. MOSFET manufacturers realize that building in some
toughness really helps the designers out and makes for a more
rugged product.

Often the high voltage on the drain will hit the gate, frying the
gate driver circuit, and any other MOSFET's that are sharing
the same driver. A chain reaction then can occur.

Sometimes, only a single FET fails, and can be removed from the circuit
with the rest of the MOSFET's still working. Although what killed the
first FET is about to do the same with the remaining ones anyways.

-Dale

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev

 

[EVDL List]

It won't be a really low resistance short, and at these
voltages/currents I'd generally expect it to create enough heat to
explode immediately.

Danny

Morgan LaMoore wrote:

>When a MOSFET "blows up", it becomes a short circuit, permenantly
>conducting electricity. That means the motor negative is permenantly
>connected to the battery negative through the dead MOSFETs.
>
>-Morgan LaMoore
>
>On 9/6/07, Sarah & Erik <[email protected]> wrote:
>
>
>>When a controller fails so it acts as a short between the batteries
>>and the motor, what is actually happening? I was discussing this with
>>a fellow EVer, but I'm at a loss to understand the actual process. In
>>the case of a Curtis, there are many paralleled MOSFETs - is the
>>failure of one of these enough to act as a strong short?
>>
>>Intuitively to me (whose practical experience with high power
>>electronics ends at the basics of how my EV works) a short across a
>>single mosfet, even if rated at 30A would blow the MOSFET before much
>>happened. Is it a cascade failure of some sort or can a single MOSFET
>>conduct well enough to cause trouble?
>>
>>Thanks,
>>
>>Erik
>>
>>_______________________________________________
>>For subscription options, see
>>http://lists.sjsu.edu/mailman/listinfo/ev
>>
>>
>>
>
>_______________________________________________
>For subscription options, see
>http://lists.sjsu.edu/mailman/listinfo/ev
>
>
>

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev

 

[EVDL List]

This is what I was expecting, but I don't understand then how the
controller can power away unless there is a cascading failure and many
MOSFETs short.

Erik
--------------------------------------------
It won't be a really low resistance short, and at these
voltages/currents I'd generally expect it to create enough heat to
explode immediately.

Danny

Morgan LaMoore wrote:

>When a MOSFET "blows up", it becomes a short circuit, permenantly
>conducting electricity. That means the motor negative is permenantly
>connected to the battery negative through the dead MOSFETs.
>
>-Morgan LaMoore
>
>On 9/6/07, Sarah & Erik <[email protected]> wrote:
>
>
>>When a controller fails so it acts as a short between the batteries
>>and the motor, what is actually happening? I was discussing this with
>>a fellow EVer, but I'm at a loss to understand the actual process. In
>>the case of a Curtis, there are many paralleled MOSFETs - is the
>>failure of one of these enough to act as a strong short?
>>
>>Intuitively to me (whose practical experience with high power
>>electronics ends at the basics of how my EV works) a short across a
>>single mosfet, even if rated at 30A would blow the MOSFET before much
>>happened. Is it a cascade failure of some sort or can a single MOSFET
>>conduct well enough to cause trouble?
>>
>>Thanks,
>>
>>Erik
>>
>>_______________________________________________
>>For subscription options, see
>>http://lists.sjsu.edu/mailman/listinfo/ev
>>
>>
>>
>
>_______________________________________________
>For subscription options, see
>http://lists.sjsu.edu/mailman/listinfo/ev
>
>
>

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev

 

[EVDL List]

Sarah & Erik wrote:
> When a controller fails so it acts as a short between the batteries
> and the motor, what is actually happening?

Semiconductors normally fail shorted. So obviously, if you have a
controller with just one or a few large transistors, and one fails, the
motor is stuck "full on". The leads on these big transistors are big
enough to carry high currents for quite a while. So the motor takes off
at full power until something else stops it (a fuse, circuit breaker,
action by the driver, etc.)

When the controller has dozens of small devices in parallel, the
situation is more complicated. You're right; if just one transistor
fails shorted, when the rest turn off, that one little device will try
to carry the entire motor current. Like a fuse, it can do so for a short
time -- a matter of seconds -- and then it explodes!

Now when it explodes, a number of things can happen. It may blow
completely open and out of the circuit, like a blown fuse. In this case,
the controller still functions, albeit with one less transistor to carry
the current. This can slightly increase the strain on the rest of the
transistors, making it more likely that one of them will soon blow, too.
If all the transistor short, and then blow open like this, you get a
"popcorn" failure. Pop... pop, pop, poppity pop. It ends up with all the
transistors exploded, and no current to the motor.

But, explosions can do lots of other things. They can throw conductive
debris into other parts of the circuit, creating shorts. They can start
fires; charring or even burning the circuit board or surrounding parts.
How the circuit behaves with random shorts or while burning is
completely unpredictable. "Stuck fully on" failures are common.

Also, when a transistor first fails shorted, it can short its gate to
its drain. This can send a command to turn all the *other* transistors
on, since all their gates are connected together. This is also a pretty
common way to get a "stuck fully on" failure.

--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev