Circuit analysis attempt of a commercial +/- 390V DC power supply

In the past several months I have been using a commercially-available DC-DC high voltage boost converter for plasma and plasma electrolysis experiments of various sorts. Several vendors sell this model for about 10–15$ on Amazon or Ebay, in a few variants with differing build quality and features. The one I have has two high voltage outputs of opposing polarity. At +/- 390V this means up to 780V.

Here are a few:

• https://www.amazon.com/Qianson-Converter-±45V-390V-Adjustable-Capacitor/dp/B01IVMU2XI/
• https://www.amazon.com/LM-YN-H…-Capacitor/dp/B01MYLWHZW/
• https://www.amazon.com/Comidox…Adjustable/dp/B07PGMBG1Q/

Quote from Specifications

Name: High Voltage boost converter module

Properties: non-isolated step-up module

Input voltage: two input voltage range selectable, (jumper selectable through the back of the PCB, default setting 10-32V input)

1, 8V-16V input

2, 10V-32V input

Input Current: 5A (MAX)

Quiescent current: 15mA (12V step up to 50V, the higher the output voltage will increase the quiescent current)

Output voltage: ± 45-390V continuously adjustable (the default output ± 50V. This module has regulator function, after adjusted output voltage is stable and unchanging, not with changes in input voltage changes)

Output Current: 0.2A MAX

Output Power: 40W (peak 70W)

Working temperature: -40℃ ~ + 85℃

Operating frequency: 75KHz

Conversion efficiency: up to 88%

Short circuit protection: Yes. (Input 10A fuse, at the output please do not short circuit arc, it may damage component, if you need please add power resistor in series for limiting)

Overcurrent protection: Yes. (Input current exceeds 4.5A, reducing the output voltage)

Overvoltage protection: Yes (Output voltage exceeds 410V, lowering the output voltage) Input reverse polarity protection: Yes (non-self-healing, reverse burning fuse, you have better not reverse it.)

Dimensions: 60(L)*50(W)*22(H)mm

Display More

To the best of my abilities I have tried looking more in detail at the components installed on the circuit board of the one I have. Unfortunately I couldn't manage to measure the capacitance of the small SMD capacitors. Here is what I came up with (source image also attached):

Besides capacitors and resistors (as well as one 50 kOhm potentiometer) there are:

• 1x 5A/10A fuse at the input. This is mostly to prevent a reverse voltage from being applied to the converter (which happened to me once).
• 1x 78L09 voltage regulator with up to 30V input and 9V output.
  • This could be bypassed by desoldering the 0 Ohm jumper from the 10–32V position to the 8–16V position.
• The Input goes into a UC3843A PWM controller which supports 500 kHZ and to 8.4–28V input. This likely defines the minimum voltage allowed by the HV converter
• These ones I believe are mostly required for PWM controller logic:
  
  • 1x LM358 op-amp
  • 1x 1AM NPN transistor
  • 2x 431 Triode
  • 3x S4 Schottky diodes
• 1x RU7088R N-channel power MOSFET
• 2x US3M ultra-fast rectifiers rated for 1000V DC voltage and 3A

Input DC appears to be PWM-pulsed into a small transformer with some voltage/current/load-limiting logic, and the transformer AC output very simply rectified into two 400V 10uF capacitors for the +/- outputs.

One thing I haven't tried yet is bypassing the 9V output voltage regulator at the input stage. Perhaps the circuit could work more effectively with e.g.12V. Also, I'm not sure if the 0.04 Ohm resistor could be safely bypassed to increase effective power or capacitor charging speed.

Rob Woudenberg

The voltage range is fine, but a higher current output at higher voltages would be desired. Since I'm using active cooling (currently, a 70mm computer fan from an old AMD heat sink, running at 12V), perhaps the converter could be pushed a little bit more.

The second reason why I posted this partial analysis that I was curious knowing why/how I'm apparently getting significant voltage spikes back to the low-voltage inputs under certain conditions, when using this device in a plasma electrolysis setup.

Curbina

The voltage spikes seem to be real and cause issues to other electrical appliances connected to the house wiring, if I use a mains-connected 12V power supply for powering the converter. These issues are mitigated if instead I use it with 12/24V battery power, but RF emission still remain strong (only under certain conditions, though).

[Speaking of different matters concerning the device] A while ago I made a graph of the current-voltage relationship with it, when powered with 12V; see graph below. I removed unnecessary graph information. The converter apparently supports up to 200 mA output, but only at relatively low voltages. This seemed consistent across several readings under various conditions.

Rob Woudenberg

Right now the converter's power MOSFET is mated to the heat sink with just a rubber thermal pad. I already have high-thermal conductivity paste and 0.10 mm mica shims, so perhaps I could try this suggestion. I think however that there are hardware limitations lowering power/current, more than heat dissipation.

For what it's worth, I tried bypassing the "78L09" 9V output voltage regulator by relocating the 0-Ohm jumper from the 10–32V to the 8–16V position as marked on the converter's PCB. Operation seems subjectively a bit steadier, and the converter appears to be acoustically quieter than before. I'm still getting intense voltage spikes back to the input under certain conditions. RF emissions from the same plasma tests seem largely unchanged.