ltc

None of the above.

Given the 'buried/non serviceable' nature of some of the modules, a fuse was rejected...go figure. High current capacity PTC's were also rejected as well.
(Interesting to note that some customers require that a module, when conencted to reverse polarity voltage, will NOT blow the vehicle inline fuse)

In the end, the proposed changes were multifaceted.

1. SMT caps (connected to B+) were to be placed in an orientation of minimal mechanical stress, depending on where the device was located on the board (distance from edge, mechanical mounting point, etc), in an attempt to minimize the formation of the micro crack.

2. SMT caps (connected to B+) larger than 0603 size were to be of either 'flexible termination' construction or a 'fail open' construction. (it was assumed that the size of a 0603 cap would make it highly unlikely to induce a micro crack)

http://www.avx.com/docs/techinfo/mlccflex.pdf
http://www.kemet.com/kemet/web/homepage/kechome.nsf/vapubfilesname/F3280_X7R_FloatElectrode/$file/F3280_X7R_FloatElectrode.pdf
http://www.kemet.com/kemet/web/homepage/kechome.nsf/vapubfilesname/F9121_X7R_100V_openmode/$file/F9121_X7R_100V_openmode.pdf

3. PCB panelization would minimize break off tabs and focus on minimal stress methods to remove the circuits from a panel (post SMT reflow/assembly/test). ICT fixture to board stress (bed of nails) to also be analyzed and probe type adjusted accordingly.

4. Tantalum caps were not to be placed on high current nodes (like B+). Ceramic caps (as above) and electrolytics instead.

That's about all I can remember from Larry told me. If you want additional details, you're going to need to get a shovel and start digging.

Bob Rouleau

Thanks Lewis - great memory! I can see that Larry had to deal with some serious constraints...

ltc

No, he usually had to deal with morons...he didn't mind design constraints

OCNYPORSCHE

You are man of many talents. Caps mean baseball caps right

I would still say that since the stress of the components are causing the issues that we should blame the mechanical engineers.

All kidding aside, very interesting stuff. Glad that they finally fixed it, I guess.

Bob Rouleau

As Lewis recounts, capacitors are commonly used as a bypass to ground. Allowing them to be exposed to as much current as the battery can supply without protection is bad design. Circuit boards are always subject to vibration. If vibration is a known condition, the use of stronger material, thicker boards and even full encapsulation are typical solutions. On the other hand no circuitry should be connected to a high current source (like a car battery) without a protection circuit. There are numerous devices which protect against overcurrent and reverse polarity. Not specifying said protection is just bad engineering. Car radios typically have this protection why not other electronic modules?

Wonder if Delco,Delphi and the like have similar weaknesses...

ltc

Yes.
They were also behind the shift to flexible termination and/or fail open capacitors.
It is quite astonishing the lengths companies will go to in order to save pennies. I suppose it matters when you are making millions of an electronic module, but the risk of an uncontained thermal event must be worth something.....

It's behind me now (Larry is dead). I'm having fun playing with fiber lasers. The 50KW one is interesting.

Sam CS 05

Andrei,

I am glad all is well with your family.

Now I hope that the "Thermal" event can't happen when the car is not running and /or parked...?correct?

JohnnyBahamas

Any mods? Thermite intake manifold with magnesium glow strips for example?

INURGRL951

Hey wirunner thanks i will have to look into that but that was also for the 05/06 ones too?

ltc

Incorrect.
Anything connected (directly) to the battery....say, for example the Bose MOST DSP audio amplifier tucked away behind the panel in the rear cargo area .....can suffer an uncontained secondary thermal event. It is 'usually' devices which have high power densities/high capacity fuses...since they can support the high fault currents for sufficient times.

I wouldn't worry, it does happen, but very infrequently. I would imagine by now most companies are following Larry's example and designing at least simple attempts at fixes.

When Larry was alive, I think the Mazda Minivan was ~ the 12th reported case in that company's history.

ltc

No, the routing was corrected at the factory post 03/04.

amaist

Brake controller.

Properly installed according to instructions found on Rennlist.

ltc

Brake controller good.
LED trailer lights on the other hand...

Dennis C

So... I guess I should tell my wife to carry marshmellows in the Cayenne, just in case...

INURGRL951

If she likes them over a open flame....haha.