Metcal MX500P blown buck converter U4 (LM2576HVT): why?

[benj38]

I have a Metcal MX500P-11 that I bought as non-working on eBay.
The unit arrived with a blown buck converter U4 (LM2576HVT), with no other signs of trouble, and no obvious signs of somebody working on it in the past.

After replacing U4 (with one from DigiKey) the unit came back to life and worked fine for a couple of weeks until yesterday when I have heard a "pop" sound and the new LM2576HVT died. Again, there seem to be no other problem (in particular, both the driver and output mosfets Q3, Q4 are just fine).

Replacing U4 every few weeks is obviously not the way to go, so there must be some subtle underlying cause.

Is this a known failure mode? Is there a usual suspect?

Any help in zooming in on the culprit would be highly appreciated.

BTW, I do have the schematic etc. from the excellent "metcal MX-500P-11 technical documentation" file.

[HB9EVI]

a spontaneous wild guess:

did you check the inductor?

I had once a case of a repeatedly dying buck converter; it turned out, that the circuit was overloaded so that the inductors enamel melted and short circuited the inductor what caused the regulator to run hot and die a thermal death after a while

[benj38]

That was my suspicion also, but the coil is fine: it measures in at 520 uH with a Q of 26, and also passes a ring test quite well.

[floobydust]

Hard to tell if overvoltage or overcurrent is causing the IC to fail. C7 at the input seems important. Otherwise I would suspect the electrolytics or the RF power setting.

edit: I would say there should be a disc cap in parallel with the buck-converter's output, across C8 to ensure no 13.56MHz oscillations make it back. LPF L11, C21, C22 might not be enough and C8 surely does not do much at that frequency.

[fzabkar]

I confess that I don't understand the PCB designer's use of the FB pin. Shouldn't this pin be sensing the converter's output via a potential divider? Why is it sensing the converter's input???

[mikerj]

I confess that I don't understand the PCB designer's use of the FB pin. Shouldn't this pin be sensing the converter's output via a potential divider? Why is it sensing the converter's input???

It's not sensing the input, note that the potential divider RV2 is across a zener diode.  The actual feedback is from R7 which connects to a peak detector on the output of the RF power amplifier.  RV2 and R6 introduce an adjustable offset to trim the RF power level.

[Psi]

Just a wild stab in the dark. Could there be an issue with the iron/cable where the RF power is sometimes reflecting back and overloading the circuit?

[Towger]

If the SWR is high the output 'should' shut off and LED turn yellow.  Requiring a power cycle to reset.

Is the regulator running very hot, so eventually fails or seems normal?

[benj38]

Sorry for the late reply, had no time to work on this until today.

I have installed a new converter that lasted for 30 minutes or so, which allowed me to make some tests, and then blew up

@Floobydust: thanks for the suggestions. I have checked the RF power setting and it is fine (around 17V-18V when idling hot, and never more than 18.5 when heating up).
All caps measure fine, in particular C7. Scope shows no 13.56MHz oscillations making their way back to the output or the feedback pins of the converter.

@Psi: The idea that the iron (i.e., tip handle) may be bad is interesting. I have two handles and multiple tips. They all check fine. In particular, the handles are simply connectors that hold the tip and have no components in them.

The funny thing is that I get slightly different measurements of converter output voltage when idling hot with the same tip in one handle then I get in the other handle! (about 0.5V difference). I can not explain this.

I have opened and looked at the connections in the handle that was connected when the converter blew up and I can not see nor measure anything wrong with it. Twisting and wriggling the cable does not indicate any intermittent fault either.

@Towger: thanks for your input! The unit never goes into a yellow LED mode. The converter does not run hot.

I can add the following three observations:

1. The unit makes an audible whine (sounds like a few KHz) when heating up a tip. It is surprising since the converter is supposed to work at inaudible 52KHz. Is this a clue?

2. The voltage at the feedback pin oscillates slightly (amplitude around 300mV pk-to-pk) around the feedback voltage of 1.23V. I suspect this may be normal. The frequency is about 4.5KHz. Maybe this is related to the whine.

3. When the converter finally blew up, it was instantly when I plugged the unit in , and with a bang that split the chip.

Help!

[oilburner]

old post, but I'm trying to fix one of these Metcal irons...  Were you able to get it repaired Benj38?

[DRM2]

Anyone have any success with this failure?