Designing a High Voltage Power Supply for Nixie Tube Projects
2014-05-04
I've posted a page describing the design of a HV PSU (High-Voltage Power Supply) that generates up to 220V from a 12V input. In addition to that, it also provides 2*Vout (so, up to 440V, for dekatrons), and two outputs for powering digital logic: 5V and 3.3V. The primary HV boost circuit reaches 88% efficiency when going from 12V to 185V at 55mA, with a 3% output ripple.
The version I'm posting online is not perfect, but works quite well in a number of my projects. I decided I'd rather publish it as it is now rather than keep it locked forever.
It is published as Open-Source Hardware, to be used however one likes. All source design files are provided. It's my way of paying back: I learned a lot from looking at other designs and by asking questions, so now it's time to give back.
Great design! Thank you for releasing the schematics and going into such great detail on your reasoning and design choices. I am curious to know, about how much does each board cost to populate?
I might just have to add a few of these boards to my next OSH Park order. :)
My BOM shows it at around $15, without the PCB, based on prices in qty 10 (you really don't want to buy single units of anything). And unfortunately, I just checked and I will need to update the design as the IRFH5025 MOSFET is becoming difficult to get — an EOL notice has been issued. Will need to find another one.
Ok, I got an answer from IRF. The part isn't EOL'd, it's just the "TR2" small-tape version. The standard tape and reel version — IRFH5025TRPBF is active and recommended for new designs. (phew!)
Hi i just signed up with itead to do my first pcb run ever because of your excellent design thanks for the great presentation of your work. One question in the zip you submit for the pcb to be manufactured the required files list one not the your zip:
Outline layer: pcbname.GKO
Do i need this before they will process my order?
Many thanks
GKO seems to be the keepout layer. Their Eagle CAM Job never generated one for me (I looked through my past orders with them). I usually submit a zip file containing all the gerber files + the .brd file, just in case they need to fix something. Had very good results so far.
Thanks for the kind words — the design isn't perfect, as I said, but I'm glad people find it useful.
Well the process was easy in the end. Itead studio made 10 boards and shipped to my home in the UK since we chatted last week. Overall the quality looks good and the experience has given me confidence to create my own boards :)
Many thanks for the design i will get around to making it in a few weeks.
In the meantime If anyone from the UK wants a board i can let you have one for the cost of postage ~£1
Cheers!
Good luck with the build!
For last two weeks I'm searching the Internet for the info about these power supplies and your article is very helpful! The graphs are just amazing and very easy to understand (second year in Institute of technology does not count :) ). Thank you very much!
From the info that I could find I understand that if I want to get (if it passable at all) some where near 400 volts and 100 mA under the load (without doubling circuit to avoid more current ripple):
* The Inductor has to be 47uf (or less) that could handle 5A of current...
Mosfet with a small Rds-on... Like ~50mΩ, 450v at min.
Low Rsense - less than 30mΩ
Is that correct?
And what about other qualities of the mosfet, like Qg and Coss? How low they should be?
If the filter cap (C1-C2) will be polypropylene type (like Solen 400v-600v) with a larger capacity, will it improve the performance, if the size of the board is not the issue?
If SMPS 12 volt power supply is "all right" for these kind of projects? Because it's very hard to find wall warts with 3A-5A or more...
Well, not to discourage you — but 400V at 100mA is 40W. That is a lot of power! Think about it this way: if you manage to get it 80% efficient (I'd consider that a good result), you are dissipating 8W. You better have a good heatsink!
Boosting voltage from 12V to 200V in a single stage is already at the limit of what is reasonable, going from 12V to 400V — I'm not sure if that's even doable.
You'll run into lots of problems — first, you will surely enter CCM, so you will need to deal with the sub-harmonic instability (I worked around this by staying in DCM). Then, your inductor current will be huge, it's likely you will need to pick a larger inductance. As for the MOSFET, in general, the larger your Vds, the larger RdsON and Qgs are. The IRFH5025 is quite exceptional in that it has a relatively low Qgs (39nC). Qgs is very important: anything larger and you will likely need a driver for the MOSFET gate. Also, I'm not sure if there is a MOSFET that has a Vds of >450V and RdsON of 50mΩ. And if there is, your Qgs will be large.
For a boost of this size and power I'd rather go with a flyback topology, or something entirely different (transformer-based in any case)— but certainly not just a single boost stage. And I think that's a very complex project — fun, but I wouldn't expect it to be done quickly or on the first try.
As for sourcing 12V, I use laptop SMPS power bricks, many can supply up to 4A without any problems.
What you might consider is starting with a larger input voltage.
What do you need this for, anyway? 400V@100mA, with a small ripple?
Thank you for the reply!
There are boost converters with NE555 and they give 310v and ~25mA for two Preamp tubes (12ax7). Same design for the Subminiature tubes for the power amp - 2 small tubes delivered ~2watts of power.
I saw on some forum that one person had reached a 800v volts from a car battery with a similar design, using only one side of the transformer (Big Inductor basically). In the end, the inductor can be made by hand. Small number of turns with a thick wire or there are 47uh 6A toroid inductors from Ebay.
On the Vishay site, in the power mosfet section, there are mosfets that are 600v rated with RdsOn less than a 50mΩ, but high Qg (250-400). So the gate can be driven by a transistor I think.
Also on Ebay there are SMPS's - 12v 8A for less than 20$ with shipping. Xbox power supply is 12volts 14A :)
400V0.1A=40watts, 12v8A=96watts. Even 70% of efficiency will be enough to supply that votlage for B+, for the heaters of the tubes, and 3 preamp tubes... (40W - pentode, 11W-3 preamp tubes, 16W-Heaters (Filaments)). 96W*0.7=67.2W.
Sorry for the long post.
MOSfets in a Class A single ended power amps deliver 30watts of output with a passive cooling (heatsink). So I think 8watts is not a problem at all and if small heatsink will not suffice, a small fan will be a good idea.
Great work mate, love the time and effort that went in to making such a beast. Also spending the time afterwards to characterize it, looks like a lot of fun!
Have you considered sending a board to Dave Jones? I am sure he would love to do a mailbag segment on this, and he has some fantastic test gear, he will be able to test a few things that you might not have been able to. Have a look at EEVBlog if you haven't already.
I'll have to make a few of these, thanks for the details. You are a legend!
Matt.
Thanks for the kind words. I think I'm too afraid to send a board to Dave — I can already imagine how he'd rip the design apart :-)
I do listen to the Amp Hour podcast and like it a lot (not enough time to watch the EEVBlog videos, though).
I am going to build a few copies of your design. I have ordered the board (3 copies) from OshPark here in the U.S. (I love their purple boards).
I am in the U.S. looking for the IRFH5025TRPBF MOSFET (3 of them), but for every site I have tried, the minimum is now 4000 (ugh!). Mouser, Digikey, Allied, Element14, etc., all of them. I see some on eBay, but those scare me since they are so cheap in small quantities and could be counterfeit. What source are you using currently?
Thanks for sharing your great design!
I'm in Europe and I've been buying those MOSFETs from RS Components — you can get them in multiples of 2. I used to buy them from Farnell UK, but they now list them as "not produced anymore", which isn't true.
Unfortunately, this part is difficult to replace with a substitution. There are very few MOSFETs which can take 250V of Vds, much less so with a still reasonable gate charge of 39nC and a reasonably low RdsON.
I tried the RS Components web site, and their U.S. counterpart is Allied Electronics which gives me the 4000 part minimum and zero in stock. So I'll keep looking.
I decided to try a vendor at AliExpress with 100% feedback since they had a lot of 5 for a reasonable price (not the cheapest, but the best feedback-score price). We'll see what happens when they arrive in a month.
Another question: C16 on your BOM list is a "100nF 250V X2" with a funny-looking part on the schematic. The board layout shows a single capacitor for C16 (2 sets of holes for different size capacitors). What does the "X2" mean? I am somewhat confused by the label.
I hope I'm not I'm asking too many questions, as I'm sure I'll have another question or two for you. 8o)
That's a single capacitor, I just wanted to provide a layout that will accomodate two popular sizes. See https://www.google.com/search?q=x2+capacitor&safe=off&tbm=isch&tbo=u&source=univ&sa=X&ei=zncxVOz5DKXa7gaa7IHQDw&ved=0CEkQsAQ&biw=1401&bih=1080 for how this part often looks.
'X2' is a designator of the safety class. We don't actually need the X2 class here (a SMD ceramic would do just fine) I just found that these MKT caps are cheap, commonly available and easy to find, so I used one.
You can actually build without this cap, if you're willing to accept more output ripple. It just provides additional filtering.
Thank you for the quick responses today. I see in my search at Mouser.com that "safety" capacitors are turning up from my "100nF 250V X2" input.
Thanks for the clarification! You learn something every day....
Jan,
Thanks much for you post. This is going to be my first delve into DIY electronics. Are the parts easily soldered to the PCB or do you recommend a different way? Apologies for the newbie question!