Power Supply Death

Power Supply Top


A few months ago I was rudely awoken in the middle of the night by the “exciting” smell of burning electronics! I got up fairly quickly and wondered what I could have possibly left on. It turns out it was not a project of mine, but (after some smelling around and unplugging of everything within sight,) my PC power supply. 2AM is a great time to catch fire!

As the pictures show there was a small scorch mark on both sides of the PCB. A blackened resistor is the culprit. I believe the root of the problem was bad capacitors possibly shorting out a power rail. The PSU was making some high frequency noise previously, notoriously caused by bad caps.

To get this far into the power supply I had to pull a lot out. It ended up being a good source of some components, but I never planned on repairing the supply. A good capacitor has already come in handy temporarily repairing an old HP 3478A.

Overall the events could have been worse than what happened, however, I would have hoped a fuse would have blown well before the power supply failed like it did.
Power Supply Bottom

ViaCar Sensors

ViaCar SensorsA quick picture: a group of sensors for ViaCar(NatCar) I made during my last year at UCSD. More to come when I get the chance to photograph everything, check out the projects page in coming weeks!




There is a video at the end of the post, check it out!

This is a pseudo-VU meter written in VHDL. I did it for an afternoon project to see what I could do with VHDL quickly. I take the line level audio output from my computer and run it through a Schottky diode ( half wave rectification ), and then straight into a serial ADC. The rest is done on the Xilinx Spartan 3E FPGA.

A true VU meter would translate the volume level of the output signal to a certain DB level. Mine just peak detects the output of the ADC and then scales the output until I get good dynamic range out of the LEDs available. The end result could use some refinement, but I was happy with the rough draft.

I need to clean up my VHDL. I will have it up on github sooner rather than later!




I have to apologize for the lack of posts for the past few months. I have been collecting a good chunk of content but haven’t uploaded it. Expect more in future weeks, I have more time on weekends now to write posts!


Non-destructively testing a heart rate monitor strap.

Polar HR strap


The polar heart rate monitor straps are completely sealed. This is the easiest way to see the output on a scope while still being able to wear the strap at a desk. Of course if the number of turns was optimized the output would be better, but for something quick this worked just fine.