A Small Friedman-ish SMPS Amp

[ax84ch]

Greetings-

[Sharing some notes about a side quest I'm on] I'd been seeing some "switched-mode power supply" products available for tube preamps and wondered how they might work in preamp projects. Specs are pretty scarce but I ordered one as a gamble.

ex: approximately 30USD

transformer_cattle.jpg

Back in my AX84 days, I would have doused myself in gas and lit a match for even thinking about things like this ("Jim Marshall and Leo Fender are rolling in their graves!"), but I'd like to think I've expanded my thinking a little bit since then. The main appeals of a properly functioning SMPS in a tube (pre)amp for me are speeding up my build time and making experimental builds less expensive. Spoiler: this does both.

After some searching, I found a couple of posts about their use in amp circuits which made me think they might be usable for a small amp. Sadly, I can't find the links at the moment, but when I got one and started to fiddle with it, I found that it didn't immediately burst into flames with a pair of 6550's loading the 6.3V regulated supply and a resistor string loading the ~280v high voltage supply to over 125mA. That is "whole amp" territory in terms of capability so I decided to see if I could get a whole amp out of it.

Executive summary: I built a full amp using a pair of russian 6P3S (6П3С) tubes and have been working on it for a few weeks. It sounds great and I've not really had any nasty computer noises to deal with except at maximum gain and volume when it is mush anyway.

So far, the only odd behavior I've really had is that the SMPS does not like being turned off and on quickly. There is also about a 30 second warmup delay that you can't apparently can't disable without surgery. I wasn't going to use a standby switch anyway but certainly not with a built-in delay.

The SMPS voltage is rather low so you're not getting a 50-100 watt amp out of it but the upside is that you don't need to buy a 150USD or more output transformer either. I'll talk about the amp itself in a later post but this turned out well enough that I am planning another build with some differences to compare.

[ax84ch]

My first SMPS build was "experimental" which meant it probably wouldn't work and I didn't want to put a bunch of money into it. Thankfully, most of the parts were already on hand except the SMPS. I planned to use a leftover 8k push-pull output transformer from a Randall RM20 2xEL84 amp. The 6p3s tubes at 280V/125mA aren't going to make 30 watts, so I gave this OT a go. "AI" suggested that 8k was reasonable for these tubes, for what that is worth.

I built a screen-fed self-split push-pull (SFSSPP?) power amp with adjustable cathode bias because I wanted to get something working fast to see if the SMPS would choke. This worked well enough for me to go ahead and build out a preamp- a variation of the Friedman BE. It sounded great but the amp was very quiet. I whipped up an LTP, thankful that I left room for it, and the amp became much louder and I was then able to play it for a while to test it out.

I had no NFB at this point, so the amp had a lot of high end, even with a variety of snubber caps ala-Friedman. I decided to implement a global negative feedback loop which brought the high end under control. I have an adjustable depth on the rear and an adjustable NFB on the front. There is no presence control because the NFB control does basically the same thing for me, in my use. Amazingly, I got the PI phase right and had no squeal.

I lived with the amp for a week or so and it sounds great. I have a "structure" switch that selects between a 10k second stage (ala JCM800) and the hotrodded 2.7k/0.68uF alternative. I moved the gain control after the first stage (after a 68k/100k divider) because I prefer it in that location and planned a second gain control (ala Hi-Octane) with a bright cap to allow some voicing control over the tightness. That ended up getting pulled out.

The amp worked so well that I decided I wanted a loop on it. I have one of Saul's excellent tube loops on hand (https://reverb.com/item/25486628-all-tube-fx-loop) but didn't want to blow a 60USD part on this build so I decided to try to implement one of Merlin's ideas for FX send with a basic recovery stage. Merlin's first preamp book opened my eyes to the variety of locations you can get a decent FX send out of without a cathode follower being added. I think that got culled from later versions of the book but one idea is to add a 25k pot under the 1M marshall-style master volume instead of a direct connection to ground. The signal is low due to the 1M/25k divider but the source impedance is evidently not bad. I would like to understand that better.

Since the master volume would be needed AFTER the loop, I broke the signal after the treble wiper, added a 1M resistor, 25k pot, 22k resistor to ground for my FX send. The recovery stage is a basic 100k/1.5k 12ax7 gain stage that comes back to the original master volume.

The loop seemed to kill some high end because the amp is a little more subdued. I am using ~32pF shielded wire for the runs from the front of the chassis to the back, but I have almost two feet of it so maybe I'm getting so attenuation from that. I ended up getting rid of one of the 500pf snubber caps in the preamp to compensate and I'm satisfied with the result. The loop is always engaged so the tone doesn't change when you use it, except based on whatever you put in the loop. I put the longest guitar cable I have in the loop and couldn't tell a difference so that is a win.

I'm close to calling this one done. The next version will hopefully use a SendCutSend chassis, a larger output transformer and and fixed bias via a tiny 120V transformer. I look forward to seeing how the chassis...err... transformer and bias changes impact the tone.

Pictures to follow...

[ax84ch]

20260406_183011.jpg

20260406_183032.jpg

20260406_b183102.jpg

[Guy77]

20260406_183011.jpg20260406_183032.jpg20260406_b183102.jpg

Great project. Congrats.
Would love to hear clips!

Cheers
Guy

[johnnyreece]

Thanks for sharing! As a budget-minded builder, this is another great tool in the arsenal!

[ax84ch]

20260406_183011.jpg20260406_183032.jpg20260406_b183102.jpg

Great project. Congrats.
Would love to hear clips!

Cheers
Guy

Thanks!

Here's an early "in the room" clip through an Egnater Rebel 1x12 with a Lynchback in it.

SMPS-amp1.mp3

. With my cheap earbuds it sounds like I'm in the room but with anything else it sounds like a cellphone recording heh.

[ax84ch]

I have a couple of question marks on this that I need to investigate, but here's the basic idea, for anyone who is curious.

Screenshot 2026-04-08 095332.jpg

Screenshot 2026-04-08 095351.jpg

[didit]

Hello --

Unfamiliar with the expansive breed of amps in the Marshall/Friedman dynasty. The hands-on closest are the Trainwreck Express and perhaps the ur progenitor tweed Bassman. Unable to offer much except to complement the detailed effort in designs. A tad unclear what value comes from that capacitance multiplier.

Fairly familiar and definitely curious about SMPS applications in this space. I've a few prototypes boxed away in storage of preamps and reverb effects used to test a smallish 300V/30mA unit. Questionable view of the inexpensive Asian stuff coming from Amazon, Alibaba, et al having carefully inspected components and manufacture quality. A dreamed up project that'll doubtless never make my "top of stack" would be designing an immortal SMPS into the approximate form factor of tradition MI amp power transformers. This would retrofit without modification into existing amplifier designs and provide various power supplies up to all four voltages required to duplicate/replace the transformer, rectifier, and a choke or two -- i.e., filament, B+1, B+2, and fixed bias.

Best ..

[ax84ch]

Hello --

Unfamiliar with the expansive breed of amps in the Marshall/Friedman dynasty. The hands-on closest are the Trainwreck Express and perhaps the ur progenitor tweed Bassman. Unable to offer much except to complement the detailed effort in designs. A tad unclear what value comes from that capacitance multiplier.

I built something alone the likes of a Trainwreck some years ago but I was not studied on all the things that make a proper one "magic". Even when AX84 ended up with a wreck-inspired project, I didn't pay a lot of attention to it.

The Friedman "thing" to me is to take the sound of a great cranked up Marshall and make it available at most any volume. Some people will say they are too tamed perhaps but others like it.

I need to document that multiplier. It is a little module someone gifted me years ago. I somewhat understand the concept at a very high level perhaps but treat it like a heavy RC stage.

Fairly familiar and definitely curious about SMPS applications in this space. I've a few prototypes boxed away in storage of preamps and reverb effects used to test a smallish 300V/30mA unit. Questionable view of the inexpensive Asian stuff coming from Amazon, Alibaba, et al having carefully inspected components and manufacture quality. A dreamed up project that'll doubtless never make my "top of stack" would be designing an immortal SMPS into the approximate form factor of tradition MI amp power transformers. This would retrofit without modification into existing amplifier designs and provide various power supplies up to all four voltages required to duplicate/replace the transformer, rectifier, and a choke or two -- i.e., filament, B+1, B+2, and fixed bias.

Best ..

I too have a questionable view of the inexpensive power supply. I would have a hard time using it in a commercial product and hope that maybe something a little more trustworthy emerges at some point. I have no plans to make commercial products so this will work for me for now.

[drewspriggs]

Hello --

Unfamiliar with the expansive breed of amps in the Marshall/Friedman dynasty. The hands-on closest are the Trainwreck Express and perhaps the ur progenitor tweed Bassman. Unable to offer much except to complement the detailed effort in designs. A tad unclear what value comes from that capacitance multiplier.

Fairly familiar and definitely curious about SMPS applications in this space. I've a few prototypes boxed away in storage of preamps and reverb effects used to test a smallish 300V/30mA unit. Questionable view of the inexpensive Asian stuff coming from Amazon, Alibaba, et al having carefully inspected components and manufacture quality. A dreamed up project that'll doubtless never make my "top of stack" would be designing an immortal SMPS into the approximate form factor of tradition MI amp power transformers. This would retrofit without modification into existing amplifier designs and provide various power supplies up to all four voltages required to duplicate/replace the transformer, rectifier, and a choke or two -- i.e., filament, B+1, B+2, and fixed bias.

Best ..

If you want less janky > https://github.com/stefaweb/audio-psu/t ... 45-MAX-1.0

485v @ 100ma, switching way out of audio frequency, total cost for this board and the carrier (assuming you can do the SMD assembly; there's not much of it and the components aren't too tiny) is under $100USD all up.

I'm reverse engineering a version with onboard DC heaters instead of the 12v auxillary circuit.

[didit]

Thanks -- that seems like a cool running start on a viable complete solution ..

Hello --

Unfamiliar with the expansive breed of amps in the Marshall/Friedman dynasty. The hands-on closest are the Trainwreck Express and perhaps the ur progenitor tweed Bassman. Unable to offer much except to complement the detailed effort in designs. A tad unclear what value comes from that capacitance multiplier.

Fairly familiar and definitely curious about SMPS applications in this space. I've a few prototypes boxed away in storage of preamps and reverb effects used to test a smallish 300V/30mA unit. Questionable view of the inexpensive Asian stuff coming from Amazon, Alibaba, et al having carefully inspected components and manufacture quality. A dreamed up project that'll doubtless never make my "top of stack" would be designing an immortal SMPS into the approximate form factor of tradition MI amp power transformers. This would retrofit without modification into existing amplifier designs and provide various power supplies up to all four voltages required to duplicate/replace the transformer, rectifier, and a choke or two -- i.e., filament, B+1, B+2, and fixed bias.

Best ..

If you want less janky > https://github.com/stefaweb/audio-psu/t ... 45-MAX-1.0

485v @ 100ma, switching way out of audio frequency, total cost for this board and the carrier (assuming you can do the SMD assembly; there's not much of it and the components aren't too tiny) is under $100USD all up.

I'm reverse engineering a version with onboard DC heaters instead of the 12v auxillary circuit.

[ax84ch]

If you want less janky > https://github.com/stefaweb/audio-psu/t ... 45-MAX-1.0

485v @ 100ma, switching way out of audio frequency, total cost for this board and the carrier (assuming you can do the SMD assembly; there's not much of it and the components aren't too tiny) is under $100USD all up.

I'm reverse engineering a version with onboard DC heaters instead of the 12v auxillary circuit.

This is probably beyond my ability at the moment but it is interesting.

[drewspriggs]

This is probably beyond my ability at the moment but it is interesting.

Gerbers are included on the older board, so if you're ok with paying for PCB Assembly you can the boards made, SMD populated and delivered to your door through companies like JLCPCB. Only THT left for you to do then.

Thanks -- that seems like a cool running start on a viable complete solution ..

I did message the designer and ask if he'd be interested in doing one with DC heaters (paying for his time obviously), but he's a hifi guy so I suspect he wants AC heaters for his 300B's. Haven't gotten a reply, but maybe I'll get lucky.

I am working on a much sketchier version using a cheap $4 inverter from your normal Chinesium sources. The board I've designed is a carrier for the inverter that accepts 12v/2.5A power, has extensive input/output filtering (as it's a dumb switcher at 60khz with no dampening) to keep noise low, and has an onboard supply for heaters that's inrush current limited. I started trying to figure out soft HV start based on heater current, but need to think about if it's worth adding the required buffer/comparator/MOSFET for soft-starting; it's absolutely possible, but I was trying to keep the whole system as dumb as possible.

It will output about 320V with a bleeder (275v with a 10w draw), and up to 1.2A for heaters. I'm specifically making them for a bolt-in tube preamp power supply so not concerned with draw, but I'm currently using the same inverter to run a small 6N2P/12AT7 Champ-ish thing I made. +B is at 308v during idle and drops to about 290v if you hammer it for about 2w output.

[printer2]

I had the idea of using a cheap SMPS high voltage modules that are advertised as having 45-390V dc and up to 60W. I used a 12V laptop brick to supply it along with the heaters to make a Champ (or at least a reasonable facsimile) using a 12AX7 and a 12V6 (NOS in the $8 range). I made the cabinet as small as could be and fit in an 8" speaker (at least the mounting holes, a squarish flange making it smaller than a round frame). Not that that matters as far as the circuit goes. The module's voltage is 350V, I have it outside the chassis to keep it cooler. The module blows the rectifier diode on the output side when it gets too hot, it gets too hot when asked to step up too high a voltage and current. Put a fan on it or feed it a higher input voltage and it is more better behaved. I think at 15V it can put out more without overheating and at 12V in I expect it to do better. I plan on making a P-P amp with one and the heaters running of the 19V. But back to this amp, I have not had any problems with the module so far once I found it can deliver 15-20W on 12V.

[WRC34]

ax84ch - this project is fascinating. I somehow stumbled across your Facebook post about the amp while researching the RCA BA-2C (go figure) and it really has been repeating on me. I know you say specs are scarce, but have you seen the bigger 300V 300mA version listed on Amazon associated with this model? The listing has more info and seems like it would power an EL34 amp comfortably, but then again who knows? The downside to that one is there is no protective cage over it.

And man, the clip you posted on fb sounds killer!