How to Make a USB-C Based -350-0+350 VDC Lab Power Supply

Inspired by my regular voltage (1-25V) lab power supply fed from a USB-C power source, thought I would build a high voltage DC supply delivering ~40-350VDC. Wanted to make this power supply in the same 3x3x2 inch form factor. Intend to replace my large electrophoresis power supply with this unit.

Parts needed are shown in the wiring diagram. The USB-C HV power supply needs the

1. USB-C decoy board that will get the required 20V from a USB-C source
2. A high voltage module with an input range of 10-32V at < 5A and an output of +/- ~40-390V at < 0.2A at maxm voltage (part no. on the module SKU: 7768)
3. A volt-amp digital meter (the one I had is 0-100V and 0-10A)
4. A 1/10 voltage divider that can handle 400+ volts. I made mine from 1/2W resistors in series.
5. A 50k Potentiometer
6. 3-way binding posts, a switch for power, wires, screws, etc.
7. A 3D printed case to hold the power supply

Small boards often listed as USB-C decoy or dummy modules can be set to get 5V, 9V, 12V, 15V or 20V from most USB-C power sources. The input of the USB-C decoy connects to the USB C power source and the output connects to the load.

I set the microswitches on the USB-C decoy to deliver 20V and confirmed that the board did deliver 20VDC.

After taking measurements of the parts, I drew out the design in the 'Parts Design' workbench of FreeCad. Text labels were added by using the 'Draft' workbench to place the text and extruded these texts in the 'Parts Design' workbench. The designs were exported to stl files and then sliced to get the gcode files to print on an Ender 2 Pro with black PLA. Settings were 3 lines for wall thickness, 30% infill, support was trees and did not use any rims or base. Temperatures were 200C for the extruder and 50C for the support base.

While these designs were printing, I removed the variable resistor from the HV power supply module and soldered in wires that were attached to a 50k ohm potentiometer.

The 3D printed top plate or 'lid' did not show any of the text that I had extruded, so an option was to laser etch the PLA plate or etch a separate black transparent layer that I would glue on. The black transparency seemed more appealing as it would cover the digital meter LEDs.

I used a laptop privacy screen and cut and etched the design on a K40 laser cutter. The privacy screen film was surprisingly resistant to being etched ad cut. I etched at a setting of 8 units which was too low, should have done 12 units. And I cut at 14 units. Again too low as had to do another pass to cut this film. Apparently the privacy screen is a laminate of multiple layers. Worth exploring what each layer is made of ...

After the laser cut and etch I did my usual fill with acrylic paint and then glued this film onto the PLA plate with E6000 glue.

The USB-C decoy had wires attached to the outputs and then was latched into place in the 3D printed base. The positive lead from the USB-C decoy was attached to the power switch and the Gnd lead was attached directly to the screw terminal input of the HV module. The other connection from the power switch was connected to the +ve of the HV module input. An additional wire was added to the +ve and to the Gnd inputs of the HV module. These additional wires would be connected to the meter later on.

Eight 91.9k ohms, 1/2 W resistors were soldered in series with the last resistor was soldered to a 100k ohm variable resistor to form a chain of 9 resistors. These series reistors were used to spread the voltage burden across them. Assuming a maximum of 500V output would mean that each resistor would see 500/9 = 56 volts only. Well within the 100-200V rating. I could have gotten away with using only five resistors instead of the nine I used.

The resistors were wrapped in black electrical tape and folded into a rectangular shape. The resultant voltage divider would connect at 3 points; one end to the red terminal that would be at >300VDC, the other end to GND and a tap two resistors from GND to the voltage sense input of the volt-amp meter.

The volt-ammeter was attached to the top/front panel and was held in place with a few strategically melted PLA scraps. A small hole was burnt into the top panel so the screw for the variable resistor for the voltage divider could be accessed to calibrate the voltage reading. The variable resistor was held in place with scraps of PLA melted in place with a soldering iron. The rest of the voltage divider was held with double sided tape. The banana plug sockets were attached and the modules were wired as shown in the diagram except for some last minute changes that I made.

Change 1: I got rid of the blue banana socket (that would have been at negative volts) and just kept the black socket (GND) and the red socket (+ve) as was worried about reverse current flow through the ammeter part of the volt-ammeter.

Change 2: I tied three of the GNDs together at the HV GND screw terminal. The meter supply Gnd (thin black wire), the ammeter thick black wire and the voltage divider Gnd (white in the photo) were connected to the HV output middle screw terminal.

Change 3: The black GND banana socket/terminal was connected to the thick red wire of the ammeter.

After most of the wiring was completed and checked the HV power supply board was screwed onto the standoffs that in the 3D printed enclosure. The power switch was attached to the top panel and the panel was then attached to the base with two screws.

I attached the USB C plug to my little power supply and after a small pause the meter came to life. Relief!

Calibrated the voltage readout by adjusting the small screw of the voltage dividers variable resistor till the voltage readings matched the multimeter readout.

The decimal point in the voltmeter readout was irritating so covered that with a sliver of black tape. Now the meter reads 50 V instead of 5.0V. Remember, the voltage divider had divided the output voltage by 10x and then fed it to the voltmeter which read 1/10th of the output voltage.

Did a quick test with an incandescent 110V lamp to see if the power drain is shown accurately. The readout matched a multimeter-ammeter placed in series with the bulb. Impressed at the performance of this volt-ammeter.

Yes, so this little amazing high-voltage power supply plugs into a USB C power cable and delivers approximately 40 to 350V to ~ 100W!