Powerful SMPS for tube ampifier

Hello, I designed a SMPS that creates 0 audiable noise in a tube ampifier. Originally this was a school project for a competition but things did not work out quite the way.

The powersupply steps up 10-30V input to whatever output voltage you desire up to the volate rating of the filter capacitor (1kV best estimate of the highest voltage). I have tested the powersupply on a resistive load and I managed to test at 450V 50W without any issues. The tube amplifier I tested the supply on is a well known 20W Mullard amplifier. The amp consists of a pair of EL34s a 12AX7 inverter and driver and a EF86 as the preamplifier. The onl change I did was to use a EF0 instead of the 86, and even it being a very microphonic tube does not create any issues whatsoever and it sound amazing. I designed the amplifier on a PCB and the powersupply too.

Right now the powersupply its been about the third prototipe, has a small issue so without a fan blowind lightly on the switching transformer an the heatsink of the IGBT, gets unbareably hot and the input filter cap too so I guess it acts as some sort of weird oscilator - I havent looked into it any deeper yet I was more interested to hear the amp.

What I would like to know, if there is anyone even lightly interested to see a switching powersupply for a tube amplifier. The current version creates the 410V for the amplifier from anywhere between 12 to 30V. It is also possible to use a different switching tranformer for rectified line votages- why would I use a swtiching powersupply istead of a regular simple transformer on the line voltage? Lets say the voltage if tightly regulated and stable-and this powersuppply is MUCH cheaper than the transformer itself.

Little about myself- Im a students I turned 17 last week and I was protiping this SMPS for about 6 months before i I got to a point of any real useable result. I made lots of amplifiers mostly transistor based or IC based AB amplifiers and Im also into analog technology. So basicaly a nerd :D

I apologise for the grammar im not a native speaking.

If anyone is interested in any measurements I can do them I do have a digital scope and other testgear to do almost any mesurements I can think of except EMI and interference measurements
 
Founder of XSA-Labs
Joined 2012
Paid Member
I simply created a Fly back SMPS and I basically made the thing bombproof. Uses a transformer of a capacitor charger of the LT7351 BUT I don't use the chip itself to drive it tho. The operating voltage is about up to 1kV if you can get such a filtercap. I can pretty much just use a 10uF 450V cap right after a diode and done. But I noticed even tho I don't hear the hiss at all or see it with a scope on the output, turns out the preamp pentode grabs that 56KHz noise and amplifies it loading the phase inverter . This does not create any issues in the amplifier however I did use a RLC filter just to make sure. Series resistor 10Ohms 470uH and a 100uF 450V cap across the output and a ceramic SMD 630V 100nF capacitor. Makes it nearly dead silent. 1V P-P noise is visible on the scope at the 450V scale. Your plan to use a 47K resistor is only good for preamp stages . It delivers no power. While I provide high power at 400V.

Hello CaterpillarSK,
I am interested in switching power supply for tube amp.
it is even better if it is stable and have adjustable voltage output .
Are you planning a kit for sale ?
No, at least not yet. The prototype version I don't even know which number has still some issues. And its also nowhere near safe to be operated for longer than 10 minutes. Not yet at least until I find out why because it seems to be just a matter of looking at it with a oscilloscope and it seems to be a simple fix
 
UPDATE: so by just about gently blowing on the switching transformer it keeps cool at 40C and I can play allday on the amp. For the ones unfamiliar the Mullard 20W is a 20W RMS amplifier. I have not done any actual measurements of what power goes in the amplifier itself, but I can tell you its a lot more than any chinese high voltage converter can do any day. Its really loud I did not expect it to be this powerfull
 
This works pretty well. Add a CRCRC with 450v 2.2uF film caps and 47k 5W resistors to clean it up.

Qianson High Voltage DC-DC Boost Converter 8V-32V to ±45V-390V Adjustable ZVS Capacitor Charging Power Supply Module Amazon.com: Qianson High Voltage DC-DC Boost Converter 8V-32V to +-45V-390V Adjustable ZVS Capacitor Charging Power Supply Module: Home Audio & Theater

The one you listed is good for a preamp and available cheaper on eBay.

This one is a much better choice for a power amplifier. I've used one for 320V@500mA and it ran cool.
500W 12VDC transfer 18VAC & 0-220v-380v inverter module | eBay

I powered it from an ATX supply.

I would be interested in an SMPS like the OP has made except at the following specs.

90-265VAC input, 300-350VDC 1A output with a 12VDC output for heaters (at least 10A, and capable of starting into cold heaters). Any ideas?
 
Last edited:
I would be interested in an SMPS like the OP has made except at the following specs.

90-265VAC input, 300-350VDC 1A output with a 12VDC output for heaters (at least 10A, and capable of starting into cold heaters). Any ideas?

That is one hell of a constructive reply thank you!

The following can be solved- I can power the SMPS with 100-400VDC its all the matter of the switching transformer. for a 10-30V input I use the Coilcraft GA3459 transformer. For a 100-400V SMPS I would need the Coilcraft HA4060 and simply use the rest of the circuit as is, EXCEPT I did not have yet thought into galvanically isolating the HV side YET, with time everything can be solved and the 12V 10A source can be used as a auxilary powersuppply from ebay - those metal shielded supplies- one thing to keep in mind is that those power supplies some of them are audiably noisy on in the way of transferring noise into the amplifier itself but the switching transformer itself inside that powersupply is actually the loud part of the supply. But I tould technically do both on a single board its just nothing I thought yet of because im using a ATX supply to heat the tubes (2X EL34 in paralel and the 12AX7 and the EF80 have the same heater current so I paralled them to each EL-34).

If you would like to see such a powersuppply I could design one, but it takes time to even grab an idea and holdon to it, and also funding. I almost ran out of money building my tube amplifier with my powersupply for a competiotion that did not even work out the way I would like to (they did not even let me to compete- they said the online registration was incorrect thus not taking my work into consideration. But I bet that one old man saw me at the cross country competiton where I won 5th place (agan because of how poeple know eachother) and he probably did that on purpose because the school that went compeeting and it was the shool he was from totally failed to even gain a place.) So I gotta recover from this realitely big spending by doing all sorts of repairs and manufacturing and fixing other devices. I even do computers so wont take terribly long.

ONe detail I did not use a common chip to drive the reansistor I made the entire controller out of 4 op amps because the TL494 has simply a too tight of a voltage regulation and creates all weird noises of raising the PWM and stuff like that. But for a line powered supply it would porbably work best as of simplicity and practicallity. I also dont have the option to make custom trasnformers so Im limited to what I can dig out from real expesive and powerfull schematics that one guy gets me of extreme spec testear that is for sourcing massive currents and the use this kind of stuff. But its easy to find some components as I said the GA3456 transformer was found to be the cheapest and best option from the LT3751 datsheet and it is a high voltage regulator/ capacitor charger. Also the HA4060 is from the same datasheet- Id did try to use that chip but the chip was useless cause I couldnt really find a good footprint for Eagle and it also died during the soldering (I did not have a proper hot air station back then)
 
One small detail- the HA4060 transformer from coilcraft would be safe to use up until 1,5KV (the specs say its islation rating if 3KV). Also the datasheet specifies 500V out max but the isolation rating should be enough to get up to 1kV making it usefull in prettymuch anything from preamplifiers-alltho quite overkill, Powerfull amplifier up to id like to say 100W. The switching frequency is double the GA3459- 100kHz which gives more power crammed into it and nearly no noise that the tubes could amplify. Prettymuch clean DC with quite the practical and light powersupply. So basically halfing the total weight of the amplifier by not needin a power transformer.

The transformer is also shielded completely instead of only flux shielding the coils-I do recommend to use a coaxial cable within the tube amplifier because hissing noises is unacceptable and with a coax cable basically not possible to pick up any noise
 
Since some poeple at EEVBlog also wanted to see the outcome of a mains powered HVPSU, I basically went into it and bought the different switching transformer. I hope that this wont be the last project im making work because I simply dont really feel okay working with mains voltages but hey, its about time to overcome some fears and lets to stuff. So expect some updates in the future about the PSU.
 
I would recommend a whole lot of caution if going down a mans ac input path.

That is a whole new world of safety and regulations for you, and even more if you want to progress that way in any commercial sense, or don't have any experience in that field.

I can recommend you grow your experience with confirming your thermal outcome, right down to junction temp level, and appreciate how to compare the source impedance of your hvdc supply to a much larger electrolytic cap.
 
Yes, I know all that whats coming for me. I read a lot of literature about this stuff and the regulations, so I have the knowledge I just dont really have "balls" to goa and experiment. First of all I aint gona be seeling this in stores, I just want it functionall and not to disturb anything arround it, if lateron there will be any demand, I may be working on a certifyable version. But I doubt any demand would even exist for a thing like this. However I do want to finally dive deep into mains powered projects. Even tho I hate to even think about powering stuff with mains, I think I do have the knowledge and its time to begin. Safety first.
 
Yes I documented the entire design, however I found a lot of flaws and found out some interesting sideeffects. Some undesireable, some desireable.

The mains powered version will use a off the shelf chip to controll the switching, also there will be implemented opto isolation between the mains side and secondary side. Im aiming for maximum simplicity and best performance, while not surpassing component ratings.

The 10-24 version is really inefficient after the power of 20W. Getting down low as 70% and prettymuch stay the same up until 50W where the switching transformer is getting saturated to the extend of not being able to extract more power from it.

Another find was that I canot use a MOSFET with a bodydiode which is the main reason to go with an IGBT. The body diode basically snubs the reverse voltage that is pointlessly loading the transformer and diissipating the power into waste for no reason.

The low voltage version was built as a proto, giving me acess to modify anything on the go. Unlike the LV converter, the mains powered one will feature a PWM block- locking the PWM on low so the switching will stop and the converter will be simply in a standby mode. This can be used as also a "safety" switch- open chassy and such.

I dont know if that answered your question, if you have any specific questions, let me know.
 
I have had good luck using the RECOM Isolated DC/DC Converters Rxx-B series for tube pre-amp power supplies. The units are only 5 watts, so not enough for a tube power amplifier. Most off the shelf high power/high voltage SMPS are more expensive than HV transformers.

Are you going to post a schematic?
 
Probably without values. The schematic is useless without the PCB. It was quite difficult to find the legit parts for my cad program that I do PCBs in (Eagle). In fact I ended up creating my own. So at this point its really not even an option to build it yourself except that if you would recieve the PCB from me, and even then, soldering is not a, tack it and done job, its quite the procedure especially with the pulse transformer. If and when I will complete the PSU, and it will be within specs and relatively safe, I can send you the finished product if you still are up for it. If you want the completed version of the PSU, youll only have to keep following this thread and when its going to be done, you can ask for a price or maybe even the full schematic, but making the PCB will be such a pain that nobody will even atempt it (I think).

For now, its a protype, currently only the 12V version works, and I have not made a physical functioning mains powered version yet. Ill post news about it as soon as it happens. Also there will be a video on youtube about my PSU.

So we will see.
 
For diyers, it is very convenient to use a commercial mains input power supply with 12VDC regulated output (especially if this is a plugpak, or module style as per a laptop power supply) as the main supply to a valve amplifier. The 12Vdc is used directly for heater powering, and as an input to and internal dc/dc converter supplying the required B+.

The dc/dc doesn't have to comply with mains ac safety standards, and only handles the B+ supply power requirement. Many options available on ebay at very low cost, such as car style inverter products that may need minimal modification. And suitable for diy designs such as your 12Vdc input supply, and others may be interested if your design shows performance benefits over what has been described in a number of threads for ebay style dc/dc options.

The ac mains supply is then (if properly chosen) compliant to all safety and base level electrical product standards for the country of use, has a quality of mass design and manufacture with performance specifications and a warranty and peace of mind.
 
You are right, but car style inverters are noisy and it really requires quite the filtering. The other problem is, generating 500V from mains is a much more efficient process than creating it from 12V. As I mentioned, 12V 50W is a really big current, and further I want to expand the power rating to 100W, which will be simply impossible to get such power from only 12V without really pushing components to the limit and generating a lot of electrical and magnetic noise.

In this case, having an aditional 12V power supply is better than powering the entire thing from the ATX PSU, because such a high current draw from the amplifier, will drag the voltage down, for the converter it doesnt matter but the tubes instead of recieving the full 12V they will recieve 11V-10V. Which does make a difference.

If youre saying, I should not even bother making this because there already exists a product but to my knowledge it does give the noise performance I desire at all. Really the point was to create the most powerfull supply with the best noise performance, and without an extreme bank of capacitance.
 
.. but car style inverters are noisy and it really requires quite the filtering.

What amps have you measured noise on the speaker output that originated from such a car style inverter? I'd be keen to see some spectrum measurements and details on the converter and amp.

The other problem is, generating 500V from mains is a much more efficient process than creating it from 12V.

I would suggest that a well designed, modern commercial switchmode generating a 12Vdc supply is going to attain a very high efficiency, and even more so as the output current rating increases. PFC is making sense at even lower power ratings, alleviating mains pollution considerably.

For a dc/dc step-up, generating one output is typically a much easier design set of conditions, and a 12V input dc/dc would use a regulated input voltage rail.

12V 50W is a really big current, and further I want to expand the power rating to 100W, which will be simply impossible to get such power from only 12V without really pushing components to the limit and generating a lot of electrical and magnetic noise.
I think you should experimentally measure and assess some relatively simple modern cheap inverters typically found on eBay. I have tried a few, and your comment doesn't ring true imho.

In this case, having an aditional 12V power supply is better than powering the entire thing from the ATX PSU, because such a high current draw from the amplifier, will drag the voltage down, for the converter it doesnt matter but the tubes instead of recieving the full 12V they will recieve 11V-10V. Which does make a difference.
12V at 10-15A is pretty benign in the world of power distribution, and many higher current switchmodes that require excellent end-point voltage regulation typically use a far end voltage sense circuit - fairly common and practical. Valve heaters are rated to cope with -10% voltage. So I'd suggest this is a non-issue.

If youre saying, I should not even bother making this because there already exists a product but to my knowledge it does give the noise performance I desire at all. Really the point was to create the most powerfull supply with the best noise performance, and without an extreme bank of capacitance.
Without the forum seeing your design and performance efforts so far, it's a bit difficult for commentary on whether there may be room for improvement and learning above and beyond what some cheap dc/dc can provide. Aspects such as acceptable B+ droop and then current limiting without hiccups are important, as well as thermal management of all parts and good layout parts and traces.
 
Hello, I designed a SMPS that creates 0 audiable noise in a tube ampifier. Originally this was a school project for a competition
......................
I designed the amplifier on a PCB and the powersupply too.
..............................
What I would like to know, if there is anyone even lightly interested to see a switching powersupply for a tube amplifier.
..................................
Little about myself- Im a students I turned 17 last week and I was protiping this SMPS for about 6 months before i I got to a point of any real useable result. I made lots of amplifiers mostly transistor based or IC based AB amplifiers and Im also into analog technology. So basicaly a nerd :D

1) WELCOME to the Forum :)

2) CONGRATULATIONS on your advanced project. :cheers:

Yes, there are some details to correct.
Yes, please be careful when dealing with Mains voltages, follow all safety rules

3) I WONDER why when the OP offers to share his work, politely asking if somebody might be interested in seeing it, he is suggested instead to "buy some E Bay supplies" :eek: ??????????????????

Including a 5W one ??????????????????