Why did this zener diode explode?

Thread Starter

fablau

Joined May 14, 2020
167
A couple of details I forgot to tell you: for some reason, instead of having 32v from the H-bridge, I took the measurements with 27 volts from the same (I must have inadvertently turned that down from the power supply by mistake). But I don't think the results would have been much different from 32v. Also, instead of using the Zener design we discussed earlier to drive the MOSFET's gate, I have used a LM317T which gives me a steady 15v to be used for the gate. Finally, I have performed the tests with a N-channel MOSFET.
 
Last edited:

click_here

Joined Sep 22, 2020
545
So the gate with nothing on is at ~11V, and the source is at ~4.2V

[edit] ~ means approx [/edit]

In the off state you have a Vgs of 6.8V

The datasheet has a Vgs(th) of <4V, so the MOSFET is *not* in an off state
 
Last edited:

Thread Starter

fablau

Joined May 14, 2020
167
So the gate with nothing on is at ~11V, and the source is at ~4.2V

[edit] ~ means approx [/edit]

In the off state you have a Vgs of 6.8V

The datasheet has a Vgs(th) of <4V, so the MOSFET is *not* in an off state
I am sorry, but where did you get those numbers? In the OFF state I have 0 volts at the gate and -8.60v at the source when measured with the positive from the rail (+27v), or 13.50 when measured with the ground (0v)

Curious to know that. Thanks!
 

click_here

Joined Sep 22, 2020
545
"Gate measurement with no PWM applied" - sits at just over 11V

"And finally, here is the source without PWM" - sits at about 4.2V

The Vgs(th) is from the datasheet

I suppose I should clarify if you wanted current to flow when no PWM is applied
 

Thread Starter

fablau

Joined May 14, 2020
167
"Gate measurement with no PWM applied" - sits at just over 11V

"And finally, here is the source without PWM" - sits at about 4.2V

The Vgs(th) is from the datasheet

I suppose I should clarify if you wanted current to flow when no PWM is applied
Oh, I see, you read those values from the pictures I uploaded for Vp-p. Well, I got confused too because that actually shows the peak. I measured everything again now and I can confirm that that's the case.

And I have great news. I have solved the heat problem. The solution was to turn off the PWM when the magnets were off! That's the glitch I forgot to address that I mentioned in my post #140. Now, everytime I turn off the magnets, I also turn off the PWM, and that's it, the MOSFETs stay cool ;)

That's because this is the cycle:

1. MOSFET turn on
2. PWM turns on
3. Everything stays on for 300-900ms.
4. MOSFET turns off
5. PWM used to stay on!!!
6. Delay of 2-3 seconds.
7. The cycle repeats

So, solved point 6 above by turning off PWM, problem solved ;)

Also, I took measurements again with the new meter of the current needed to turn all the 9 magnets on/off at their gate, which was driven by the Zener we discussed at the beginning that I have now replaced temporarily with a LM317T, and the max amps I get there is not more than 4 mA! Therefore I think I can still use the original Zeners (the one that exploded). I am 99% sure that Zener exploded because several MOSFETs shorted because of the heat, and the Zener had to absorb the high current coming from the drain of the MOSFETs.

So, to conclude: heat problem solved, Zener problem solved.

Thank you again very much for all your help, and I must admit, even if this scope is cheap and small, it gave me the missing clue ;)

If anything else comes up, I'll let you know!
 

click_here

Joined Sep 22, 2020
545
A scope will make design/debugging so much easier - Even confirm microcontroller outputs and things like UART baudrates

What you said goes with what I saw on your scopes - The MOSFET was conducting even with no PWM signal.

I'm happy to help
 

Thread Starter

fablau

Joined May 14, 2020
167
A scope will make design/debugging so much easier - Even confirm microcontroller outputs and things like UART baudrates

What you said goes with what I saw on your scopes - The MOSFET was conducting even with no PWM signal.

I'm happy to help
Exactly! You got it right ;) Thank you again very much :)
 
Top