DIY Power Supply for Gel Electrophoresis

I was using my variable power supply modification of an ATX supply for gel electrophoresis but thought a dedicated power supply would be better. And just for the heck of it, I decided to make one out of passive components only, no power supply chips or power transistors or digital displays. Go retro to the early days of electronics.

I needed multiple outputs, 12-120V would be ideal. Current limiting would be useful. The power supply for an electrophoresis unit does not need to be regulated but knowing the voltage and current would be valuable.

The images show the version I ended up building. It looks rather cute with its old fashioned moving coil panel meter, a neon light, and mechanical switches. The final specs are not too shabby.

The basic principle is to divide the AC voltage down with a voltage divider made out of capacitors so heat dissipation is kept to a minimum. If a chain of three equal value capacitors are connected in series then the voltage at the junction of the lowest capacitor in the chain will be 1/3rd of the full AC voltage, 2/3rd at the next junction and 3/3rd at the top of the capacitor chain. the total available current is determined by calculating the individual reactance of each capacitor, summing up the reactance and and dividing this from the total voltage applied to the capacitor chain.

As I used 2.2uF capacitors, the reactance would be 1200 ohms at 60 Hz. For three capacitors in series the total reactance would be 3600 ohms and the available current would be 120V/3600R = 33mA.

Just to make sure it works, I mocked up a simple circuit and connected it to an electrophoresis tank to see if it works. Yes, the voltages were dropped as expected and DNA moved in the gel.

To make the breadboard into something practical, I would need a fuse, a power switch, a switch to select voltages, and a meter to show how much current would be drawn. The component count is quite low. If everything is sealed from a user, the circuit would be relatively safe but for electrophoresis there is a risk someone may dip their finger into the electrophoresis tank so to ensure isolation from mains an isolation transformer would help. I could not find such a transformer at low enough rating and at a reasonable price so decided to hack one by using two 12 volt transformers by connecting them back to back. The final circuit diagram is shown. I also decide to double the capacitance by putting two capacitors in parallel resulting in increased current capacity and adding one more layer of capacitors for a total of four voltages: 12 V from the transformer secondaries, 24V, 48V, 72V and 96V from the capacitor bank (at load) and the total current would be about 50 mA from the capacitor banks and about 500mA from the 12V secondary.

I decide to build a case that was 3 inches wide 6 inches tall and 6 inches deep. Cut the panels out of plywood and assembled them together. Had ordered X2 2.2uF 275V non-polarized capacitors from ebay (yellow) plus, dpdt toggle switches, again from ebay (that turned out to be defective) and other parts from my parts bin including a VU meter. My goal was not to use any active devices like transistors and IC's. An exception, I guess, is the bridge rectifier. For the transformers I ended up using Ikea light bases which were gathering dust at home. The transformers were 120V:12V so I could connect these back to back to make an isolation transformer.

The back panel held just two components, a fuse holder and an IEC socket. These were placed into position and wired according to the circuit diagram.

Required holes were cut into a thin plywood panel which was then painted with black acrylic paint, labels (printed on an inkjet) were applied and the meter switch, the voltage selector switch, the neon indicator, the power switch, and the two banana sockets were added.

The old VU meter was transformed into a milliammeter by: Taking a picture of the meter face with a ruler, importing this image into the Xara drawing program and then drawing a new scale on the old image, scaling it to the right size and printing them out.

I had analyzed the meter by measuring the full scale deflection current and the voltage drop across the meter at full scale defection. Values were 0.5mA FSD and 0.197V. That means the resistance is R = V/I = 0.197V/0.0005A =394 ohm.

To calculate the shunt resistance required for the meter to read 50mA (0.05A) I used the formula, Resistance of shunt = Meter FSD voltage/(Required current readout x FSD current) = 0.197V/(50mA x 0.5mA) = 3.98 ohms

So if I put a 3.98 ohm resistance across the meter I would convert a 0.5mA into a 50mA meter.

The capacitors could provide a maximum of 50mA but the 12V directly from the transformer was capable of supplying upto 500mA so I decided to add a switch that would allow the meter to measure 50mA or 500mA. For 500mA, a 0.39 Ohm shunt is required. My pile of rescued resistors delivered a 3.3 ohm + 0.68 ohm to equal 3.98 ohms for the 50mA shunt. To get to 0.39 ohms, the closed option was to put a 0.47 ohm resistor in parallel to the 3.98 ohm series resistors giving me 0.42 ohms by calculation. Close enough. The switch, resistors and meter were wired as shown in the circuit diagram. tested. worked.

I first prewired the bridge rectifier and tried to glue it to the front panel but without success. The pin numbers were hard to see on the switch so I scribbled a yellow marker t make them more prominent. I wired the switch pin terminals according to the circuit diagram. Switch terminals 2-5 were connected together. This single wire would then be connected to the bottom of the chain of capacitors. The two central terminals of the switch were wired to the AC in pins of the bridge rectifier. Separate wires were attached to pins 1, 7, 8, 9, 10 and 11. Pins 6 and 12 were unused. These separate wires would connect to the 12 volt terminals of the transformers (pin 1 and 7), and to the capacitor chain (wire from pins 2-5, pin 7, 8, 9, 10 and 11).

Fortunately, the capacitors fit end to end in the top of the enclosure. So I taped these together as one block, connected the leads from two capacitors together (in parallel) and then joined the leads of one pair of parallel-connected-capacitors to the next pair of capacitors and so on. A 1 Meg ohm 1 W resistor was connected across the whole capacitor chain to discharge the capacitor chain when power was turned off. A 120 ohm 3W resistor was connected to the top end of the capacitor chain. The other end of this resistor would be connected to the 120V AC output from the transformer. Each junction of the capacitor chain was then connected to the corresponding wire of the voltage selector switch. The top of the capacitor chain at ~ 100V was connected to switch pin 11, 2nd junction at 72V was connected to pin 10, 3rd junction at ~48V was connected to 9, and 4th junction at ~ 24V was connected to pin 8. The bottommost part of the capacitor chain was connected to the wire that had been soldered to switch pins 2, 3, 4, 5 and to another wire that would be connected to the second 120V output of the transformer. The wired capacitors were attached to the side panel with double sided tape.

Had glued the two transformers to a small plywood piece and wrapped a wrist rubberband around the transformers so they were a bit easier to handle. I then soldered the two 12V secondaries in parallel and wired the 12 volt transformer output to the voltage selector switch terminal 1 and 7 (as in the circuit diagram).

The 120V AC input to one of the transformer was wired to the power switch outputs. I also connected the neon indicator across the transformer AC input.

The 120V AC outputs from the second transformer was wired across the capacitor banks.

The earth wire from the IEC socket was soldered to the laminations of both transformers and then covered with the black rubberband.

The jumble of wires were taped together to clean up the mess.

After connecting the transformer in, I tried to power the system up but just could not as the power switch would not toggle. I cut the wires connecting this switch to the rest of the circuitry (first image highlights the cut ends of these wires). This was quite irritating and I did not want to replace with another of the same kind of switcg so went through my parts bin and found a heftier switch. The barrel of this switch was a larger diameter so expanded the hole in the front panel with a reamer and then spliced this switch in. Tested the power supply and it now worked.

After replacing the power switch repeated the tests and the expected voltages showed up. The hooked it to the DIY gel electrophoresis setup with and agarose gel containing DNA and tested it with that. DNA migrated as expected. Re-applied the label for the voltage selector switch and another coat of paint to some area.

So quite happy with the small retro- electrophoresis power supply that is essentially short circuit proof with current limiting.