Split-rail power supply

Building an adjustable linear split-rail PSU.

Posted by revmecha on June 13, 2016

I wanted to build a split rail power supply to use on small current projects. It's pretty much going to be used for quick and simple experiments so I didn't add over voltage or a constant current circuit. I have the HP E3611A Power Supply for anything that requires more current so I'll look at getting a better supply if this project isn't stable enough or has too many problems.

Veroboard Layout

The design is pretty simple and is based of two schematics I found online. I don't need more than +/- 15 Volts at one Amp so I am leaving it to the LM317 to handle over current protection(it shuts down). I tested it running two fans on both rails and the small heat sinks seem to provide enough heat dissapation.

Testing the regulator heat sinks with 12 Volt fans.

Preparing the Case

Using a carboard template, I cut out the holes for the panel meters, bananna jacks, and multiturn potentiometers with a dremel tool and a cutting disk. I made sure to mount everything in the carboard first and test it out with most of it wired up. I've run into issues before where my layout seemed fine until I actually started using the device.

Inserted panel meters, jacks, and knobs.

Completing Component Installation

It ended up being easier to use batteries for the two panel displays. I could have made an additional board to power the meters but decided to use a relay to turn them on/off since the battery usage is low. There are two fuses and I didn't solder the transformer wires to the board. Instead, I opted for a screw terminal jack.

Top view of the final layout with everything installed and fastended to the case.

The switch in the back powers the unit from the mains and turns on the panel meters, then the front switch is used to actually enable the front jacks to the psu. This is similiar to a "load" switch except I can't see the voltage level until I turn it on. I can see the need for that now.

Calibration and Testing

I knew the Fluke meter was accurate and used it for testing. The BK Precision DMM was calibrated but seemed slightly off(but still within spec) during testing. This was all still under load too and I noticed the panel meters flickered some. I'm hoping it's not as bad when I use it for OpAmp circuits since that will draw a lot less power.

Taking measurements with a Fluke 115 and the BK DMM in the background.

I haven't put the device through any stress tests besides leaving it on a while and running some fans. Anything that draws a lot of current will immediately trigger the protection built into the regulator chips.