Active cooling for a cheap power supply

Post date: Jan 09, 2017 3:18:14 AM

It happened -- an accident killed my big power supply. Alexander (KF7ZIM) was disconnecting some stuff he had connected to it, and apparently bumped the bus hard enough to short something out (not hard to do with the big clips, etc. standing up exposed around the connection bus), and poof! Suddenly the overload alarm on the power supply was sounding continuously. The output voltage now stayed at about 15.5V regardless of what you did to the adjusting knob, (We later found that that was only true in a no-load situation; otherwise the voltage quickly went to nothing, so it was unable to power any useful load.) After ruling out some basic problems (e.g. something still connected to it that shouldn't be), probing around inside of it (with a really weak schematic diagram, and no other diagnostic info) and doing a lot of research online, I determined that it wasn't going to be up for a while. Not wanting to buy another $123-300 power supply, I placed an order for what I felt were the most likely cheap suspect components (the SCR and the 723 voltage regulator IC), and sat back hoping for the best. After a couple of days running off the battery, I adopted Alexander's suggestion that we hook up the little 3A supply that was being stored on the bench with its cord wrapped around it.

Power Monitor

I put my cheap Chinese DC power analyzer (pictured below) inline, so I could monitor the current, both to assess the battery charge, and to ensure that I didn't overload the power supply. A word about this meter: I originally bought it last February for $12 intending to put it in full-time between the radios and the power source (battery + bench supply) to monitor the current draw and the battery voltage (charge and discharge). Two things changed my mind:

• • The thing draws quite a bit of current (nice bright backlight!), which would continually drain the battery

  • It's horribly inaccurate, despite what the 2-digit decimal precision might lead you to believe -- e.g. at 12V it reads around 0.4V low, and it's like 30-40% low on its current readings.

So, it's useful for getting a ballpark number, and for monitoring something when you've already checked the real value and can mentally correct what you're seeing. It's really nifty for the money, but I wish their quality control had been a little better.

Hot Stuff

It took quite a bit of time to charge the big battery back up, and despite the meter reading <2A most of the time, the supply started to stink pretty quickly. After reading online, it appears that the supply's claimed 3A rating is for surge, but you really don't want to push it over 2A -- and I was. I was surprised to find that it wasn't the main power transistor on the back (mounted to an external heatsink) that smelled, nor anything on the PCB; it was the power transformer! This supply doesn't have a fan, and I decided that if it was going to be useful, it was time to add one. I found that an 80mm fan's screws aligned nicely with the outer ventilation holes in the top of the case! Since computer fans run on 12V, it was a match made in heaven.

I drilled two of the holes a little bigger to accept the fan mounting screws, and passed the wires through two adjacent holes. Alexander donated a SPDT toggle switch from his collection and added a 2W 100 ohm resistor to it, and presto! 2-speed fan. It makes a huge difference -- running it on low (roughly half-speed), it quickly cools the interior of the supply. Aside from me losing my white-on-black label tape and using black-on-white instead, it even looks good!

Now it's the pass transistor on the back that's the hot spot. Maybe we'll have to design a shroud for the fan to direct the air down and over the back of the case to cool that heat sink so we can push it a little harder without significantly shortening its life.

Stay tuned to find out what happens with the PS-36KX. In the meantime, the new PS-3PLUS is keeping us on the air!