Hello forum, I'm here because I have to build a 70kv, 18 stages C-W multiplier with cathode grounded. I can't use liquid insulators so I don't know what to use, polyurethane, epoxy, nothing...
What do you guys recommend?
I find this kind of thing is hard epoxy encapsulated. In air, I find a lot of losses from corona discharge because parts are close to each other.
Potting something, well it's not repairable.
What output current you are trying to achieve. 70kV/18 stages =3.9kV per stage, am I right. Seems high, as far as diode/cap ratings.
I find this kind of thing is hard epoxy encapsulated. In air, I find a lot of losses from corona discharge because parts are close to each other.
Potting something, well it's not repairable.
What output current you are trying to achieve. 70kV/18 stages =3.9kV per stage, am I right. Seems high, as far as diode/cap ratings.
Peak current when capacitors are charging is 40mA, nominal current is only 2mA.
Edit: and the diodes are 2CL2FM
The 70kV output terminal is where the insulation is really needed. Usually a HV wire exits the potting compound - no screws, terminal blocks etc.
For a commercial product, a potted pcb would be best.
For hobby stuff, you can build this point-point wiring on perfboard in air, example:
250kV HV multiplier.Image below from
High Voltage Shop of multiplier PCB with 15mm diodes
There are gel-type silicone compounds which might be suitable. They don't set hard and can be (messily) dug out for service if necessary.
Silicone is used on some multipliers. But a lot just use lots of air space.
What frequency of operation and capacitance per stage are you aiming for?
With 18 stages and 2 mA load current, there will be a lot of voltage drop under load, and any stage beyond the tenth will likely not add much, for any practical input frequency and capacitor value. Practical multipliers often have less than 10 stages, I would aim for 6 or 8 maximum.
What frequency of operation and capacitance per stage are you aiming for?
With 18 stages and 2 mA load current, there will be a lot of voltage drop under load, and any stage beyond the tenth will likely not add much, for any practical input frequency and capacitor value. Practical multipliers often have less than 10 stages, I would aim for 6 or 8 maximum.
Frequency will be between 350kHz and 600kHz. I just want to build a multiplier like this one:
http://www.cei-xray.it/content/uploads/prodotto_image/ceiTrasformatorA3_1.pdfEdit: I don't know exactly what people understand as stage. When I said 16 I meant a CW multiplier with 8 diodes and 16 capacitors.
One stage is two diodes plus two capacitors, so that would be a four stage multiplier, with two capacitors in series per stage. This is a much more reasonable number, and very typical for dental X-ray supplies, see the attached image.
The main challenge with operating it in air is that high voltage diodes often don't have enough creepage distance across the package to stand off the rated voltage, so you need something other than air for insulation. If liquids are out, that leaves gases and solids. Gases are difficult to get (SF6) or have other dangers (butane), and sealing them is a pain. Silicone or epoxy can be an option, but they will not remove heat as well as oil so the available power is lower. How much power are you aiming for, and for how long? A dental x-ray tube can typically handle 500 watts for a few seconds, and 250 watts for a few tens of seconds, medical diagnostic tubes can handle more.
Note that 350 - 600 kHz is a pretty high frequency for a multiplier like this, and diode recovery might be problematic, a more typical frequency is in the 50 - 200 kHz range, the multiplier board with the onboard transformer in my picture was made for 150 kHz operation.
One stage is two diodes plus two capacitors, so that would be a four stage multiplier, with two capacitors in series per stage. This is a much more reasonable number, and very typical for dental X-ray supplies, see the attached image.
The main challenge with operating it in air is that high voltage diodes often don't have enough creepage distance across the package to stand off the rated voltage, so you need something other than air for insulation. If liquids are out, that leaves gases and solids. Gases are difficult to get (SF6) or have other dangers (butane), and sealing them is a pain. Silicone or epoxy can be an option, but they will not remove heat as well as oil so the available power is lower. How much power are you aiming for, and for how long? A dental x-ray tube can typically handle 500 watts for a few seconds, and 250 watts for a few tens of seconds, medical diagnostic tubes can handle more.
Note that 350 - 600 kHz is a pretty high frequency for a multiplier like this, and diode recovery might be problematic, a more typical frequency is in the 50 - 200 kHz range, the multiplier board with the onboard transformer in my picture was made for 150 kHz operation.
Diode recovery for the 2CL2FM is faster than my switching frequency, so I don't expect it to be a problem. Power is below 200W for a few seconds, that leaves 2mA of current so I don't expect diodes to heat either, because the voltage drop is 35V, 35*2*10^-3=70mW converted into heat in each diode.
Epoxy resin is a bit expensive, can I use polyurethane resin? Nominal voltage is 70kV.
Epoxy resin is a bit expensive, can I use polyurethane resin? Nominal voltage is 70kV.
It's hard to believe that plain old air won't provide enough insulation at only 70 kV. That's not a very high voltage in the grand scheme of things.
If you do use two-part epoxy, the secret is to avoid buying it by the milliliter as "epoxy." Instead, buy it by the quart/liter in home-improvement stores in the floor coating department. Same stuff, very different price.
Air can work if humidity is low, ozone does not build up and there is decent clearance on the 70kV.
I find the sharp points on the trimmed leads has a lot of corona, not so much between CW stages, but end-to-end.
This guy used a lunch sandwich box but filled it with oil....
then went to 200kV
Wire in air with good mechanical connections which are then soldered and allow no sharp points, coat exposed conductors with corona dope, and then vacuum pot the entire assembly in wax.
Thank you guys very much for your advices. Today I went to the store and I bought a can (do you say "can" for this?) of polyester resin. I can try to leave a small piece of this resin to dry and then connect it to a flyback transformer to see how it can handle hv. I think I will just submerge the pcb in resin.
I have tried Google to find articles about polyester resin but I have not found any useful data.
In general, with wax or resin potting, the trick is to avoid bubbles. From elementary electrostatics, the E-field inside an air bubble surrounded by bulk material with dielectric constant c will be c times higher than the field in the bulk material, and the dielectric strength of air is also less than that of a good insuilator. When the air breaks down, it can damage the material and possibly lead to a track of carbon from one electrode to another.
I seem to remember from some HV mailing lists that LDPE (cut up and cleaned milk jugs) can be dissolved into wax to make something a bit more robust than plain paraffin wax (not sure what its called outside North America, where kerosene here is called paraffin there...) I can't seem to find any solid references at the moment though.
EDIT:
http://www.pupman.com/listarchives/2002/May/msg00378.html http://www.pupman.com/listarchives/2002/May/msg00527.html of course it was pupman mailing lists for something HV related.
Looks like it can be either hot glue or polyethylene strips that you can add to the wax.