For those who don’t already know, a variac is an AC transformer which has an adjustable output voltage. It’s useful when you’re not sure a device is working properly and you want to power it up slowly so as to not destroy components. With a bridge rectifier and a couple filter capacitors, it also allows me to have any DC voltage I need. I’ve used this many times to test LED strips. I chose this particular variac, a PHC Enterprise model SC-10T, because it is isolated. I wanted a variac, but I needed an isolation transformer so I could safely use my oscilloscope. This allowed me to have both, and because I opted for a scratch and dent model, I got it for for about $150. Even today, it’s still about $200 with shipping. The bad thing was that the volt meter was damaged and I couldn’t find one which was the same size. Instead of letting that ruin it for me, I decided to modify it and add a digital display. Here’s what the variac originally looked like.
And here’s pics of the damage mine had
That little dent saved me about $50.
For the mod, I ordered a model DL69-2042 digital volt/ammeter from Aliexpress.
What I actually received was a model D52-2048.
Some pictures of the internals.
I ended up getting a full refund, and I got to keep the meter. Truthfully, I liked the D52-2048 better.
Here’s what the inside of the panel originally looked like.
There was no way this meter would fit unless I modified it. I literally hacked away every bit I didn’t need. Then I cut the opening to make it fit. All I had was tin snips and a file, so at this point it’s pretty rough.
The problem with this meter was that it wouldn’t go below 80 volts. The reason was because it was powered by the line it was sensing, so at some point there wouldn’t be enough voltage to drive the meter. The solution was to provide the meter with it’s own power source. Simply powering the meter from the primary side of the variac would have worked, but would have caused it to no longer be isolated. Instead, I used a mini switch mode power supply to provide the 5 volts DC to power the meter.
In order for that and the sense board to fit, I had to make changes to the remaining components in the panel. So, I decided to eliminate the 10 amp fuse holder and replace the switch with an illuminated 10 amp panel mount breaker. Model number was W51-A121B1-10, and cost about $5 at Digikey.com.
I didn’t have a welder so I just screwed a plate over the old holes, and then filed them flush on the other side.
I had to make brackets so the panel could be mounted. Originally, two screws went through the front but there wasn’t room now. Plus, it just looks better with them on the side and bottom.
I used Bondo to make the front smooth and ready for paint.
I had to strip everything to bare metal because the new paint made the old paint bubble. I primed everything first, but forgot to take pictures of that step.
I made a schematic of the power and sense circuits.
After some testing, I found the supposed 1% resistors to be way out of specs. I upgraded the ones in the sense circuit to 0.1%, and replaced R6 with a 866Ω resistor in series with a 200Ω, 1W, 25 turn cermet pot. This required me to cut the trace, as seen in the second image. I also cut the current sense circuit off the board to make it smaller.
Unfortunately, the meter still wasn’t accurate and, though I could adjust it with the pot and get it perfect, it wouldn’t be perfect across the entire voltage range. I removed the pot and 866Ω resistor and put a jumper across the cut in the trace. I put in a 0.1% resistor to replace the last 1 Meg resistor in the circuit. I probed the voltage sense circuit with my oscilloscope and this was the waveform I observed. Yes, it’s just a 60Hz sinewave. I was actually surprised, I’d expected it to be DC because of the bridge rectifier. It was actually two circuits which were in parallel.
I removed everything from the board which wasn’t part of the voltage sense circuit. I left the bridge rectifier in place because I thought it might need a diode in the circuit. This is the waveform I observed after doing that.
I realized at that point that the bridge rectifier needed to go, and I needed a jumper to complete the circuit between neutral and R6. This is the board with everything removed and the jumper in place. I also replaced a few SMD resistors on the control board with 0.01% ones.
And this is the waveform I got. Exactly what it was before.
I used some self-stick nylon standoffs to secure the mini SMPS and sense board, then wired everything up. It was a tight fit. I used a lot of hot glue to secure all the wires. I also disconnected the chassis ground because it defeated the isolation.
You’ll also notice I moved the power cord to the side, replaced the receptacle with a black one, and carefully painted the bezel of the meter with black model paint. I’m very happy with how the completed assembly looks.
Initially, it wouldn’t go below 0.7 volts.
But a bit of filing on the stop, and I was able to get it down to 0.2v. I couldn’t go any farther because the brushes would have dropped off a ledge. I got it down a half a volt, I can’t complain.
It is very accurate across the voltage range. It’s within 0.03 at the lower voltages. By the time it gets to 70 volts, it’s off by about 0.2 and at it’s maximum, about 145v, it’s off by about 0.3v. I’m sure I could play with it and get it perfect, but I think this is close enough. It’s not like I’m planning on using it as a multimeter.
What’s really nice is that I can see just how many amps my device under test is drawing, without having to use my multimeter.