Pixsys ATR-171 teardown.

[ResR]

This unit broke just under 6 months and I took it apart to see why. I see a traces of magic smoke, but I can't find the source. No cracks, no burns, just some black stuff on some components. Main processor MC9S08 is a 8-bit MCU by Freescale, there is also AD7795 6-Channel, low noise low power, 24-/16-Bit ?-? ADC with on-chip in-amp and reference and other misc. ST, AD and TI components. The PSU driver chip is TOP 246 PN. It has custom LED display with pretty bright red, yellow and green LED's.
I thought this supposed to withstand for years, because it controls industrial processes, but I guess I was wrong.

[PA0PBZ]

Hard to see how this thing folds up when assembled, but the cap in the top right of the picture (next to the transformer) has had a lot of the magic smoke looking at how it is discolored.
Maybe have a look around there.

[free_electron]

that chip is toast ...
right hand side of picture just below the cap.
the cap is discolored from the chip smoke.
chip has a charred trail going downwards...

[ResR]

That's TOP 246. 132kHz switching regulator DIL-8, 3,11 euros from elfa distrelec. I was suspicious about that too, will test if I get enough to buy a soldering iron and a spare part mentioned above . I wonder, why those regulators burn out so easily, maybe the power fluxuation has something to do with?
Edit: Sneaky little ... There is a hole at the bottom of the chip, top looked solid.

[SeanB]

You are looking for trouble having a 1kV pulse on a tiny little chip with so little isolation. Slightest contamination on the die surface ( breath of the person doing a QC visual check on the wafer, or just on the die before encapsulation, or a spot of water trapped in a crack in the leadframe) and it becomes a time bomb. I prefer a separate power device and a switch die, even in the same TO package they work better.