Fried VB921 for no good reason?

[16vboost]

Hello,

I'm running MSII, V3 using the VB921 for direct coil drive and got it to smoke after the second day of trying to start the car. It's a 16v Scirocco, I'm using the hall input from the distributor, and have done the following mods to the tach input circuit (see picture below)



My hall sensor sinks the signal line to ground, so instead of adding a pullup to the signal line to source current to the opto-isolator, I just sink current to the opto-isolator and tie the other end to S12. I've verified that the voltage at Tsel goes from 0 to 5 volts, as I turn the distributor with the engine at TDC. Other than that, I think my installation is very typical.

So yesterday, I saw some smoke coming from the VB921 chip as I'm thinking of what to do next between starting attempts (actually the smoke was from the screw, not the chip). I had been trying to start the car for at least an hour that day, and 4 hours the day before. I've had the ignition on for plenty of time. I checked the temperature of the VB921 when I first turned the ignition on, and after the first few starting attempts to make sure it wasn't getting too hot, and didn't check after that. But after hours of having the ignition on, and several minutes of combined cranking attempts (with spark) the VB921 smokes! Why is it that it would smoke after having worked for so long?

I was messing with the trigger offset when this happened, the settings below haven't changed since day 1

Ignition Input capture: Rising Edge
Cranking Trigger: Trigger Rise
Coil Charging Scheme: Standard Coil Charge
Spark Output: Going High (inverted)
Maximum Dwell Duration: 2.5ms

I'm running Megatune 2.25 and the 2.89 Embedded Code.

Oh and I don't know if this matters, but when I hook up an LED in place of the ignition coil, I can still see it flash when I crank the engine. But there is definitely no spark any more. Here's a pic of what the VB921 looks like now.



Any help would be greatly appreciated, I have two BIP373 chips coming next week from DIYAutoTune and I don't want to burn those too.

-Alex

[Bernard Fife]

16vboost,

Usually this is caused by either having the settings wrong, or the dwell too long. However, you look good on both counts.

So the first thing I would check is the coil itself. If it is shorting internally, it may still work but it will fry the VB921. So I would try to find the spec for the coil resistance and check the coil to see if it is in range, or if it is lower (in which case it needs to be replaced). Most repair manuals will have this specifications, and google will often find it too.

And when you replace the VB921, be sure to be generous with the heat sink compound (the current one looks good for that, though).

Lance.

[16vboost]

Hi Lance,

I checked the primary and secondary windings on my coil:

Primary measured .8ohm; spec .6-.8ohm
Secondary measured 3170ohm; spec 6500-8500ohm

This doesn't meet the spec for the 16v coil, but it does meet the spec for the 8v coil. I'm suspecting that someone got the wrong part number a while back and I was driving with this coil before I started the megasquirt project. I'll look for a new one either way, but it doesn't sound like the coil caused my VB921 to overheat.

Could entering a negative value for trigger offset cause a VB921 to stay on and overheat? What could cause it to stay on?

-Alex

[Matt Cramer]

Almost always when you kill a VB921, it's from too much current. BIP373s have a built in current limiting and thermal shutdown feature that makes them a lot harder to kill. The only time I've seen one of them get smoked was when somebody wired the middle pin to ground and the right pin to the coil.

[16vboost]

While waiting for my BIP373 to come in, I did a little experiment. I hooked up a spark plug, my ignition coil, and my megastim to my megasquirt, all on the bench. I could get the VB921 to flash an LED, but I confirmed with this setup, no spark. (though the coil would make a clicking sound). I think I "baked" the VB921. Either way, the BIP373 kit came in today from DIYAutoTune and I didn't waste any time in replacing it. The bench setup now had a SOLID spark. I played with dwell settings on the bench and came to the seat-of-the-pants conclusion that there wasn't much more spark energy to be had past 2.2ms of dwell (not far from my initial 2.5ms).

The coil and the megasquirt are back in the car and I'm back to cranking and popping. Man it WANTS to start so bad. I'll post in the correct category if I need help tuning. I think I will.

So I still aren't sure exactly why the VB921 baked in the first place but I think the BIP373 shouldn't have the same problem. Thanks for the tips and advise though.

-Alex