Burning of resist. R154 of PM8834 auxiliary power to MOSFET driver of PFC of STDES-3KWTLCP ref dsgn

[DSHAR.31]

I have assembled PFC and LLC circuit of STDES-3KWTLCP reference design. I followed the procedure as described in the document to assemble whole circuit. When we gave AC voltage to PFC, resistance R154 gets heat up and eventually burn out. I double checked all the components value and their placement but couldn't find the cause for this behavior.

Can anyone please assist us in finding which part of the circuit or component is creating this issue so that it would help us in rectifying the cause?

Thanks.

[AScha.3]

>D112, D117, D122 which may be drawing too much current

Sure - possible, but thats why i recommended IR-cam : you should "see" , where current is going...

suspicious are also the Schottky bridge diodes, if one is 4° more than all others - replace it.

These are 40V diodes and the supply gets 18V, so about 36...39V on them reverse. 

(Bad design idea, usually rule of thumb is using 2..3 x max. voltage of supply (ie if i choose a rectifier for 400V mains, 1200V type is minimum, 1600V better.) So here i would use 60V diodes minimum, or 80V if avail.

But as said: see on IR , who is the "bad guy" . If no zener D112, D117, D122 and no R 152/56/58 shows somw degree more, than others, then remove Schottky bridge diodes; if ok then (R154 still will show some "warm" temp, transformer needs energy even without load), use 60V Schottky or fast Si- diodes and see then...

diodes at mouser:

https://www.mouser.de/c/semiconductors/discrete-semiconductors/diodes-rectifiers/schottky-diodes-rectifiers/?if%20-%20forward%20current=2%20A&type=Schottky%20Diode&vrrm%20-%20repetitive%20reverse%20voltage=60%20V

maybe : 78-SS26SHE3_B/H  (look, if case/package matching !)

[DSHAR.31]

I removed D112, D117, D122 Zeners and checked the circuit again. Now, R154 is not heating up, neither any other component.

But after removing zeners, the Vcc_18VA went upto 50V. Which is obvious because zener was clamping the voltage to 18V-20V. Similar behavior observed on other 2 VCCs. 

Although R154 is not heating now, but we cannot run the circuit without Zeners because the all 3 Vcc should be in the range.

What can we do to rectify the issue with the Zeners?

[Peter BENSCH]

If you get a voltage of 50V at VCC_+18VA after removing D112, the maximum reverse voltage of one or all Schottky of the bridge rectifiers will be too low. Either replace the bridge rectifiers with 60V or 80V Schottky, as @AScha.3 suggested, or with ultrafast rectifiers with >80V.

[AScha.3]

50V is because no load, so put some resistors on every output, loading 2mA -> about 6 kohm at 12v , 2x 10 k for 18V outputs.

Voltages then ?

+

Transformer is for 12V  + 2x 21V outputs :

https://www.we-online.com/components/products/datasheet/750344767.pdf

So i would try: change zener diodes to 24V type: (same case )

https://www.mouser.de/c/semiconductors/discrete-semiconductors/diodes-rectifiers/zener-diodes/?pd%20-%20power%20dissipation=95%20mW~~250%20mW&vz%20-%20zener%20voltage=24%20V&instock=y&rp=semiconductors%2Fdiscrete-semiconductors%2Fdiodes-rectifiers%2Fzener-diodes%7C~Pd%20-%20Power%20Dissipation

-> STGAP2D driver chip has 26V max. , so i see no problem with zeners limiting to about 24V .

+ maybe the 12V zener -> 15V type .

[DSHAR.31]

Just to be sure that the issue is not with auxiliary supply, I removed the auxiliary circuit and gave all 3 VCCs externally using 3 different regulated power supply (2 X 18V and 1 X 12V).

What I observed is , both gate drivers STGAP2D  draws too much current that one of them burnt with flames, while other one also got damaged completely, with external VCCs dropping to 2V for 12V and 9V for 18V..

I still cannot understand which section is wrong.

[AScha.3]

Upps...now no need to understand which -- now repair ALL .

You have to replace all gate drivers, and when removed the old, check gates of mosfet/igbts for good or dead (most time short or low resistance, if damaged).

Was this 3kW inverter running ever ? What happened before ? 

[DSHAR.31]

Only once we got output as 400V from this PFC. But this output remained for only couple of seconds and R154 got burnt. 

After replacing R154, it got burnt several times but 400V output never came and MOSFET also got shorted couple of times.      

[AScha.3]

So we have here an exploded PFC , which is/was never fully repaired.

You should have said that from the start.

If we assume that there is little experience for repairing an SMPS, the best way is:

- replace all parts in the "way of destruction" : driver chips, gate resistors+zeners, mosfet/igbts.

Then test without HV power : driver etc running (without any part getting hot) ?

If ok, apply mains, but with "safety current limiter" : i use a simple, old (230V) bulb (60...100W) for this,

and a variable transformer , to rise the voltage slowly and check at even low input voltage : running, ok ? or not.

Then look, whats still damaged , replace and try again.

[DSHAR.31]

Before burning of gate drivers, we replaced all MOSFETs, gate drivers and zeners and checked all resistance and capacitors.

But what surprises me is the heating of R154 even when auxillary supply is not connected to the gate drivers. To double check, theis auxillary supply was assembled on a new PCB. Still, R154 was getting burnt when not connected to the gate drivers.

Also, I dont fully understand how PWMs from microcontroller can run this auxillary supply. Can microcontroller deliver this much high power? 

Is it possible that in the documentation available on ST website for this reference design is having some misprints in the schematic or,

the firmware available of this PFC as downloaded from ST website is not updated?

[AScha.3]

>I dont fully understand how PWMs from microcontroller can run this auxillary supply. Can microcontroller deliver this much high power? 

No. The PM8834 is the driver, that drives the transformer; its clock/pwm is from cpu :

+

>the firmware available of this PFC as downloaded from ST website is not updated?

Possible - but it should work, as it is, because STM dont want to show a "demo board", that just blows up, if you test it - we can assume .

+

>But what surprises me is the heating of R154 even when auxillary supply is not connected to the gate drivers.

Right, this is no "nice feature". But as i showed more early, there are zeners, that (intentionally) burn some current, to keep voltage at desired level without using some active regulators. So there must be some resistor, to limit the input current to all this, R154. It will burn some power, but if everything ok, it should be in the limits, it can do long time. (So maybe heating up to 70° is no problem for a resistor. ) If it begins to burn, something is wrong...