Introduction: High Voltage Power Supply for Nixie and Valve Tubes

Picture of High Voltage Power Supply for Nixie and Valve Tubes

Since in last years nixies displays attracted my attention so much that I bought many of them, I finally managed to design a good power supply and my nixie clock is on the way... [UPDATE: done! read instructable here)
With this circuit you can test a nixie display, or power a multiplexed nixie clock. The potentiometer lets you change the output voltage to fit different types of displays (you can obtain from 100V to 300V DC). My prototype has been built on a breadboard, but I also made a pcb from the schematic so you can etch your own board and solder the components.

WARNING: this circuit raises the voltage to deadly 300V so you must avoid to touch contacts while working, I'm not kidding, please BE CAREFUL!

Step 1: The Schematic

Picture of The Schematic

To understand the way a circuit can raise voltage from 9V to about 180V I had to make some research in the web, because I'm not so skilled in electronics to design my own circuit. Anyway I merged some information from the most exhaustive sources, and my circuit tries to be simple and efficient, but most of you readers can probably add some revision or advice.
The two best sources, which are must-see, have been (device assembling instructable here) and Threeneuron's Pile o'Poo, and both their circuits are a bit different from mine, since from Arduinix I've taken the NE555 application, and from Threeneuron's the use of a pull-down transistor to improve the efficiency and reduce the mosfet heat. You can also read ian's instructable to better understand working principle and components details. Another source was this document from

On Threeneuron's website you also can find informations about the calculations to find the right R36 anode resistance value to limit the current for each type of nixie display. Let's say that 15 Kohm will probably work for 90% of the displays.
Look the notes on the image for more informations about the circuit. 

I also report the explanation of circuit functioning as -max- wrote in his comment:
An inductor is used to create high voltage kickback. This configuration is often found in boost converters, called that because the boost voltage. It works because one can not change the current flowing through an inductor spontaneously, inductors resist change in current. This means when the MOSFET turns on and passes a current through your inductor, the current flowing though it will steadily rise, "charging" it. It will eventually reach saturation where the magnetic flux can no longer increase and the inductor acts like a short. This is sort of like stretching a rubber band or spring. There is a point where you can no longer pull it tighter (otherwise it will break). Once this point is reached (or often just before it), the transistor turns off, and the inductor is allowed to "snap" back. Remember, the current absolutly cannot change instantaneously. so what happens is when the current is removed, the magnetic field in the inductor created collapses and induces a potential across the inductor. This voltage will slam up to infinity until a current flows between the 2 terminals.
However, in the real world, there is stray capacitance within the inductor as well as outside it, especially if a capacitor is connected. This creates a tank circuit, so the output will actually "ring," it is the equivalent to a bell, tuning fork, spring on a string instrument like guitar, etc. This can be seen on an oscilloscope with the probes connected across the inductor. You will see the ringing and even the initial voltage spike, known as a "transient voltage spike." The diode simply rectifiers this spike and transient so the voltage is DC, although quite a dirty one. (The voltage is all over the place as the inductor oscillates) the capacitor smoothes this out, offering a clean DC high voltage to the nixie tube.

Step 2: The Components

Picture of The Components

Some detail regard the components.
L1 is a fixed inductor 100 uH 1A, threeneuron's lists some similar models of it, and they could be a little different in dimensions, so my pcb has space for long and short components.
R17 and R18 should be 1% accuracy metal film resistors, to achieve a better voltage stability.
D1 has to be a ultra-fast 400V diode, as BAV21, UF4004, UF4007, MUR140, or MUR160 (thanks threeneuron's again).

Step 3: The Pcb

Picture of The Pcb

The pcb has been obtained in DipTrace from my schematic. Maybe you can design a better arrangement for components, I didn't etched this board because I'll incorporate the supply circuit in my nixie clock board. There is space for a big heatsink, also if probably you don't need it. In alternative you can lay down the mosfet. On the left you can see the long shape of the inductance alternative. On the lower right you find the Anode and Cathode contacts, where connect the display.
Attached to this step there are schematics, board and the ready-to-print pdf for toner transfer method.

Step 4: The Breadboard Prototype

Picture of The Breadboard Prototype

Otherwise you obviously can insert components on a breadboard and test the circuit.
To build your breadboard prototype follow the schematic (not the board) and start adding the IC (the NE555 in this case), connect one pin at time, adding components when they're needed, then connect other components between them and to Vin and ground. Double check everything before connecting the circuit. Also check that resistors pins don't touch anything else.

WARNING (again): this circuit raises the voltage to deadly 300V so you must avoid to touch contacts while working, I'm not kidding, please BE CAREFUL!

I used a bigger pot to ease the voltage setting, and I left my circuit working with a nixie for some time (maybe half an hour) to test the heat dissipation, and neither mosfet or inductor were warm at all... so I didn't add the heatsink.

Step 5: The Glow in the Dark

Picture of The Glow in the Dark

Here you can see some different nixie displays, the red glow is very fascinating.
IN-1 are unpopular between hobbyists because they have a shorter life and they're not transparent on the side, but I like them, and once solved the life duration problem they will be great in a nixie clock. Ampoule displays as IN-16 or the very expensive IN-18 are nice, but I love more the top-view nixies as IN-4 and IN-12 (russian tubes had been produced further than american ones, so they're cheaper and easier to find). If you want a exhaustive description about many nixies see here.

Step 6: This Is Only the Beginning

Picture of This Is Only the Beginning

So this is my contribution to anyone wants to build his own nixie clock, or needs a high-voltage power source to light his plasma devices (as in this impressive instructable, or also this other one) or some valve tubes.
Please comment this instructable and help me to improve the design. I'm waiting for some components to finish my clock, so that you will be able to see it very soon, anyway I know that some improvement is certainly needed and V2 of the power supply can born now... 
[UPDATE: my clock is completed and published:]


EwaJ5 (author)2017-08-27

hi there,

could you please add gerber files if you have them?

CraigH5 made it! (author)2017-07-21

I had a few issues getting this to work correctly from overheating issues caused by me not getting the A55 the correct way around to only having 110v output. Changed to using a bigger potentiometer (a chunky 10k one instead of one of the slim 1k blue things) and I now have voltages upto 300+v although I have not had it that high for longer than a few seconds. I only need 170v for what I am doing. Many thanks Andrea for yet another 'Ible. I used a 2n3904 and a 2n2907 but changed these for A55 and BC547 to try and sort my 110v issue but I got the same results until the Potentiometer was changed so they would have probably worked fine.

andrea biffi (author)CraigH52017-07-21

that's interesting... thanks to you for sharing your experience!

ymralin (author)2017-02-21

i finally managed to make it work. max voltage i got from it is around 300V. works perfectly for my nixie in-1, that requires 170V.

it can power 2x IN-1 in parallel. i couldn't check more, i don't have enough cables with crocodiles (or however they're called in english) to connect them and today i'm too tired to make up anything else.

andrea biffi (author)ymralin2017-07-10

well done, I'm glad you made it!

PierreL29 made it! (author)2017-01-31

I MADE IT , im so happy ! max voltage is 296 Vdc and the voltage adjust is pretty accurate ! Thank you for this instructable !

andrea biffi (author)PierreL292017-07-10

I'm glad of that! :-)

EwaJ5 (author)2017-07-09

what current does the 9v power supply need?

EwaJ5 (author)2017-07-08

hey, it outputs only 65VDc... does anyone know why??

mr.fusion (author)2017-06-18

i made in pcb. i used MPSA92 replacement of mpsa55. and other values are right. but i cannot take 200vdc only 20v. an MOSFET getting extremly hot after 10 seconds. someone help me?

mr.fusion (author)mr.fusion2017-06-18

use mpsa55

ymralin (author)2017-02-03

can mpsa55 be replaced wieth bc327? some website listed it as possible replacement. i used it and it's not working and now i don't know if it's because of bad transistor or if i screwed up something else.

at first i fried that bc327, because i soldered it the wrong side (collector and emiter are on other sides as in mpsa55). when i replaced it, olny thing i can observe when i turn the power on is that mosfetis getting really hot in seconds.

ymralin (author)2017-01-24

what voltage should be capacitors c3 and c4?

SonnyAngell (author)2017-01-18

How many tubes could this circuit power? Im looking to build a simple circut, jsut 6 tubes, All they have to do is light up.. nothing else..

andrea biffi (author)SonnyAngell2017-01-18

I'm not sure... but it probably could work

Mertzi (author)2016-05-30

This works like a charm, awesome guide! Giving me everything between 96VDC and 288VDC with quite okay precision when tuning. If anyone could give me some pointers as to what to change to get it up to even higher voltages I'd be very happy. Have seen some thyratrons that require everything between 600-1200V, so that would be awesome. Also, whats the best way of making it deliver more amps? And is there a limit to the current draw as it is now, or will it just burn up if I overload it?

andrea biffi (author)Mertzi2016-08-23

hi Mertzi, well done, glad you like the project. I hope someone can give you all the answers :)

melon_lemon made it! (author)2016-08-23

Yes, this circut has nice precision
I've read some trouble reports below, so i'd like to share my own problems if this can help somebody
1) induction should be reeled on coil at least 10x8 like that:
This one is 14x13
All other inductions (including smd) are too weak and didn't work for me
2) if Mosfet is hot, check mpsa55 and check its connections at the pinboard
3) while assembling, put ohmmeter between vcc and gnd reel. If resistance becomes lower than several kiloohms, start checking components
4) i've replaced some components:
Mpsa55 with 2sb644 panasonic (if you use it, dont trust datasheet and check pins names by multimeter, it looks like there is mistake BCE instead of right ECB)
Irf740 with irf840

andrea biffi (author)melon_lemon2016-08-23

Thanks melon-lemon! glad you made it!

Mertzi made it! (author)2016-05-30

Here is what it looks like, meassuring at max voltage

tomsamps (author)2015-05-17

Great article - just joined so I can pursue this as I am attempting to do the other half of the title "Value Tubes". The goal is a audio preamp circuit that will drive its tubes at their recommended voltage, similar to the "starved" type preamps that are common using the 6N3 or 6J1 tube. I picked up a couple of the kits on eBay that are covered in Threeneuron's Pile o'Poo and want to look into your modifications. I however only need to drive 1 tube's filament (leaning towards the 6J1 as good sound and lower voltage), but I need an ultra-clean power source. Looking at some those preamp kits (that use a voltage doubler to get them to around 30v) I see about 5 filtering caps to do the cleanup work, so I need to get a similar filter at around 120-150v. Since I am brand new at circuit design I was wondering if there were suggestions on how to clean up the voltage for audio use.


Mertzi (author)tomsamps2016-05-28


Im looking to do the exact same thing, wanting to use this supply to make tube oscillators. Did you have any luck cleaning up the DC voltage for audio use? Anyone else have any tips on how to?


Shaan743 made it! (author)2016-04-09

I just made it and it works like a charm. But I used a higher resistor at the output, instead of 15kohm for R36, I used 133kohm and it works just right. I recommend before putting any nixie to it, check the voltage and the current with a multimeter and check if they are correct for your nixie.
I used a IN-16 and it works perfect. If nothing is connected I have a power use of 0.7 to 1.8W, depending on the Voltage setting, which I control over a big slider poti, making the voltage setting very easy to set to the right value.
I used all parts like in this instruction, exept the very fast diode, there I used 8ETU04, because the electronic shop of my trust doesn't sell any other. But this here works too.

What I have to say to the author of this instructable: Thank you for this awesome plan. Finally I can work on my own nixie-tubes projects.

andrea biffi (author)Shaan7432016-05-09

8ETU04 appears to be an ultra-fast 400V 8A rectifier so it's perfect, just a bit more expensive.
I'm glad you are now able to power your Nixie tubes, and good luck for your projects. don't forget to keep us updated!

MahmadO (author)2016-05-09

can this circuit produce 400volts

andrea biffi (author)MahmadO2016-05-09

yes, probably it can, but you have to make calculations and change some component...

edward.chamberlain (author)2015-12-08

Is there a way to use +5V instead of +9V as the input?

EmillG (author)2015-12-02

Is it possible to use this circuit as alternative to HVPS-H 1364?

edward.chamberlain (author)2015-10-28

My attempt is so close to being right... 9V in but 348V out... changing the pot doesn't do anything to the output and the MOSFET is extremely hot extremely quickly. Also an audible 'switching' sound can be heard, although very quiet. Any ideas whats wrong?

Ok i noticed the mistake that my BC547 pin 3 wasnt connected to ground... however connecting this now limits me back to 9V? any ideas?

SamM77 made it! (author)edward.chamberlain2015-11-20

I'm having the same problem! 9V in, 8.4V out. I've had to use some alternate components as they are not too easy to source in the UK.

For Q15 I tried both BC557 (45V) and BC558 (@rtoselli had success with this- rated -30V... I have a separate question here about what negative voltage ratings mean for a transistor?) I also wondered if the D1 I am using is fast enough (1N4937 instead of BYT01).

I've posted my bill of materials online if anyone wants to take a look: - very open to suggestions about components to swap out.

I'm also uk based, try I find they have a better range! And are usually cheaper. I'll have a proper look tomorrow at your photo and send you mine

I had a similar bug but I only remember that it's extremely difficult find the mistake..
Actually most of times is a solder point failed.
Now that you found the problem and you fixed it you should test all transistors and some other components, because some of them could have been damaged with the high voltage of before...
Keep going, your clock will be great!

ouch I was thinking to my nixie clock instructable. this is only the nixie power supply, you should reach the goal with not too much pain. check that all ground traces are connected, and check transistor or change components one by one.

Have tested all components and they all are functioning correctly. This is very confusing! Can you confirm that this is a suitable inductor:

I think the ringing and low voltage can be caused by the (100mh) inductors size/amp rating. I had to change my inductor to 5.4 AMP ( ), this is despite my current circuit drawing less than 350 mA. Once I started driving more than one Burroughs B-7971 (one of larger Nixie tube types) the ringing would occur (increase). Once the ringing occurs the MOSFET heats up and is destroyed in short order if nothing is done. I think the RF (Radio Frequency) pros would have some pointers to give on inductors that this article doesn't cover. I think your inductor frequency rating may also play a part, the model you've used is rated for VHF (Very High Frequency) which is above 30MHz, this circuit only switches at 15-200KHz so this could also play a part, on second thought I noticed some strange behaviour moving this circuit to various breadboards, made be the larger leads on the 5.4 AMP inductor is just making better contact. Lastly check out I think Ronald was one of the first to put this base 555 circuit on the web, from his explanation I simply changed the MOSFET in his circuit to a modern low RDS(on) version (Infineon SPP20N60C3 or STP10NK60Z) and the fast recovery Diode (STTH2R06), thus getting the efficiency of the Flyback with his basic (non-flyback) circuit. No I'm not using the exact circuit above as I've omitted C4,R15,D2,Q15,R25. I added a 50K (R25) resistor between the gate and Ground in Ronald's circuit as the newer Mosfets "float" on without it.

Hope this helps.

I didn't test that model but it should work... I don't know how to help you

Hi Andrea! I'll run some tests on the components now, I actually have a spare so might try substituting them in as well. ( little hesitant as don't want 2 sets of broken components!!)

Also this is currently on a breadboard, I redesigned your supply to include a 5v voltage regulator as well so that I can power 5V components.

PierceK (author)2015-02-28

I made this so I would have something to test nixies I bought on ebay. It works great, and it's instructive to drop the voltage with the trimmer, see the nixie still illuminated, turn it off and back on, see no illumination, move the voltage back up, see the nixie relight, etc, so I'm learning about startup current vs sustain current. It's helping me decide on a working voltage without shortening the life of my nixies. I had the exact same problem that Jerdy described, and a thorough check of my soldering showed two bad joints. Fixed 'em and was in business. I'm using a 9v transistor battery on mine.

andrea biffi (author)PierceK2015-10-29

great! :)

diafour made it! (author)2015-04-20

Made it with analogous components:

BC547 - S9014


IRF730 - IRF840

BYT01 - FR107 (1000V, 1A)

1N914 - 1N4148

I have connected 2 IN-1 and 2 IN-4 nixies and efficiency was about 65%. Input: 11.4V, 175mA, output: 168V, 7.7mA. Wallwart has label 9V 1000mA but output is 11.4V =)

One of my IN-1 has digit "2" partially lighted. Bigbigesauce i think it is "cathode poisoning" especially if tube has used condition.

andrea biffi (author)diafour2015-10-29

thanks this is extremely useful!

for cathode poisoning you can try to play with potentiometer and alternatively reach a very high voltage for half second the going back to low voltage. cathode poisoning will disappear, but I imagine the tube will be a bit stressed ;)

rtoselli made it! (author)2015-05-27

This HV supply with the ominous 555 timer ROCKS. Thanks for sharing, now i'll finally put my nixies to use!

rtoselli (author)rtoselli2015-05-27

Oh yeah, i made it with analogous components too:

1N914 - 1N4148

MPSA55 - Bc558

andrea biffi (author)rtoselli2015-10-29

thanks! I will try it!

andrea biffi (author)rtoselli2015-10-29

yeah I think this circuit works so great!!

Jerdy (author)2015-02-04

I made the circuit as shown in the schematic and applied a 9v 1A power input. However, I can only get an output voltage about the same as I put in. I used parts with all the correct parameters and (I believe) wired them up correctly as near as I can tell, given how many times I double checked. I notice the diodes, C3, and C4 that I used look different from what the author used, even though the values are the same. I'm trying to figure out where I went wrong, but I'm not familiar enough with boost converters to diagnose the symptoms. Any thoughts?

bigbigesauce (author)2015-01-28

Hey, I am using IN-14s, and they don't seem to be lighting up all the way. Notice how the "6" doesn't illuminate fully? Any ideas?

organicelectrics (author)2014-08-23

Cool! I was playing with a flash charger circuit that generates ~300V so when disposable cameras with flash disappear forever, glad I know the trusty 555 works.

Yes I also was looking for a good HV supply to replace disappearing flash circuits, so I managed to put this circuit together

About This Instructable




Bio: I'm an Italian freelance structural engineer, graphic designer and photographer. I'm also investigating electronics, robotics and science in general. I enjoy hacking and ... More »
More by andrea biffi:Lasercut Little Wood BoxesSafe Clasps for Bike Pannier Bagsrevived CD rack as vintage box
Add instructable to: