Author Topic: Cap blows cooling fan transistor in power-supply unit.  (Read 481 times)

0 Members and 1 Guest are viewing this topic.

Offline paul8fTopic starter

  • Regular Contributor
  • *
  • Posts: 109
  • Country: ie
Cap blows cooling fan transistor in power-supply unit.
« on: July 25, 2020, 09:00:53 am »
Came across an interesting fail recently. In a power-supply cooling circuit, a small surface-mount 10uF electrolytic (C95) went short-circuit and burned out the main power transistor (Q51) that drives a parallel pair of DC cooling fans.

The cap looked fine until I got out my magnifier and spotted a hairline crack across it. The component literally fell apart as soon as I brought my iron next to it. With the cap now out of circuit, and the blown transistor replaced, the fans started spinning again.

At first glance at the schematic, I thought that C95 was there for some sort of soft-start/inrush function, but now I think it's there for electrical noise suppression of the motors. It's strange that C95 seems to float in the circuit (i.e. its positive side is permanently connected to 12 volts, but its negative side only ever gets connected to GND when the micrcontroller sends the 5V signal to Q50). The power-supply unit wasn't used that much, and when it was it was for light use in a cool environment, so the fans could have easily gone a year without getting a signal to turn on. Could the electrolyte have dried out?

The +12 and FAN+12 rails are derived from different transformers, but share a common GND. The fans use identical motors (see photos), and the SMD electrolytic that cracked actually looked like a regular 1206 ceramic or MLCC cap.

I found a separate issue with the power-supply, whereby the NTC thermistor spec had drifted with age, possibly causing the fans to continually cycle on and off repeatedly. I'm just curious as to what may have caused C95 to fail like that. Any suggestions?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf