SMPS for vacuum tube power amplifiers.(status: back at it)

[SK_Caterpilar_SK]

Hello. I designed a switchmode PSU for tube amplifiers. The PSU is capable delivering upt to 70W of power and runs of 10 to 30VDC on the input. The output can be adjust to anywhere from 200V up to whatever your filtercap is rated for (450V). This was originally an idea of mine to switch out a really expensive transformer for the High voltage which would be nowhere near regulated and really heavy. The Powersupply has its advantages and its problems but those seem to be so far easy fixes. The big deal is, would anyone be even slightly interested in such a supply. I know that SMPS were frowned upon because of their noise, while my supply generates 0 audiable noise at the speakear even when looking at it with a scope.

I originally intended it for my school project for a Tube amplifier. 20W RMS Mullard amplifier more specifically. Workse great no hiss nothing, I also posted the same question on a site called DIY Audio and one of the guys messaged me if it would be possible to power it with basically line voltages. My design has some problems and those would be fixed right away if working with line voltage, but I kept myself away from such a job because I simply dont feel comfortable enought to work arround line voltages.

Its basically a Flyback converter that stepts the voltage up to 450V with a lot of power to back it up.

Little bit bout me- Im a student and I just turned 17 last week and I basically consider myself as a  home grade enginerd. Im into mostly analog stuff, high power regulators semiconductor amplifiers and many other things. Before you askt, the entire high voltage power supply was created by me. THere is no existing schematic for such a powerfull flyback converter from said input voltages to the said output voltages

I wonder if it is worth it a all to continue the developement of the supply at all.

Btw im not a native speaking so I apologise for any grammar mistakes I made.
Thanks

[Benta]

First, I applaud someone really enthusiastic about electronics at 17. Keep it up!
About your current project, I have my doubts. Tube enthusiasts tend to be purists and like big transformers etc. I don't think you'll find customers in that area.

[TimNJ]

Most "tube enthusiasts" are perfectly content in the year 1955. It's audio, so there's a lot of ~feelings~ involved, as we know. So, you're unlikely to have much appeal with that crowd.

That said, an SMPS would allow you to cut down on size, weight, and maybe cost of a tube amplifier. Could be an interesting design challenge to see how small you can go.

I'm not entirely sure how you deal with the output transformer, though.

[SK_Caterpilar_SK]

Most "tube enthusiasts" are perfectly content in the year 1955. It's audio, so there's a lot of ~feelings~ involved, as we know. So, you're unlikely to have much appeal with that crowd.

That said, an SMPS would allow you to cut down on size, weight, and maybe cost of a tube amplifier. Could be an interesting design challenge to see how small you can go.

I'm not entirely sure how you deal with the output transformer, though.

The output transformer has to stay- its there to convert the tubes high impedance to low impedance that matches the speakers. Its a necessity BUT getting the power transformer out of the way it basically halfs the total weight of the amp with some advantages as of powering it from 12V batteries, cars, and more, also its tightly regulated and it has absolutely no noise whatsoever, and maybe even improved efficiency, but converting to 450V from 12VDC aint a an effective process. But I had a rather intresting requiest by one individual asking me to make a line voltage version. It would be much simpler in terms of managing losses and heat problems since more power but less current is involved. I had to use a IGBT for switching the transformer since a regular MOSFET has a body diode which creates unwated effects in the circuit. I know that people have feelings for the good ol transformers, and I dont blame them, go linear go quiet. But these days technology can achieve same, and the line being noisy aswell the transformers are not so different from my supply I designed. Im hoping to draw some more poeple into this, currently im recovering from the financial loss of making the tube amplifier itself and getting the HV PSU done. Once Im up again Im going to work further on the PSU. It will work of of proper mains rated from 90VAC to 265VAC with a power of abou 100W thats my goal to reach and up to 1kV of voltage. I dont have really the proper tes equipment for that except a old VTVM (vacuum tube voltmeter) I love dearly but the filament current source "barretor" seems to be not in the best shape at all. And it seems to soon require a simconductor fix. However it goes to 1kV full scale so no problems there. I just hope this project aint my last one xD. The 1kV output would be really catching for tube RF transmitters.

I know the stigma arround anything else than true linear transformers, but at some point..someone has to come up with something that might be worth a shot. I could not find ANY comercial device that could provide this amounth of power or anything close. Not even an IC that do the work so I had to design my own SMPS controller based on the literature of MAXIM integrated and making the ctriangle gen and PWM gen and error amp so experimental purposes. As of now the GA3459 switching transformer is rated for 56kHz switching. The line voltage transformer could go even higher up to 100kHz would really improve noise basically eleminating even the electrical noise I can see with my old HP digital scope with a divider probe amd the fluctuation is about 1V pulses. Again nothing noticeable at that voltage. I also have a RLC filter there fo it does a real good job at filtering (10R series  470uH series and 100uF in paralel. The previous stge has only a 8uF filtercap after the diode and a very big fat 630V SMD cap biggest one I have ever seen in my life.). If anyone would be interest I could upload a video on youtube abou this. But the biggest problem im afraid of of the people to really hold onto those tranformers like even if my solution fits into 50 dollars and thats well within the manufacturing cost the engineering cost and the component cost.

[SK_Caterpilar_SK]

First, I applaud someone really enthusiastic about electronics at 17. Keep it up!
About your current project, I have my doubts. Tube enthusiasts tend to be purists and like big transformers etc. I don't think you'll find customers in that area.

Well hey thank you its been my passion ever since I was 8- from blowing stuff up into making making functional custom made products for special use and other things. I consider myself as a ENGINERD

Yeah I have the same fear. Regardless I think people woul reconsider buying a 150 dollar transformer shipping not included or a 50 dollar same solution just lighter and not traditional as we could say. More versatile lighter maybe even more effective But I havent done math arround efficiency yet. I wanted to get it working for the competiton but it did not work out the way it was suposed to- the competition. You know how it works.. That one old guy saw me at the cross country competition this year totally crushing the school he brought hes students from and he probably made damn sure to not even let me compete xD. Ahwell screw competitions those aint bring money only pets my ego. But yeah it would have been the 6th competiton this year. Still got two more to go .

About trannsformers- one would have to be left and thats the output transformer. It is necessary for a tube amplifier because vacujm tubes are high impedance devices- they cannot drive speakers directly like transistors. Tubes need a way of transforming high impedance to low impedance that matches the speakers there is the big fat output audio tranformer for.

[T3sl4co1l]

Cool!

I did a few things like that at the same age, too, although I don't remember making an SMPS for that purpose as I always had mains transformers handy, and didn't need battery or mobile operation.

Here's an example experiment, merely a multivibrator, switch and rectifier.  Probably circa 2001 or so:



I didn't have any schottky diodes handy at the time, not in that voltage anyway (indeed, 600V and higher parts -- SiC schottky -- were introduced since then!), so I actually used a tube for the rectification itself, too.  17CT3 was another I played with, convenient to use with the 15V bench supply I had at the time.

As rectifiers go, tubes are actually pretty impressive, if you can ignore the pesky heater power requirement that is.   Like schottky diodes, tubes have essentially no recovery losses.  Unlike schottky, they don't have an exaggerated nonlinear capacitance response.  They also have no junction voltage drop (an advantage of the heated cathode, which actually has a negative zero-offset current -- yes, it generates electrical power from a temperature difference: it's a heat engine, if an extremely poor one!), which makes them better signal diodes than any semiconductor, for sufficiently small signals (under 200mV say).  Finally, they have massive internal resistance, due to the voltage drop of all that vacuum.  And, again, a massive heater just sitting there cooking away.

Since then, I've made many SMPSs.  Here's a dual output module for hybrid* supply duty:



*As in, a low voltage linear supply powering a switching supply, as opposed to an offline switcher.

I've got one of these in my ever-evolving hand built Theremin, https://www.seventransistorlabs.com/Images/Theremin7.jpg in the shielded module on the right.

It's powered by a 15V 1A wall wart, hence the hybrid title.

There are some small differences from the above circuit, like joint regulation (a feedback resistor from 6.3V as well as 100V, to the TL431; regulating just 100V prevents it from starting up because the 6.3V rail browns out from the cold heater load).

Note the juxtaposition of a switching supply in a radio frequency project -- a Theremin is very sensitive to RF noise around it, acting like an electromagnetic microphone at its operating frequency.  This is why the power supply is shielded, and it helps that it is common ground, not isolated -- when isolation is required (like an offline supply), filtering must be very carefully implemented, otherwise it just leaks out everywhere.  Understanding how to do this, is one of the most important and rewarding lessons of switching supply design!

Cheers,
Tim

[TERRA Operative]

I am definitely interested, myself and a musician friend have just been talking about this exact subject.
He wants to build himself a guitar amp that is as small and lightweight as possible for taking to gigs. (He used to build guitar amps as a side business years ago).

I'll be following along for sure.

[SK_Caterpilar_SK]

I am definitely interested, myself and a musician friend have just been talking about this exact subject.
He wants to build himself a guitar amp that is as small and lightweight as possible for taking to gigs. (He used to build guitar amps as a side business years ago).

I'll be following along for sure.

Good to know. Considering people want to see such a power supply even on the audio forum, I made the decision to go with the developement, leave the 10 to 30V input to 450V converter and go for the mains powered version.

[SK_Caterpilar_SK]

Cool!

I did a few things like that at the same age, too, although I don't remember making an SMPS for that purpose as I always had mains transformers handy, and didn't need battery or mobile operation...

Thats looks interesting- but it seems not to provide almost just very minimal power if any at all. Youre right about tube diodes. In fact my VTVM (vacuum tube voltmeter) uses a pair for rectifying the AC voltage so it can provide a reading. I probably have one of the rarest VTVms ever existing. Back in the 50s Tesla in czechoslovakia was the main manufacturer for really thee entire soviet european block. So this very rare voltmeter does not even have a schemati available really on the internet xD Not even the manual. Its called the Tesla BM388E (not the round style but the rectangular big box style). It was quite common to use batteries back in the days for ohms measurements this one has a tap o its transformer for just that. So yeah very advanced VTVM. Regulated heatersupply and such- lotsa fun restoring. The only thing im sad about is that the scale is getting crusty and the paint wants to flake of badly. The one I use is in relatively good condition but the other one I also have has its scale almost unreadable.


Back to the subject - I actually use fast recovery rectifiers instead of shottky diodes simply because locallly theyre not available for me and fast recovery rectifiers considering the high voltage and low current flowing trought its not an issue.
And oh boi that shielding tho. Lotsa PCB material there xD.

So yeah and considering people actually want to see a line powered version, well heck why not I ordered the switching transformer today and lets get going. I just hope it wont be my last project xD. SO the thing by ma estimation will be capable to go up to 100W no problems. This really scares me cause I always try to keep myself away from line powered projects. I mean no doubt that the 12 to 450V converter is nowhere near a safe device. But I at least know that there aint anything before that can supply A LOT MORE. But over the years I think its finally time to overcome that fear of running away from line voltages. And its really noticeable by looking at the eficiency of the 12V in version. By having it runing on a higher input voltage, it will allow me to double the power safely and make losses less of a problem.

So yeah, Im in it now. Going further to make this a really good option. So far the whole thing has been designed on a single sided board and THT components. The shear PCB size can be smaller as soon as ill use SMD components and probably go for the TL494 chip instead of making the whole thing myself. Also by going double sided it will help a lot. Im going to attach some images for you to look at.
(pics will be in a separate post)

[TimNJ]

I work on offline AC-DC SMPSs all day. Shocked myself many times, but haven't been seriously hurt. Classic electrician's rule is to "keep one hand in your pocket". That's a good rule of thumb here too. Try to work with just one hand, especially when the circuit is powered up.

Of course, there are many other preventative measures to prevent yourself from getting cooked, but that's a good start.

[T3sl4co1l]

How is a device intrinsically unsafe?

Energized, stick it up your bum?  Not very safe, no, wouldn't recommend that.

Energized, under a cover, or inside an insulated box, connected with insulated terminals, connectors, whatever -- not a problem.

We handle hundreds of volts on a daily basis -- to varying degrees of safety mind, but mains outlets are hardly mass murders.  YMMV by country -- US plugs being notorious (you can easily touch the prongs when partially inserted), but 120V also arguably not being very dangerous by itself.  UK and AUS plugs being safer, with the metal bit on the end, atop an insulated shank, and preferably with the outlets having shutters that retract when a ground prong is inserted.  And arguably being more justified in the increased cost, at the higher voltage.

All a matter of safe practice.  And where possible, safe hardware and tools.  You don't have to get obnoxious with rubber gloves and face shields and finger guards all over -- after all, the most unsafe protocol is the one that doesn't get followed! -- but there are some very reasonable and affordable steps that can be taken to improve confidence in things.

Tim

[SK_Caterpilar_SK]

I work on offline AC-DC SMPSs all day. Shocked myself many times, but haven't been seriously hurt. Classic electrician's rule is to "keep one hand in your pocket". That's a good rule of thumb here too. Try to work with just one hand, especially when the circuit is powered up.

Of course, there are many other preventative measures to prevent yourself from getting cooked, but that's a good start.

I know that rule very well. I did get bit quite a bit of times and usually it was because of my stupidity. I did not expect the inductor to zap me but the voltage was so hight it penetrated the isolation. Something that is so obvious, but for some reason im still an idiot sometimes. Nothing to major tho. Since that day I use speacial isolated gloves if I really want to look at stuff and handle electronics while turned on, I use covered probes right to the tip. And no the gloves wont be penetrated by the sharp leads sticking out the bottom of the PCB. And me having dry hands as the sands of sahara is also a bonus.

Learned my lesson lol. But the question still remains. I see a lot of feedback circuits, but Im afraid I dont get their function(how does it work exactly). I mainly look at the TL431 feedback loop with an opto. So far doing research about the regulation.

[SK_Caterpilar_SK]

How is a device intrinsically unsafe?

Energized, stick it up your bum?  Not very safe, no, wouldn't recommend that.

Energized, under a cover, or inside an insulated box, connected with insulated terminals, connectors, whatever -- not a problem.

We handle hundreds of volts on a daily basis -- to varying degrees of safety mind, but mains outlets are hardly mass murders.  YMMV by country -- US plugs being notorious (you can easily touch the prongs when partially inserted), but 120V also arguably not being very dangerous by itself.  UK and AUS plugs being safer, with the metal bit on the end, atop an insulated shank, and preferably with the outlets having shutters that retract when a ground prong is inserted.  And arguably being more justified in the increased cost, at the higher voltage.

All a matter of safe practice.  And where possible, safe hardware and tools.  You don't have to get obnoxious with rubber gloves and face shields and finger guards all over -- after all, the most unsafe protocol is the one that doesn't get followed! -- but there are some very reasonable and affordable steps that can be taken to improve confidence in things.

Tim

Unsafe..well there aint nothing safe on 450V of DC with abou 60W of power backing it, no matter if its generated from 12V or 240V. The thing is I dont like working arround mains because ITS UNFORGIVING. You do a mistake you pay for it. Doesnt have to be about touching live stuff. Even simple mistakes can turn into disasters. The deal is I simply never was comfortable working arround those voltages. Maybe fixing powersupplies, its not a problem, but powering my own circuit with mains? Something I havent done yet and also I dont really "have the balls" to just turn it on. Sometimes shit can go haywire and I simply hate to risk anything.

[Zero999]

Do away with the power transformer altogether and go with good old school hot chassis. This sounds dangerous, but it needn't be: the inputs and outputs can be coupled with transformers, with sufficient isolation for mains use.

[SK_Caterpilar_SK]

Do away with the power transformer altogether and go with good old school hot chassis. This sounds dangerous, but it needn't be: the inputs and outputs can be coupled with transformers, with sufficient isolation for mains use.

Now that is what engineers were thinking of when they were designing all american 5 and 6 radios. Make the chassis hot or not- aint a good idea if you want the thing to be showed of not just working. And if you want a metal enclosure for the thing to make it premium looking.

Now I would use a 3 wire plug anyways. But ground loops are still gona persist as an issue, and isolating the output from the input mains seems to logical since you dont want to spend tons of money for tranformers that will put your sound into the amplifier, afterall the catch of a tube amp is its high input impedance. Your solution, is not what Im after.

[chris_leyson]

Well done, a really neat idea to get the weight down 

Unsafe..well there aint nothing safe on 450V of DC with about 60W of power backing it

Know what you mean. If I'm powering up a small off line 15W flyback design for the first time I use a small 24V to 500V flyback set to something like 325VDC and drive it from a bench supply limited to < 50W. At least that way there are no fireworks and breakers don't get tripped if shit does go haywire.
One tip, put bleed resistors across the HV bulk storage caps and maybe a small neon indicator lamp as well. Simple reason is that if my off line flyback doesn't start up for whatever reason then there is nothing to discharge the bulk storage cap, apart from me and that is usually followed by an expletive. Always double check things and if in doubt just put a meter across anything that could be hot, play safe and enjoy.

[SK_Caterpilar_SK]

Well done, a really neat idea to get the weight down 

Unsafe..well there aint nothing safe on 450V of DC with about 60W of power backing it

Know what you mean. If I'm powering up a small off line 15W flyback design for the first time I use a small 24V to 500V flyback set to something like 325VDC and drive it from a bench supply limited to < 50W. At least that way there are no fireworks and breakers don't get tripped if shit does go haywire.
One tip, put bleed resistors across the HV bulk storage caps and maybe a small neon indicator lamp as well. Simple reason is that if my off line flyback doesn't start up for whatever reason then there is nothing to discharge the bulk storage cap, apart from me and that is usually followed by an expletive. Always double check things and if in doubt just put a meter across anything that could be hot, play safe and enjoy.

Yeah thanks. Im planning to series 2x450V caps so the bleeders will be the balancing resistors themselves. The reason why I want to series them is to raise the voltage capability. Running a 450V single cap at that voltage can be little too close to the danger zone. If something goes oh shoot and the voltage shoots up into the sky, fireworks are expected..or at least a mist maker.

[chris_leyson]

Do away with the power transformer altogether and go with good old school hot chassis.

 

That reminds me, I've been looking for a small HV transformer to retrofit to a old DAC90 receiver, two reasons, it's safe and I get rid of the hot dropper resistors. Haven't found anything that will fit yet.

[SK_Caterpilar_SK]

On the left side is the PCB of the tube amplifier. On the right is the HV converter.

Just a showoff of the working 10-30Vin converter module

[chris_leyson]

Nice 

[SK_Caterpilar_SK]

Heres the finished thing. Only the bottom PCB view for the SMPS.  .

[SK_Caterpilar_SK]

And some tube goodness. Warm orange glow with some blue powdery accent to it on the inside of the power pentodes.

[SK_Caterpilar_SK]

Heres another look at one of the power pentodes.

[SK_Caterpilar_SK]

And the last one.

[Zero999]

Do away with the power transformer altogether and go with good old school hot chassis. This sounds dangerous, but it needn't be: the inputs and outputs can be coupled with transformers, with sufficient isolation for mains use.

Now that is what engineers were thinking of when they were designing all american 5 and 6 radios. Make the chassis hot or not- aint a good idea if you want the thing to be showed of not just working. And if you want a metal enclosure for the thing to make it premium looking.

Now I would use a 3 wire plug anyways. But ground loops are still gona persist as an issue, and isolating the output from the input mains seems to logical since you dont want to spend tons of money for tranformers that will put your sound into the amplifier, afterall the catch of a tube amp is its high input impedance. Your solution, is not what Im after.

If input impedance is an issue, you could also use a low power pre-amplifier before the transformer, which would only need a small power transformer. Another possibility is to have a differential input stage for the valve amplif and couple the signal via Y-rated capacitors, which are safe to connect between the mains and exposed metalwork or a low voltage signal.

I knew you wouldn't like my suggestion, but that was the point. I personally like your project, but as mentioned above, many thermionic purists won't. The fact that it sounds the same and the oscilloscope output is identical doesn't matter. It's phycological. If I told you it's possible to use digital signal processing to make an amplifier with exactly the same characteristics as a one with glowing filaments, would you be interested? Probably not. Many people like old technology. This is no different to steam enthusiasts vs those working on cutting edge electric locomotives.