SMPS for tube guitar amp?

[soapfoot]

Has anyone experimented with a SMPS for powering a tube guitar amp in the most weight-efficient way possible? Do any commercial solutions exist for this, or example circuits from which to take inspiration?

It seems like it might be nice to have:

Up to 400-450vdc for B+
6vdc or 6vac for heaters
size- and weight-efficient

Might also be nice to have a circuit to simulate the “sag” of a tube-rectified linear supply if desired

The goal would be to make an extremely lightweight gigging amplifier that still sounds and reacts like a tube amp.

Neodymium speakers and lightweight cabinet materials are easy enough, but I’ve yet to see anyone dispensing with the heaviest component in the amp: the power transformer.

I know Manley have gone to SMPS in their tube rack gear, so I feel like it may be possible.

Any input?

 

[ccaudle]

Blackstar Amps have a line called St. James that they are advertising as the lightest 50W tube amps available.
The specs say 90V to 264V AC input, which sure sounds like the usual universal AC input range for switching supplies.
Blackstar Amps St. James line
Also has the "sag" feature you mentioned.

Although presumably you were asking about commercial SMPS available, thought I would mention these amps just to point out it is certainly feasible.

 

[soapfoot]

That’s a start! Thanks

 

[mad.ax]

Maybe this... 400V max tho, and lacks 'sag sim"...
Ali tube psu

Axel

 

[soapfoot]

That looks interesting!

 

[gyraf]

..and it has a "customize" option, which means you deal with whoever really makes it (or close)..

 

[soapfoot]

I found the same supply on eBay. Apparently you can customize it up to “<350 VDC” which would surely be enough for a pair of EL84 in class AB.

I wonder how well it works (and sounds) under a heavy dynamic load… what sorts of misbehaviors can one expect from a SMPS under guitar amp conditions?

 

[ccaudle]

lacks 'sag sim"

As a start I would just add an extra resistor inline with the HT supply, and play around with C values in RC filtering so that peaks get through with the supply relatively intact, but sustained load starts dropping the plate supply.

I will admit that is just an off-the-cuff guess, I have not actually looked at the dynamic output impedance of tube rectifier circuits to see how they behave under load, but my view on new designs is you don't really have to match exactly the behavior of some older piece of equipment, you just have to capture the essence of the sound with your own twist on it, so if RC or RLC power filters get something reminiscent of the old behavior, call it good enough and move on to tweaking other things.

 

[soapfoot]

True enough.

The goal here would be to create a lightweight amp to carry around town. Obviously should sound good enough to be inspiring, but a direct 1:1 clone of the performance of a spectacular vintage thing isn't the priority.

 

[ccaudle]

Because Google is a creepy stalker, I got this helpful reading suggestion this morning:
How Blackstar Ditched the Mains Transformer to Make...the Lightest 50W Valve Amps Ever
(article/interview at guitar.com)

 

[soapfoot]

I would assume Blackstar would rather not disclose their SMPS topology. And it's not quite inexpensive enough to make me want to buy one and reverse-engineer.

But at the very least it's an interesting proof-of-concept.

I wish I knew more about SMPS and could have a go at designing my own, but I've never messed with switching power before so would have a lot to learn.

 

[ccaudle]

I would assume Blackstar would rather not disclose their SMPS topology

SMPS is not magic, but getting SMPS transformers off the shelf for 300V, +/-15V, 12V, 6V, etc. outputs that they needed would have to be customized. Tube amp requirements don't match up with any common industrial or IT products.
That Alibaba link someone posted earlier is probably the closest thing to an off the shelf SMPS for tube designs you will find.
There are a few vendors who make modular converters, so you might be able to find something like a Vicor front end module, and then pair it up with Vicor output regulators and build up what you need. Will likely be larger and more expensive than a purpose built design, but at least I would have a little more confidence in Vicor modules than some no-name vendor from China selling on Alibaba. The Alibaba thing might be just fine, but hard to know without a track record to go by.

 

[spaceludwig]

The Crate V33 uses an SMPS. Powers 3 preamp and 4 power tubes.

 

[Matador]

SMPS is not magic, but getting SMPS transformers off the shelf for 300V, +/-15V, 12V, 6V, etc.

Only with a flyback topology - standard common-ground boost topologies only need a simple compact inductor, a diode, and a cap.

 

[ccaudle]

standard common-ground boost topologies only need a simple compact inductor

Ah, so get an off-the-shelf isolated supply to generate 12V, 24V or 48V, then build a boost converter to generate HT? That might be something to consider so you can find an off the shelf supply for heater supply, and then you only have to design a boost converter for 300V at a few mA (for preamps at least; you would need about 0.5A for a power section, right?).

 

[Matador]

you would need about 0.5A for a power section, right?

Depends on the output power: a standard 6V6 pair operating around 300V requires about 100mA peak to produce 15-20W (Fender juices this quite a bit by pushing B+ way up). This is easily doable. A 6L6 pair can need nearly 200mA peak to produce 50W when operated near 400V. For 100W designs you would be looking at close to 500mA (peak).

The real challenge is component selection to maintain these currents: MOSFETS with super high surge currents and low channel resistances are needed, and can be very expensive (sub 20 mOhm MOSFET's can run $20 EACH in low quantities), as can be the inductors.

Just to get a sense of scale: to produce 500mA at 400V minimum, at 100kHz switching frequency, requires a 95% duty cycle, with a 33uH inductor capable of greater than 17A saturation current, a 470uF low impedance capacitance, and a rather beefy Schottky diode capable of 17A as well! The layout and component selection would be critical, and likely outside the capabilities of a DIY'er at these currents (requires a lot of specialized layout knowledge).

 

[soapfoot]

@Matador, are those power requirements assuming rated power (e.g. 1% THD)?

What's the best way to calculate what kind of current would we'd need if we wanted to be able to drive the 6V6 or 6L6 outputs into pretty serious clipping without weird behaviors from the PSU (or, obviously, PSU failure)?

 

[Matador]

You can guess from the load lines: worst case AC current should be when one side of a push-pull is driven to near 0V grid to cathode potential. For a 6L6 that's around 200mA, so you'd multiply by the number of pairs of tubes. So the 500mA rating should cover all the bases.

As you push past 0V, you enter AB2 mode, which isn't typical in (most) guitar or music amps.

 

[trobbins]

This may be too DIY for you, but I use a generic AC mains smps plugpack with 12Vdc regulated output to power heaters directly, and through an ebay inverter for B+. Plugpacks with up to 8A are common, so maybe ok for 6V6 PP, depending on the heater requirement and then the B+ power requirement. I use an inverter pcb with an isolation transformer with various HV secondary tappings that can be diy rectified and filtered to suit the B+ needed. This path does require a broad range of experience to manage powering design, smps rectification, and noise control.

 

[Somkereki]

Has anyone experimented with a SMPS for powering a tube guitar amp in the most weight-efficient way possible? Do any commercial solutions exist for this, or example circuits from which to take inspiration?

It seems like it might be nice to have:

Up to 400-450vdc for B+
6vdc or 6vac for heaters
size- and weight-efficient

Might also be nice to have a circuit to simulate the “sag” of a tube-rectified linear supply if desired

The goal would be to make an extremely lightweight gigging amplifier that still sounds and reacts like a tube amp.

Neodymium speakers and lightweight cabinet materials are easy enough, but I’ve yet to see anyone dispensing with the heaviest component in the amp: the power transformer.

I know Manley have gone to SMPS in their tube rack gear, so I feel like it may be possible.

Any input?

Maybe good, but it MUST be shielded well (also magnetically) because it can have an unpleasant effect! A guitar amplifier has a high input resistance, it easily picks up all disturbing signals! Although I'm an old engineer, I haven't seen such a solution yet! The caps also wear out much sooner than at 50 (60) Hz! Worth it ? Only the weight decreases, the quality deteriorates!