TPS54335A: Chip burns after at ignition control

For information D50 should protect against wrong polarisation and D51 against Overvoltage.

Hi Dirk, 

We'll have to analyze the waveforms on Vin & Vout  paths. We also need to probe the current drawn.

We cannot assume stuffs here. 

Regards,

Gautam

Thank you. Could it be I did a mistake in the circuit?

Schematic looks fine, just that I'm not sure Zener D51 is able to clamp any input transients. Esp. during cranking or ignition as you might be aware, large transients get absorbed by the battery.

Thanks for sharing your thoughts. Yes, I also suspect the diode, so I removed it. It's not there in my circuit with the EN1584 either.

Also, if you observe the damage signature above - Pin 2 is the one that seems to be damaged ie Vin pin. We need to probe and capture the Vin transients well. 

Correct. My problem is that the motorcycle is many kilometers away, where this problem has already occurred twice. I can't measure it. But maybe I can rebuild it and measure it. How big would a capacitor on EN have to be to delay it by 0.5 seconds? Thank you very much.

Why do you want to delay EN by 0.5sec? FYI, it will be a big capacitor bank of ~5F. 

I thought it could minimize my isssue...But 5F is too much

That will not help. We'll have to arrest the transients on Vin pin first. 

I guess I found the issue, without D51 I could messure 34V at the input with a 16.8V Lithium akku. After it the chip exploded.

Are there any suggestions to solve it

I played a little with different loads. It seems the load is too low at starting, because the controller starts much later.

Yes, exactly. Its very clear now. Looks like this is simply not a transient, it takes 800us to reach the absolute max of the chip. Why is the zener not clamping? 

The value is 40V.

Can you find something lower? Clamping around <30V?

yes, but I think it is too unsecure. I guess I will change the chip. The MP1584EN doesn't show this behavior.

Dirk, but the absolute max with MP1584 too is 30V and might be risky. You can have a look at our Wide Vin parts that support >30V. 

yes, but I dont't measure this behavior. But the MP1584 is out of date, therfore I tried the TPS54335A.

If I would have a TI Chip with the same shematic I would try it.

Thank you for your all responses. I will look.