# Winding transformer for -40KV power supply.

Last edited:

#### Lash432

##### New Member
Would a full wave rectified, 40Kv peak signal be acceptable? If so the reservoir capacitor can be eliminated.
Yes it should be, however, I believe I would be using half wave CW multiplier circuits. I am wondering if a re-wound MOT transformer may be able to supply the 10kv needed (however I am once again wondering about safety). after that it is simply a matter of putting that voltage through a voltage quadrupler correct?

If so then the capacitors for that would be about .42 uf right? I also take it the whole of the CW multiplier will need to be separate and under mineral oil to prevent arcing. Would I use a circuit board or breadboard to build the multiplier? The diodes and capacitors also need to be rated for about 25kv correct?

#### KeepItSimpleStupid

##### Well-Known Member
"We" got DC plasmas with about 300 VDC at about 2-5 mTorr of pressure, Argon atmosphere if my memory is correct. Occasionally, you had to raise the pressure to get a strike. Your using a MFC (nass flow controller) for the Argon and a pressure controller. The main pump was a Cryopump.

==

You generally put the current measuring ground potential. Did simple current-voltage converter stuff to measure electron beam current on an SEM with a home-made faraday cup. This was used in an EBIC (Electron Beam Induced Current) measurement setup that I did. Currents were tiny though (pA etc)

Edit: fixed typo - flow

Last edited:

#### Lash432

##### New Member
Yes it should be, however, I believe I would be using half wave CW multiplier circuits. I am wondering if a re-wound MOT transformer may be able to supply the 10kv needed (however I am once again wondering about safety). after that it is simply a matter of putting that voltage through a voltage quadrupler correct?

If so then the capacitors for that would be about .42 uf right? I also take it the whole of the CW multiplier will need to be separate and under mineral oil to prevent arcing. Would I use a circuit board or breadboard to build the multiplier? The diodes and capacitors also need to be rated for about 25kv correct?

To add to this, I went into an an online circuit simulator and made this. This is all wired to have a positive output. I was fiddling around, but I could not figure out how to get the output's polarity negative.

CWM2.png

"We" got DC plasmas with about 300 VDC at about 2-5 mTorr of pressure, Argon atmosphere if my memory is correct. Occasionally, you had to raise the pressure to get a strike. Your using a MFC (nass fkiw controller) for the Argon and a pressure controller. The main pump was a Cryopump.
Nice! I am planning on experiment with hydrogen, helium, and argon, at pressures form 1-10 millitorr. I am however using a diffusion pump.

#### Attachments

• 11.1 KB Views: 247

#### Lash432

##### New Member
To add to this, I went into an an online circuit simulator and made this. This is all wired to have a positive output. I was fiddling around, but I could not figure out how to get the output's polarity negative.
Well, I sure have become obsessed with designing this power supply now. After doing much more experiment in the simulator, and research, I finally figured out how to get a negative DC output. Here is the picture of the design in the simulator, along with what I plan on doing for voltage measurement.

This is with the AC input being 10kv. For whatever reason it is saying the voltage is unusually high. I believe this is an error

#### spec

##### Well-Known Member
I am wondering if a re-wound MOT transformer ...
Yes, a microwave oven transformer (MOT) could be re-purposed for your application, but MOTs do not have the best characteristics for a bench (voltage) power supply. And the core dimensions and arrangement are not ideal for an extra high tension (EHT) secondary.

Also, if you used the original primary of a 120V, 60Hz MOT, the MOT would overheat in your application and the output sine wave would be distorted (limited). This is because MOT cores are run into saturation in microwave ovens, to minimize iron and copper and thus save money.

The solution is,
(1) Wind a new primary with twice the number of turns.
(2) Run the primary at around 60V RMS by limiting the output voltage from the Variac that you intend to use.
(3) Get a 240V RMS, 50Hz MOT (UK version for example) and run it at 120V RMS, 60Hz. The higher frequency will be OK.
(4) Use two similar 120V, 60Hz MOTs with the primaries connected in series across the 120V 60Hz mains supply. This approach has the advantage of giving more room and flexibility for the EHT design.

Also, with an MOT of any type, the magnetic shunts should be removed for your application to reduce the leakage inductance and, thus, reduce the secondary output impedance, which would improve the secondary output voltage regulation.

spec

Last edited:

#### spec

##### Well-Known Member
Well, I sure have become obsessed with designing this power supply now. After doing much more experiment in the simulator, and research, I finally figured out how to get a negative DC output. Here is the picture of the design in the simulator, along with what I plan on doing for voltage measurement.View attachment 103270

This is with the AC input being 10kv. For whatever reason it is saying the voltage is unusually high. I believe this is an error
Nice work Lash.

spec

(PS: have a look at the power dissipation in the 1M resistor connected to the voltmeter.
http://www.digikey.co.uk/product-detail/en/ohmite/SM108035007FE/SM108FE-5000M-ND/824231)

#### spec

##### Well-Known Member
I also take it the whole of the CW multiplier will need to be separate and under mineral oil to prevent arcing.
It is doubtful that immersion in insulating oil will be necessary, but a conformal coating, or even potting may be a wise precaution, especially if the EHT Generator is to be for general use where high humidity will be encountered.
Would I use a circuit board or breadboard to build the multiplier?
Afraid not. There will be some very high voltages associated with the secondary electronics, so EHT wire will need to be used for any cable runs. Electronic components will need to be mounted on high-voltage ceramic stand-off pillars, and the components will need to be well spaced out and laid-out to minimize voltage/space gradients between components. This means that the EHT electronics will need to be much bigger than you may expect.

In any event, it is advisable to place the EHT circuitry in an earthed metal case, but make sure you provide full access to all components and connections for servicing/development- don't bury the components in a box; instead, mount the circuit on the lid and use the box as the lid.

Preventing arcing due to the high voltage will be the main challenge with this project. 40KV is one hell of a high voltage and it will arc over if it gets the slightest chance. If you are not fully familiar with high voltage design, it would be wise to investigate the sustaining voltage of air, at the various temperatures and humidities that the EHT generator will be exposed to. Also, check the insulating properties of any components and materials that you are considering for the EHT circuits (for example normal resistors will only take 200V DC).

One advantage of using a low frequency, 60Hz to 120Hz, EHT generator is that arcing is less of a problem than with a high frequency EHT generator of, 10KHz to 100KHz.

spec

Last edited:

#### ericgibbs

##### Well-Known Member
I agree, safety for me, and especially those around is the most important factor to consider. I do actually have people who are going to help me this. I have been recently talking to a guy who does work with x-ray machines, and also does a lot of electronics projects. I also have been talking to the dean of engineering at my local university, and he said he is going to put me in touch with some people who can help with certain aspects of the project.
As to asking on this forum, I absolutely was not going to use this as my only recourse, but it does help me learn some information, get advice, and put things into perspective on what I am trying to do, while also introducing new ideas.
hi L432,
Pleased to hear that you have already taken the wise precaution of seeking 'face to face' professional advice, by discussing your HV project with a local Engineering University.
Eric

#### spec

##### Well-Known Member
POST ISSUE 06 of 2016_12_26

Hi Lash,

Below is a schematic showing a notional circuit for a 40KV EHT generator for your consideration.

spec.

NOTES
(1) The current through R1 and R2 will discharge the capacitors when the mains is removed.
(2) The peak to peak ripple voltage is defined by 166.6/C, where C (C1 and C2) is in uF. With C1 and C2 = 470nF the peak to peak ripple voltage will be 723V.

COMPONENTS

(1) Diodes are 40.4KV min, 40mA min. http://www.semtech.com/images/datasheet/schjxk.pdf
(2) Resistors are 20.2KV min, 2.2W min. http://www.digikey.co.uk/product-detail/en/ohmite/SM108035007FE/SM108FE-5000M-ND/824231
(3) Capacitors are 20.2 KV min. capacitors can be made up from any number of smaller capacitors in parallel.

Last edited:

#### KeepItSimpleStupid

##### Well-Known Member
The diodes are usually stacked in the systems I worked on. It takes multiple diodes to get to the voltage you need. The diode drops were on the order of 10 V on the systems I worked on.

spec: We generally don't switch the neutral in the US. I've only seen one system that did and, guess what, there was a High voltage thing in it. It wasn't the main breaker either.

#### ronsimpson

##### Well-Known Member
The diodes are usually stacked in the systems I worked on.
In production we stack diodes that have the same amount of leakage current. Usually we pick diodes with the same date code.
In some very old products we put resistors across each diode to help keep the voltage spread out evenly across all the diodes.
When the factory makes 10,000 volt diodes they probably stack up 10 to 15 diodes that came from the same silicon wafer and probably form the same side of the wafer. That way they will work well together.

The problem is; the leaky diodes have little voltage across them and the good diodes have much more voltage. With resistors we (more or less) cause all the diodes to have the same amount of reverse leakage. (reverse current flow)

#### spec

##### Well-Known Member
From a person that has made many 40kv power supplies;
Where are you going to get 60kv diodes?
Parts like this are hard to find.
http://www.semtech.com/images/datasheet/sckvxk12.pdf
Hi Ron,

Yes, you are right. I meant the circuit to be conceptual rather than practical and should have made that clear. With a 60Hz mains frequency a practical design is a problem from a component view.

I recon that the diodes need to be 40.4KV minimum peak inverse voltage (PIV) rather than 60KV. The Semitech SCH45KV has a PIV of 45KV and should do the job. They were just the first diodes that showed after a quick web search. By the way, many of the extra high voltage diodes are, in fact, lower voltage diodes connected in series in an encasing tube.

The main problem, from a cost point of view, are the two reservoir capacitors.

Hmm, interesting that you designed and built EHT supplies. I have done quite a few too but using Royer inverters rather than 50Hz/60Hz mains frequency.

The highest voltage PSU I built, at that equivalent frequency (50Hz), was a -1KV plus 6KV supply for a valve (tube) scope that I tried to make in the early days. The EHT supply was a frightening thing with a massive, re-purposed transformer and similar sized reservoir capacitors.

In the mean time, I have been looking at a 6oHz EHT supply from a component point of view- I may post the results if it turns out OK.

spec

Last edited:

#### spec

##### Well-Known Member
Hi Keep,

The diodes are usually stacked in the systems I worked on. It takes multiple diodes to get to the voltage you need. The diode drops were on the order of 10 V on the systems I worked on.
.
That is true and, like Ron and I said, many EHT diodes are, in fact, a series of diodes connected together.

spec: We generally don't switch the neutral in the US. I've only seen one system that did and, guess what, there was a High voltage thing in it. It wasn't the main breaker either.
Yes, you have mentioned that before- in the UK, high end/industrial gear has the live and neutral switched and, for this application, switching both live and neutral is probably a good move.

spec

Last edited:

#### spec

##### Well-Known Member
POST ISSUE 04 of 2016_12_27

Hi Lash,

Below is a design for a 40KV EHT generator that focuses on component availability/cost.

You could put three secondary windings on one MOT and three secondary windings on another MOT. You could then connect the two MOT primaries in series so that you could use the existing 120V RMS primaries from MOTs designed for the mains supply in your area.

Also, you could take a risk and run the 10KV 10nF capacitors at 10KV. In that case you would only need four secondary windings of 7.07KV RMS. This approach should be successful as most capacitors will stand at lest 10% over their rated voltage.

spec

COMPONENT COSTS
(1) Diodes (HMV12): 4 * 6 = 24 at $1.43 US each =$34.32US
(2) Capacitors: 10 * 6 = 60 at $0.57US =$28.50US
(3) Instead of using 10 * 10nF, 10KV capacitors per bank, at a higher cost (£45UK), you could use 2 * 47nF, 10KV capacitor per bank (see link below).

(1) HVM12 diode datasheet: http://www.rectron.com/data_sheets/hvm5-hvm16.pdf
(2) HVM12 diode supplier: http://www.ebay.com/itm/10pcs-HVM12-Microwave-Oven-High-Voltage-Diode-Rectifier-/261403720417?hash=item3cdce03ae1:g:8dMAAOSwyjBW47cD
(3) 10nF 10KV ceramic capacitor supplier: https://www.aliexpress.com/item/5PCS-Ceramic-capacitor-10KV-103M-10NF/32378972602.html?spm=2114.10010108.1000014.46.1NkkQW&scm=1007.13338.69414.000000000000000&pvid=b6b86805-99d3-46e1-976b-76cb9e123a39&tpp=1
(4) 47nF, 10KV ceramic capacitor supplier: http://www.ebay.co.uk/itm/High-voltage-capacitors-47nF-10Kv-X4-/252687037262?hash=item3ad552334e:g:abEAAOSwtnpXpkUI

Last edited:

#### Lash432

##### New Member
(3) Instead of using 10 * 10nF, 10KV capacitors per bank, at a higher cost (£45UK), you could use 2 * 47nF, 10KV capacitor per bank (see link below).
Hi, sorry for not responding, I have been out of town for a bit. Something I was recently wondering about the capacitors, was why at 60hz they need to be at 470nF. I tried looking this up, but I could not find anything.

#### spec

##### Well-Known Member
Hi, sorry for not responding, I have been out of town for a bit. Something I was recently wondering about the capacitors, was why at 60hz they need to be at 470nF. I tried looking this up, but I could not find anything.
The choice of reservoir capacitor depends on the ripple voltage that you can tolerate, according to the formula,
Q = C * V = I * T ...(F1)
Where,
Q = Charge stored by a capacitor in Coulombs (number of electrons)
C = Capacitor capacitance value in Farads
I = Current drain from capacitor in Amps
T = Time in seconds.

For example, you have a 60Hz power supply which has a full-wave rectifier (bridge) and a reservoir capacitor. You want to drain 1 Amp from the reservoir capacitor and you can tolerate a maximum ripple voltage of 2V peak to peak. And you would like to calculate the minimum value reservoir capacitor that you will need.

From formula 1 (F1) you get,
C = (I * T)/V ...(F2)
Where,
I = 1 Amp (the current drain from the reservoir capacitor)
T = 1/(60 * 2) = 0.0083 seconds ( for a full wave rectifier the reservoir capacitor gets charged back up to maximum voltage twice for very cycle of the supply sine wave)
V = 2 Volts (the allowable ripple voltage)

Thus,
C = (1 * 0.0083)/2 = 0.0045 Farads or 4,500 micro Farads (4,500uF)

There are other factors that affect the ripple voltage but, for a typical power supply, the above formulae will not be too far off.

spec

Last edited: