Converting a SMPS into an amp

[earthtonesaudio]

I'd like to convert a SMPS into a guitar amp.  Basically something like this 90W model but with an additional audio input, some preamp/controls, and the DC cable replaced with a speaker jack.

My first thought is to simply add audio modulation to the LED current in the optocoupler section.  Or perhaps replace the stock optocoupler with a two-input type.  Obviously you can't have DC running through your speakers so perhaps two identical PSUs with out-of-phase audio inputs and a quiescent balance control.
Also, it should go without saying (but I'll say it anyway) that this would be done safely and sanely, with great care taken to protect the builder/user/whoever from line voltages and other hazards.  I'm nowhere near actually opening a PSU to start tinkering, just want some opinions about whether the theory makes sense.

So... is this a practical idea?

[R.G.]

I'd like to convert a SMPS into a guitar amp.  Basically something like this 90W model but with an additional audio input, some preamp/controls, and the DC cable replaced with a speaker jack.

My first thought is to simply add audio modulation to the LED current in the optocoupler section.  Or perhaps replace the stock optocoupler with a two-input type.  Obviously you can't have DC running through your speakers so perhaps two identical PSUs with out-of-phase audio inputs and a quiescent balance control.
Also, it should go without saying (but I'll say it anyway) that this would be done safely and sanely, with great care taken to protect the builder/user/whoever from line voltages and other hazards.  I'm nowhere near actually opening a PSU to start tinkering, just want some opinions about whether the theory makes sense.

So... is this a practical idea?

I think you would wind up replacing everything, perhaps several times, to get anything that would work, and that operation would be marginal at best.

The idea of two switching power supplies running with variable PWM modulation to amplify audio has been around for a while. It's the essence of all "Class D" amplifiers. But the practicalities of getting this technique to (a) work at all, and then (b) work well have taken decades of effort. There are some Class D switching power amp chips out there. It would be far more practical to start from one of those app notes and work toward an amp than to start with a power supply and work toward an audio amp.

Power supplies have become highly specialized in how they operate. Un-converting these back to more generalized may be quite difficult as a practical matter. Maybe not impossible, but quite difficult.

[earthtonesaudio]

I hear you, R.G.

I wonder if a better starting point would be to find a brand of SMPS manufacturer that offers various output voltage models, and see if the only differences are in the output section.  If that's the case it implies that the line rectification/active PFC section is independent of the output section, and would be a better candidate for variable output.

Or maybe variable-output SMPS products are already available?  I don't think I've seen one yet.  Hm... back to Google I go.

[JKowalski]

They are typically built to land on a narrow range of voltages. This isn't just because of the feedback, but the construction of the transformer, the limits of the SMPS controller+switch, and depending on the circuit many small details. With a fixed output SMPS you are going to be doing some major reworking to get decent P-P output from these things, overall like RG I would not recommend it. Start from scratch. SMPS's are made to be SMPS's and amplifiers are made to be amplifiers.

[R.G.]

I hear you, R.G.

I wonder if a better starting point would be to find a brand of SMPS manufacturer that offers various output voltage models, and see if the only differences are in the output section.  If that's the case it implies that the line rectification/active PFC section is independent of the output section, and would be a better candidate for variable output.

Or maybe variable-output SMPS products are already available?  I don't think I've seen one yet.  Hm... back to Google I go.

You do know that there are Class D amplifier chip sets available specifically to do what you're looking for, right?

[earthtonesaudio]

The key idea I have in mind here is this:

An ideal fixed-output SMPS consumes some power x;
An ideal Class D amp consumes some power y;
therefore,
The idealized SMPS plus idealized amp consume power x+y.

Now consider a variable-output SMPS.  It consumes some power z.  To get a variable output you can no longer have the ideal SMPS, so it follows that z is greater than x.  By the same token, this new device is certainly not an ideal amp, so z must also be greater than y.
What I would like to know is if z is greater or less than (x+y).

I'm not averse to starting from scratch, but the fact that a SMPS already takes care of the line rectification and active PFC makes it attractive.  The active PFC in particular is out of my knowledge comfort zone but it's something I want to include.