Dead Topping MX3 -- overloading power supply

[wpmcnamara]

This is the amp I had been using for patio speakers.  It generally lived on the patio unless we were going to get rain.  It is only a about four months old.  We had been sitting on the patio one evening and it was working fine.  Came out the next evening and it was dead.  The power LED on the power brick would briefly flicker, which I interpreted as the power supply shutting down due to overload.

This evening, I got it on the bench and connected it to a current limited supply and sure enough, the supply immediately went into current limit.  I had it set at 12V/0.5A, which is well below the maximum for the amp.  Visual inspection revealed nothing obvious.    I left it connected to the power supply and started probing with my Mk1 finger and quickly found a warm component.



D17 was getting pretty warm.  I pulled it from the board and it tested good, but the overload went away.  It is a series diode, feeding into a buck converter that generates 3.3v.  It is fed directly from the power jack.  D1, seen at the top of the picture is the flyback diode.

Better picture



Measuring across it's terminals shows about 2ohms either direction.  Pulling it off showed it to be good and 2ohms across its pads.  I went ahead and pulled the TPS54231 off as well, to get all the active parts out of the circuit.  Still a couple ohms across the pads of D1.

This is where I sit now.


At this point, I'm thinking about pulling C132 and C57.  They are the output filter caps and if either of them cracked and shorted, that would explain the behavior.  I don't really see any point in removing L3 as one side is already isolated and the other feeds the output caps.  I'm not sure where I'll go next if I pull the filter caps and the short is still there.  Maybe feed 3.3V from my bench supply and see if anything else gets hot?

[John.S]

On the picture it looks like FB16 is cracked. If that is a kind of protection diode it may be your short.

[wpmcnamara]

On the picture it looks like FB16 is cracked. If that is a kind of protection diode it may be your short.

I had a similar thought, however FB stands for Ferrite Bead.

I pulled the two big ceramics off and they both tested just fine -- 22uf -- and the short was still there.  I fed current limited 3.3V into the 3.3V line and found it to still behave the same way.  I turned the current limit up to about 1A.  Probing around with my Mk1 finger again, I found the Bluetooth module to be very hot.  Pulled it off the board.  Note castellated boards are cool, but a PIA to unsolder.  With the BT module removed, the short was gone.  3.3V to ground measured in the 900ohm range.  Unforuntately that didn't fix things.  I put the caps, diode and TPS54231 back and it still immediately puts the power supply into current limit mode.

Unfortunately, there isn't a way to isolated the buck converter output from the rest of the board, so I can't easily tell if the converter would function if not for the overload.  I'm starting to think not though.  I am starting to think that a failure in the converter may have effectively dumped 24V into the 3.3V line, in which case "its dead Jim".  I'm thinking I'll pull L3 and build up the output of the converter on a breadboard.  It's effectively L3, the two filter caps and two resistors for voltage feedback.  It will let me test the converter in isolation. 

[cmstein]

Did you solve the problem?
My mx3 has the same symptom (the power supply goes into current limitation).

How did you open it? Apparently the front part is just fitted, but I didn't want to force it.

Thanks