Close

The dangers of surplus wall warts

A project log for GPS Disciplined xCXO

A DIY GPS disciplined 10 MHz reference clock

nick-sayerNick Sayer 07/29/2015 at 07:060 Comments

I knew that I was going to need to procure a high current, low voltage power supply for this project. To that end, I went down the row of surplus wall warts at Halted a few days ago and picked what appeared to be a suitable candidate. It was rated 1A @ 6V.

Well, it's been powering the complete prototype project for a few hours now, but I began to hear a high pitched whining noise coming from the oscillator module.

Well, whenever your project makes a noise it's not supposed to make, the first thing you should do is put an AC coupled scope on the power supply lead at the component that's having trouble. Sure enough, there was 9 kHz ripple at ~600 mV P-P in 0.5 Hz pulses perfectly in sync with the noises coming out of the module.

So... that power brick is fired.

I've moved things back temporarily to my desktop switcher, and the problem went away. I'll go back to Halted tomorrow and try to find another candidate.

With the switcher on duty, the power looks very nice - < 30 mV P-P ripple. That ought to do.

EDIT: Turns out, it doesn't look like it was the wall wart after all. I came home and the noise and oscillations had returned. Perhaps it's the LT1117 regulator going into cyclic partial thermal shutdown or becoming unstable or something like that. I dunno. I'm less interested now in tracking that problem down since I've more or less decided to go with VCTCXOs for the project.

EDIT: Adding a larger heatsink to the LT1117 made the oscillations go away for good (on the board with the VCOCXO and the trimpot). But that heatsink is the size of Montana. I'm glad I decided to go with VCTCXOs instead. My hope is that the total current draw for the whole project will be under 100 mA, which means that a D2PAK NCP1117 being heat-sunk by the PCB will be acceptable.

Discussions