Need more voltage than your supply can deliver? This is the device to build.

Based upon the Villard doubler, cascaded in multiple stages, this circuit will multiply a small AC voltage into a huge potential.

With a 6000 VAC Oil Burner Transformer and just three stages, you can multiply the potential from 6000 to 51,000 VDC!

Also, the higher the voltage ratings on your components, the higher the input voltage can be and therefore, more bang for your buck.

One side benefit of a stack like this is you can tap off various voltages from stage to stage. In this example, three outputs of 17000, 34000, and 51000 volts are available. Nice!

Step 1: The Theory

Each stage of this multiplier is a doubler circuit and is made up of (2) Capacitors and (2) Diodes.

On the positive half-cycle of the input, the capacitors charge in parallel to the peak value of the voltage presented through the forward biased diode. The other diode is reversed biased. Each cap charges to 8484 peak volts.

On the negative half-cycle of the input, the forward biased diode becomes reversed biased while the reversed biased diode becomes forward biased. This effectively connects the capacitors in series allowing them to discharge into the load at the output. The result is a doubled voltage presented to the load or 16968 peak volts.

Each connected stage adds its potential to the total output.

To calculate the expected voltage at the output with a given input and number of stages, plug the numbers into this formula:

Eout = (2 x Ein) x S x 1.414

Eout is the Output Voltage, Ein is the Input Voltage, and S is the number of stages in your design. I used a 6,000 VAC Oil Burner Transformer for my input and built 3 stages.

Eout = (2 x 6000) x 3 x 1.414

Eout = 12000 x 3 x 1.414

Eout = 36000 x 1.414

Eout = 50,904 volts

Step 2: The Parts List

All of the critical parts are available on EBay. Here's what you will need:

(6) 470 picofarad, 20KV Ceramic Doorknob Capacitors.

(6) HV03-12 12KV PIV High Voltage Diodes.

(7) Brass or copper wiring posts (homemade).

(8) 8-32 x 1 in. brass screws.

A length of 50KV High Voltage wire for the output.

A length of 20KV High Voltage wire for the input.

A length of 12ga. Stranded wire for ground.

(3) #8 Wire lugs.

(1) Ziploc Food Storage Container big enough to hold the multiplier(not a bag).

(1) qt. Mineral Oil.

RTV Silicon Rubber Sealant.

NOTE: Only the voltage rating on the components is critical.
The higher the rating, the higher the input voltage can be, resulting in a higher output voltage. The input limit on this design is 10KV (with safety margin built in). Ignition wire or coax cable with the shield removed may be substituted for high voltage wiring if need be.

Step 3: Build It!

This step isn't rocket science, just common sense building techniques. Just copy the design shown with the binding studs added.

Start by cutting (7) pieces of 16 guage brass or copper wire two inches long. Bend each piece around a drill bit and shape with a loop at one end.

Next, cut the heads off of four of the screws to make the studs that join the capacitors together.

Screw a stud into one of the caps. Place a binding post over the stud and screw on another capacitor. Screw in another stud, place a binding post over it and screw on the last capacitor.

Repeat the previous step to create two stacks of three caps each.

Lay the two stacks side by side with the binding posts pointing up.

Cut the four remaining screws to 1/2 inch long.

Attach binding posts to the right side end of each stack using a screw.

Attach a binding post to the left end of the bottom stack using a screw.

Attach the diodes as shown and solder them to the binding posts.

Cut off any excess binding post tops and file them flush to the soldered diodes. No sharp edge here or you will get an unwelcomed breakover which could ruin your project.

Attach the wires with lugs. Tighten the screws and make sure the entire assembly is tight.

Step 4: Bury It!

Now that the multilpier is finished, we need to encase it in oil.

With the lid still on the Ziploc container, punch three holes in the lid at the approximate locations where the leads will exit it. Remove the lid from the container and set it aside.

Slowly fill the empty, clean Ziploc container half way with mineral oil without creating any bubbles in the oil. Oil good! Air bad!

Hold the multiplier by its leads and slowly lower it into the oil bath.

Continue holding the leads with one hand and slowly pour more oil over the multiplier until it has at least a 1/2 inch covering and about 1/2 inch head space between the oil and the lid.

Now pass the leads one at a time through the holes in the lid then work the lid down onto the container while holding the leads. Do Not get oil on the leads at the lid level or the sealant won't stick or seal the wires. Snap the lid onto the container.

Finally, seal the leads and if you like, the lid to the container with the RTV sealant and set the project aside for 24 hours until the RTV is cured.

Your Done!


Step 5: Notes, Warnings, Safety Procedures and Disclaimer


This multiplier is part of a 0 to 50KV adjustable High Voltage Power Supply capable of sourcing 20 milliamps of current.

An instructable on building the supply is my next project, so hold on!


  • * * W A R N I N G * * *

This device produces LETHAL CURRENTS at HIGH VOLTAGE. The output of this device WILL KILL YOU if you do not follow standard common sense safety procedures.

Safety Precautions and Procedures

1. Wear safety goggles or glasses, rubber safety gloves, and stand on a rubber safety mat when powering or using this device.

2. GROUND the output of this device after powering it off. It can retain a lethal voltage for several minutes after powering down.

3. DO NOT TOUCH the device AT ANY TIME during operation.

4. Do not allow others who are not familiar with high voltage devices to touch or use this device without proper supervision.

5. Do Not operate this device alone! Have an emergency person available when performing your experiments who is familiar with proper rescue procedures.


By building and operating this device, you acknowledge that you understand the dangers improper operation can pose and you accept all risks associated with the operation of this device.

You also acknowledge that I am not nor will be responsible for any death or dismemberment by this device whatsoever and that you assume all risk by the use of this device.

<p>Here's my little guy, seems like this is the only reasonably priced type of cap you can get like this, unless anyone knows where to get the doorknobs for less than like $70 a piece? <br><br>Anyway, the first build was a lot tighter than this one and testing it burned out the OBT's primary and killed one of the diodes, so I got new diodes and remade it with more room and tested it with a microwave oven transformer which, obviously not as much voltage, but enough to test and it works great! Now to replace my OBT....</p>
<p>Try ebay :) like 25 dollars for 8</p>
<p>if i want to make voltage multiplier 20kv with input 220V 50hz, what the best components to mak it?</p>
<p>You won't be able to make a 100x voltage multiplier like this. The high voltage diodes consume a considerable amount of the voltage. Further the cost of just the diodes would be more than $/&euro; 2000. <br><br>If you want 20kV from 220V mains, just use a transformer. Its FAR cheaper, and many times more efficient. Plus playing with 20kV, you need some serious insulation layers between the capacitors and diodes to prevent unwanted arcing.</p><p>2 12kV neon sign transformers would run you around $/&euro; 100-150.</p><p>Then rectifying that would be about another $/&euro; 100-150.</p><p>So using a multiplier would be more failure prone, much more dangerous, and cost easily 20 times a dual transformer setup. </p><p>If you want to really don't want to go the transformer route, don't use this circuit. Use a Marx generator. It doesn't suffer from the same diminishing returns as a multiplier. Though it has other drawbacks as well.</p><p>Be careful though. 20mA at 20kV will likely instantly kill you, and a Marx generator can easily exceed 100 to 200 times that current. One touch, anywhere on your body, and you're very dead.</p>
<p>do you ever made voltage multiplier, do you have id line? my id : hikmatsb</p><p>or my email hikmatsaefulbahri@gmail.com please answer</p>
<p>do you ever made voltage multiplier, do you have id line? my id : hikmatsb</p><p>or my email hikmatsaefulbahri@gmail.com please answer</p>
<p>thanks a lot for your information</p>
<p>...and you don't experience any arcs between those unprotected solder joints?</p>
Very nice article.
<p>Will the multiplier work if I directly connect it the AC mains power supply?</p>
<p>hi, can I use different output capacitors in same multiplier? how to built a multiplier with different output capacitors? </p>
<p>Encased in 3&quot; PVC and filled with Paraffin wax. Caps are 20kv 470pf and diodes are 15kv 100ma. That's why I used 3 per connection. As far as the &quot;ground&quot; terminal I am assuming when using a 2 pole transformer (type 619 10kv OBT) I just connect 1 lead to HV in and the other to ground?</p>
<p>I just finished building one to your spec and was wondering how to connect my 10kv OBT to it. There are 2 secondary terminals. </p>
Alright Ive thought on this for some time and I absolutely HAVE to ask... partly because some Im confused about and partly because no one else has. <br> <br>First, my OBT is similar to the one used in this instructable, RMS 6kV. Peak to peak its 20kV, half wave obviously 10kV. Now burningsuntech, you said the RMS value is .707 of the peak value, and/or the peak value is 30% higher than the RMS, but my numbers dont equate, according to what the label on the OBT says my RMS should be 7070v, and yet it says its RMS is 6kV. So whats up with that? <br> <br>The other part of my question is in the formula you give, Output voltage, Input voltage, and the number of stages are all represented as well as the action of each stage (that being Ein x 2), i.e. Eout = (2 x Ein) x (# of stages), but where are you getting the extra 41.4% output at the end there, represented by x 1.414? Its not mentioned, its just there? <br> <br>I also wanted to say that even though Im confused on those minor details (not minor to me cause I wanna know :P) I have constructed the multiplier, connected the variac, which is an external one that I can keep outside the case and plug the supply into or use for other things as well, and connected the OBT to the multiplier using 20kV If its labelled TV-20/40/50 thats the kV rating) HV wire I got from flybacks, as well as a short length of 50kV (TV-50) for the output, and while I still have to fill the project box for the multiplier with wax (it DOES arc!!!! Blew a diode that way... just one though!) it works PERFECTLY! Its a very elegant project, and I am very proud of it. Thank you for your instructable and lessons on the theory of operation, its great!!!
Oh I forgot, I meant to ask if the x 1.414 was because of ripple or is it something else entirely?
Nevermind, I got it, the caps charging to the peak forward 8484v, which is 1.414 times the input of 6000v
<p>It would have been really easy for me if someone would have just said the 1.414 is the square root of 2, part of the Root Mean Square lol</p>
Can you guys that are also making this multiplier tell me what diodes you use and why? I had some that worked, but I only had a couple extras and had some arcing that fried one a couple times so I have to get more and revise my construction and I'd like to hear what you other guys are using for this if not HV03-12 (hard to find and expensive, as far as I can see)
Hi, if for example i want to use this multiplier to multiple the output voltage from high frequency ( say, in the kHz range) alternating current (say from ignition coil), would it be reliable to use diodes 1N4007? Or in simpler words, is 1N4007 suitable for hi freq AC?
Hey Ymasamune <br>I think your answer is no because the 1n4007 does not have a fast enough switching speed. At high enough frequency like in the kHz range the diode would conduct backwards. Here a youtube link i saw this at. It starts at 3:02 <br>http://www.youtube.com/watch?v=cyhzpFqXwdA&amp;feature=endscreen&amp;NR=1
Couldn't he just get fast switching diodes though? I think the ones Im using aren't 1n4007, they're rated for enough voltage but the amperage is only 350mA, looking at the datasheet the switching speed is in the nanoseconds
The 1N4007 is rated at 1000 volts P.I.V., 1 amp. UF4007 is rated the same except the recovery is much faster than 1N4007. UF4007 recovery speed is 75 nS, the 1N4007 is 35 uS. The UF4007 is best on high speed switching circuits. See datasheets for UF4001 to UF4007.
Ah yeah I see
Hi, nice tutorial, I have a DST half wave modern type Flyback from CRt like this: <br>http://www.electronicrepairguide.com/images/flyback1.jpg <br> <br> Will it work for this multiplier? I have always wondered this but I coudnt find any answer <br> <br>Thanks .
will this also work with capacitors which aren't made out of ceramic?? i can't find 470pF ceramic ones, vut i found iothers for 470pF
The capacitance doesn't have to be spot on, I saw a lot of caps that were like 580pF 20kV that would work just fine, it doesnt have to be exact but dont stray too much from what the design calls for. Ill try to find you a link where I saw them, but it seems much easier to get the correct style doorknob caps at a slightly higher capacitance @ 20kV, I ended up finding some russian 20kV 470 pF caps on eBay, the guy sells them often so look there. Theyre more cylindrical than what suntech uses, and you'll have to get creative with the assembly because they don't fit together real nice like his do, not to mention the metric thread size, but its doable, and theyre a hell of a lot cheaper than the actual &quot;doorknob&quot; style used in the pics.
Here is a link to a lot of four that I purchased from the seller I purchased them from. They came all the way from Ukraine, but the specs are exactly right and I didnt have a lot of trouble with assembly, I used thick solid copper wire like burningsuntech does, and threaded the end that isnt threaded with the metric size needed, and got the right nuts to fit so I could loop the wire from cap to cap, and had copper posts on the ends, and just wrapped/soldered the diodes to the copper. You may find a way to do it that is better for you, and to be honest, it wasn't worth the trouble of threading and finding the correct size die and nuts for it, so stay away from that, but here you go: <br> <br>http://www.ebay.com/itm/4x-470pF-20kV-High-Voltage-Doorknob-Capacitors-K15-4-/200889917153?pt=Vintage_Electronics_R2&amp;hash=item2ec5f8dee1 <br> <br>And check out places like these: http://vintageaudioandvideo.com/high-voltage-ceramic.html <br> <br>Just run a google search, eventually you'll find some you can use. DO NOT buy them if they say 20kVAR!! Thats not the same thing! 20kV or as high as you want, and Id say anywhere from 330pF to 880pF should do alright, but theres far too many 470pF and 500/550/570pF to have to stray that far. Good luck with the hunt! They are rare and can be expensive!
hey i'd like to make a device i can attach to say my foot ,that if i touch someone they would get a shock basically like when you walk with socks over a carpet but the effect is automated. A guy made an ible https://www.instructables.com/id/The-mini-electrostatic-generator/ where he makes what im describing but its not very detailed. any ideas?
have you had anyone use this voltage multiplier and or your variable voltage supply in corona poling?
Im wondering if someone here can help me out....<br><br>I have a transformer, an OBT, with an RMS of 6000v, and youd think, hey! This is exactly what the design calls for! But Lo and behold, Im looking at it in my hands, and it says on the label: &quot;PRI: 120V, 60Hz, 35VA SEC: 20Kvpk, 35mA, Midpt Grnd&quot;<br><br>If the RMS is 6kV, how can the peak be 20kV? And can I still use this with 12kV rated diodes? Or should I get different ones rated for like, 25kV?<br><br>For the record, the OBT is a Beckett 51771U, A-type oil burner transformer, and everywhere Ive checked, its 6kV RMS
I think I figured it out:<br><br>I emailed the company, and they told me that they use half the sine wave for the ratings, but I found out that isn't completely true. <br><br>20kV peak is actually a peak-to-peak value of the transformer using the full sine wave, negative peak to positive peak. The RMS value is found using only peak value, which uses only half the sine wave. So that leaves me with 10kV peak. If you go by burningsuntech's math (which I am inclined to believe), the peak value will be 30% higher than the RMS value, or the RMS is 60% of the peak value. 60% of 10kV happens to be 6kV. Problem solved!
OK I see I have some math errors, they're not really, but how they are measuring the sine wave values and etc is more than I wish to go through to correct myself, but either way, if anyone has this same question/issue, its OK to use your transformer so long as the RMS is 6k!
i have 1 question. i only have 100pF capacitors with rating ranging from 2kV-5kV. i intend to use it to make an output of about 20kV. i'll be using an output of about 200-600V. how many stages can i make out of the it. because rating of the capacitor seems to be quite low
Hi Arjoon,<br> I'm not an electrical engineer or anything, but I've had a bit of experience with high voltage as of late.<br> <br> I think that you meant to say that you will be using an INPUT of about 200-600V, and if so, your 2kV-5kV capacitors should work. I assume that your INPUT power is in AC, because if its not, you're out of luck.<br> <br> I believe that your capacitors will work because their voltage rating only needs to exceed the peak voltage of your INPUT (200-600V r.m.s. = 282-848V peak).<br> <br> Unfortunately, for your very small input voltage, you will need A LOT of stages to get all the way to 20kV. If your input is 200V, you'll need about <strong>35 stages!</strong> If your input is 600V, you will need 11 or 12 stages.
how can I double a 6V 4.5Ah battery into 12v battery with 2A current or more.
Success! Well, sort of . . .<br><br>I finished my multiplier, and hooked it up to a 7,500 volt neon sign transformer. I used the output to power an ion &quot;thruster,&quot; but the results were disappointing. Using the HV outout of an old CRT monitor, the thruster moves vigorously. Using the multiplier/NST combo, I got a little bit of thrust, hissing, and a stream of ions, but not enough to make the thrusters turn. Since the output of the CRT is around 25,000-27,000 volts, I suspect I am not getting as much from the multiplier. According to calculations, I should be getting 60,000 volts, but I don't think I am. Since I am using recycled diodes, maybe some of them are defective? I checked the doorknob caps and found them within spec. I did not check the diodes.<br><br>Paul
will a 10kv 23ma transformer work?
Good Question, Q. The voltages on all AC devices are usually shown as the R.M.S. (root mean square) of the actual Peak value of that voltage. Your 10,000 volts R.M.S. would be 14,100 volts peak-to-peak or just 12,293 volts peak. The average voltage (R.M.S.) is .707 of the peak voltage. We know that 10,000 volts is .707 of the peak or 70% of the peak voltage. That means that the peak is 30% more than the average or in this case about 13,000 volts. I need to know the peak figure so I can determine if the P.I.V (peak inverse voltage) of the diodes can withstand that high a voltage and in this case... They cannot!. The diodes have a PIV of 12,000 volts. If you put a peak voltage of 13,000 volts across these diodes, all you will get is lots of smoke and not much more. What I telling you is that the peak voltage of your transformer is higher that the PIV of the diodes by 1000 volts and you will destroy them. If you get diodes with a PIV of say 15,000 volts, then you will be safe using this transformer. The capacitors in this multiplier have a much higher rating and you dont need to worry about them until you reach 20,000 volts on the input. The diodes, however, have to be changed. RA
Pardon me for breaking in here . . . if I follow you, then I could use a 7,500 volt NST with a multiplier built with 20kV doorknob caps and microwave oven (HVR-1X3) diodes, which are rated at 12kV? <br><br>The (max) result would be output of 63,630 V DC?<br><br>thanks,<br><br>Paul
Correct. <br>Give the man a kupie doll. <br>One change I would maKE in the build is to use parafin instead of oil in the multiplier case. Its easier to use and it is easier to seal the container. I left mine open with no problems and was able to get it to maximum voltage <br> <br>You might consider using a higher amperage Variac as well. Mine was too close to tolerance with the transformer running at 2.25 amps and my variac was for 2.5 amps. A bit too tight. <br> <br>Have fun. <br>RA
Thanks for your reply. I put the multiplier together today. Haven't tried it yet, but I have just a couple of simple questions. <br><br>The input Ground goes to a real ground, like a pipe in the back yard? <br><br>The low voltage (relatively speaking) input is going to come from an NST--the NST has two output lines; which one do I use, and what do I do with the NST line I don't use? Just isolate it?<br><br>The output is + DC, correct? <br><br>Thanks,<br><br>Paul
So you're saying all I have to do to get my transformer working with your design is buy 15,000 volt diodes instead of 12,000 volt diodes?
Q. Yep. Thats about it. Something else to consider is how you are going to &quot;pot&quot; or cover the assembly to remove air and prevent breakdown from occurring. i have tried potting in epoxy, oil, and wax (parafin). Epoxy does not allow you to change anthing or to harvest components if you should decide to change the number of stages. Oil is hard to control and to prevent leakage you will need to pay attention to the seal on the container you put it in. That leave Parafin. This is what I am currently using and it seems to be the best of all worlds. Enjoy your build. Good Luck opening that stargate worm hole. RA. :)
You could try starting with voltage from a Old tube TV or Monitor
JJ.inc <br>The output of the high voltage section of a tv or monitor is too high. Typically, they output from 25,000 to 50,000 volts in the microamp range. The current won't be a problem. But the design of my HVM is for low inputs up to 10,000 volts. The project would fail on the first try. <br> <br>To overcome the problem, you would need to change the values of the caps and diodes for it to withstand the much higher voltage of a tv or monitor output. <br> <br>Raving Apache
Yea I figured as much thanks
burningsuntech,<br>what's the best way to discharge the capacitors after use?
bs1500 <br>Short them to ground. <br>RA
burningsuntech,<br>I have a 9030v, 30 ma neon sign transformer. I see the 20kv capacitors should be enough to handle the voltage, but the diodes may not be enough, particularly since the HV03-12 reverse voltage is 8400, possibly leading to a lot of smoke. I found some 15000v, .30 forward current, 100ns recovery time rectifiers (part number Z150UFG) at http://www.voltagemultipliers.com/html/selection_guide_hv_diodes.html. How do these look to build this for my transformer?<br>Thanks!<br>bs1500
bs1500 <br>Looks like they will work fine. <br>Good luck <br>RA

About This Instructable




More by burningsuntech:Maxwell House Wireless Antenna Make your own Eco-friendly soldering flux Build a Variable High Voltage Power Supply 
Add instructable to: