The issue is, in my opinion, a bit of both. The way Stern drives their mosfets lends itself to self-destruction if_ the code isn't carefully done.

However, the circuit on the aux board is pretty much identical to what they use on the main board, so I would think that bad code would also be responsible for destroying main driver boards too, but I haven't heard anyone really talk about that.

Maybe it's because they are doing all their code monkeying on the aux board lines only. Don't know. Stern could probably modify the code, but that may defeat some of the effects they were trying to achieve. So in my mind, they need to fix the hardware.

My offer is still on the table if anyone wants to send me their blown up (or working) board for me to remake. I am pretty confident that I can fix what I see as the hardware problem, that way Stern can keep rolling out bad code to their heart's content.

Send me a PM or email if interested.

Tony

Jackal2001:

I completely agree with you.

Rotordave:

It's not about the rating of the transistors, they should be plenty beefy enough. Should be blowing fuses before the transistors fail.

Putting beefier mosfets in there might potentially help but that still may not be enough.

Tony

Maybe if someone who's got a current dead board can take some hi-res pictures of it and then take pictures of the replacement board Stern sends out. Maybe if we compare a blown board against a replacement that works, then we may know if the problem was just bad parts or if Stern updated the hardware.

If there is a hardware change, maybe look for a change in the board revision number printed on it.

Just some suggestions.

Tony

markmon:

Interesting. If that's the case, then it could just all be due to a bad batch of parts. Scary!

Tony

I'm quite interested to see what Stern comes up with as a solution.

Please post pictures of any new board that Stern ships!

Tony

Looks like they added a larger cap to the logic power line?

Maybe transient spikes were killing the logic chip?

Tony

You can destroy a mosfet without ever running enough current through it to blow the fuse.

All mosfets have the capability of operating in a "partially on" (linear) state. In this state, the mosfet loses its near-zero impedance and has to dissipate far more heat than it was designed to handle.

Ultimately if you operate a mosfet this way, it will overheat and fail. And most likely your fuse will still be perfectly fine.

Those burn marks in the picture above might indicate that the mosfet overheated.

Tony

That would be a very unlikely situation. Most coils don't become dead shorts, typically they just end up with lower resistance, which in turn draws more current, which ultimately leads to blowing the fuse first, not the mosfet.

I don't believe that is what is occurring here. Any coils that are going bad is most likely due to the mosfet having failed first, not the other way around.

Quite possible. I have been suspecting the software since I became aware of this issue. Again, it's primarily due to the way Stern designed and controls the mosfet circuitry. Any wrong move in software can lead to those mosfets ending up in a "partially on" state. Because of the unique elements in this game and possibly also the way they are being used or controlled, it may explain why this board has not failed in other games.

Tony

Not necessarily.

<techie on>

Switching mosfets generally have a fairly loose "gate threshold", that is the voltage at which the manufacturer specifies that the device basically starts functioning. But that threshold point doesn't mean that the device is running at full rated capacity. Generally you need to be as far above or below that threshold voltage as possible to ensure proper operation.

Let's say for example we have a IRL530 "logic level" mosfet. The gate threshold can be anywhere from 1V to 2V per the manufacturer's datasheet. According to the datasheet, the IRL530 can't start handling 10A of current (which we'd like it to do) unless the gate voltage is approximately 3.5 to 4V (assuming a 2V gate threshold).

It is quite possible then, that mosfets with a lower gate threshold of say 1V may be surviving, while those with a 2V threshold don't. That 1V of difference (according to the datasheet) might equate to upwards of 6 Amps of additional operational capacity!

The problem is that the Stern circuitry allows the software to potentially have those mosfets operating too close to that critical threshold level. The programmers have the capability of writing software that could be destroying these mosfets and not even understand what is occurring.

<techie off>

Stern either has to correct the hardware to remove the possibility of bad software code from potentially causing problems, or change the software to remove the code that "could be" causing the current hardware configuration to fail. The hardware is easy to fix (IMHO). The software may not be since some of the cool features in this game may rely on the particular way they wrote it.

Certainly I understand the skepticism about my view on this situation, and certainly I could ultimately proved to be wrong. But every time I look at the schematics, I see a glaring and obvious problem staring back at me.

FWIW

Tony

Once I have had a chance to examine and run tests on one of these boards (and I hope to be able to do that very soon), I should have a much better understanding of the hardware and what can be done with it.

Thanks!

Tony