SVT type tube amp with Switching PSU?

[okabass]

I just wondering how difficult would it be to make a switching PSU to 300 W tube amp (SVT type). If you suspect my mental health: I've already built an SVT clone.
http://www.okabass.suntuubi.com/?cat=8
Does it already exist some suitable modules like that style (link). The purpose is to save weight, make it smaller and keeping the tube circuit "as it is".

http://www.coldamp.com/opencms/opencms/coldamp/en/products/power_supplies/SPS80HV/index.html

But with c. 700 V, 1 A  for the anode , some 350 to 400 V to Vsg and the rest of the tubes. Plus some 6,3 V, 12 A for the heaters.

Is it reasonably idea? Is it overall possible? How much would it save the weight using a switching PSU and a toroidal output transformer? My clone (with toroidal mains trafo) is c. 30 kg.

 

[Junction]

Don't have an answer to your question, but I love your DIY SVT. Did you build this on your own or was it from a kit? I want to build one, but nervous about doing it without layout drawings/photos etc.

Cheers
Michael

 

[okabass]

Don't have an answer to your question, but I love your DIY SVT. Did you build this on your own or was it from a kit? I want to build one, but nervous about doing it without layout drawings/photos etc.

Cheers
Michael

HI

Thanks a lot! I love it, too.  It was built on my own, not from a kit. It wasn't an easy project, but the most rewarding of all my tubeprojects.

I havent seen any original SVT layouts. Photos can be found from eBay, and the schematic is no problem.

It is wise to be some nervous with the SVT.  Because of the complexity and the hi voltage you must be rather experienced tubeman. London Power's TUT 3 book has instructions for a  bit simplified SVT.

 

[schmidlin]

I have asked about switching PSUs for tube equipment for years now.  For some reason, the PSU guys don't talk to the tube guys.  Must be a sheep-don't-mix-with-cattle thing.  It seems like a viable idea, but I guess it won't handle the current transients.

Starting this conversation with an SVT amp is like asking to tackle Everest when you still can't walk.  Maybe someone should try this with a Deluxe or someting.

Congrats on the clone: you are a better man than I.  I tackled a 200 watter and it is still giving me static.

 

[nickt]

Have a look at the www.soundskulptor.com MP66. I know it's a preamp but it does include a switcher to make 300V from 24V (or there abouts). This particular circuit is real simple but might be a good starting point for mucking about. It makes a lot of sense for preamps/compressors etc. Not much sense for a power amp but you never know. Hope it helps.

cheers
Nick

 

[Svart]

I'm a PSU guy.

It seems like a viable idea, but I guess it won't handle the current transients.

Why not?  Add a ton of small-ish value caps(instead of a couple big value caps) on the output of the powersupply and you should be able to handle transients just fine.

As for the HV, use a high amperage output PSU and use a diode/cap/resistor multiplier.

Technically there is NO difference between a switcher and a linear if you are looking at the same things.  However, some amps like the Single Rectifier actually use rail sag during hard transients to create their "sound", thus when you use a dual/triple rectifier and change between the valve diode and the silicon diodes you can get totally different tones..

So make sure you actually WANT to change over to a switcher because it might actually be TOO GOOD for the circuit.

;D

 

[PRR]

> Add a ton of small-ish value caps(instead of a couple big value caps) on the output of the powersupply and you should be able to handle transients just fine

The audio transients can be swallowed fine with one big cap.

The many small cap technique is only needed for the hypersonic switching rate. "Clean" audio suggests switching onto a HF bank, a small choke, and then a big Atom or other fine audio bulk-cap.

> As for the HV, use a high amperage output PSU and use a diode/cap/resistor multiplier.

If I tap into a "12VDC" switcher at the AC side of the 12V diodes, I need 300V/12V= 25-stage multiplier. For SVT voltage, 50 or 70 stages. Such multipliers are not impossible but usually a last resort.

What we need is a cheap commodity PC power supply, with a transformer which may be easily re-wound with more turns of smaller gauge. But they are always jammed-in and saturated with shellac. Every time I trash a PC PS I look at the guts, but never see anything I want to try to re-purpose.

> How much would it save the weight using a switching PSU and a toroidal output transformer?

The switcher could be only a pound or a kilogram; compare with SVT power transformer for savings. You might or might not be able to use less main B+ filter cap; that's not the Big Weight.

Toroidial audio iron is NOT IMHO correct for an SVT. Get a little unbalance on a lighter-weight toroid and it saturates. SVT type Rock AND Roll is all about unbalanced sound and gigantic grind.

There is a patent for a "tube output transformer" which takes current from the tube, modulates to hypersonic, transfers across a high-frequency (small) transformer, and converts back to audio. The stunt should work on test-bench. It avoids most of the "coloration" of audio transformers. But "color" is part of the point of using an SVT. If you just wanted 300 extra-clean Watts, you could buy a transistor amp, lose the weight of OT and heater-supply, and even find some with no 50Hz/60Hz main DC supply for impossibly low weight. Carver had a SCR flux transformatator back in the 1980s, others use routine switching supplies. And still others claim that audio is no good without big 50Hz/60Hz iron.

 

[Svart]

The audio transients can be swallowed fine with one big cap.

The many small cap technique is only needed for the hypersonic switching rate. "Clean" audio suggests switching onto a HF bank, a small choke, and then a big Atom or other fine audio bulk-cap.

You're right.  It hard for me to not "overdesign" when my career has been based on making designs totally bulletproof and technically transparent to the circuit that is being powered.

I was thinking more along the lines of being able to power the devices in a way that they can reproduce the crazy rise time/slew rates created when squaring audio signals and being able to reproduce their upper harmonics as well since a lot of the sound of distortion is more about harmonics than the fundamental frequency.  But again I suppose we need to know more about how the PSU reacts with the rest of the circuits.  On the linear side of things, the rails will sag as the transformers saturates but you won't have hard transients created by this whereas a switcher, by nature, might go into current limit and give you strange behaviors like bouncing rails while the current limit moves the PWM frequency around wildly.

If I tap into a "12VDC" switcher at the AC side of the 12V diodes, I need 300V/12V= 25-stage multiplier. For SVT voltage, 50 or 70 stages. Such multipliers are not impossible but usually a last resort.

True again.  I was thinking of something that was a little more easy to do than rewinding a HV tranny.  One thing that might be do-able is to tap into the HV side of the switcher and regulate down(depending on the input voltage and the resulting HV potential) but the DIYer would have to be extremely careful doing this since the currents on the primary side can be significant and only the rectifiers/fuses/chokes/other assorted front end parts would be your current limiting at that point.  Depending on how loaded the switcher is the HV side can vary a lot too so it would have to be regulated heavily as well.

The best way is to charge-pump up from the output voltage in some way.  How much current are we talking about for the B+ anyway?  Maybe we can whip something up using a CCFL inverter as the basis for the B+ generation?  Digikey stocks some CCFL inverters that output 260-380(selectable via the "dimming" circuit) at 4ma driven from 3.3v.  The only problem is that there is a "strike" voltage which is a short burst of ~1400V that would excite the phospors in a CCFL.  You could clamp that down with a couple zeners I would think..

 

[jdbakker]

How much current are we talking about for the B+ anyway?

1A @ 700V, according to the OP. Doesn't sound too far off for a 300W tube amp.

While I agree that some of the frontend/intermediate stages may be low-current enough for a charge pump or (hefty) CCFL inverter,  I suspect neither will sanely scale to a level which could supply the output stage (and that's where all the weight gain can be achieved).

JD 'DIY 700V 1A SMPS == scary stuff' B.

 

[Svart]

Oh wow I didn't know we needed that much.. 

 

[okabass]

HI

Thank you all! Good points and ideas.

 

[schmidlin]

That crazy thing draws almost 900 watts out of the wall.  If someone can pull this off you will be my hero, for what that's worth  .

Again, any size tube amp switching PSU would be a big step forward, but of course, it would have to be cost effective and sound right.

 

[Gus]

http://www.qscaudio.com/products/products.htm

 

[PRR]

> Oh wow I didn't know we needed that much..

Sorry; I thot you knew what "SVT" spelled.

"Portable" (ha!) stage amp, intended for bass but has been used for guitar. Re-issued in many versions, some quite lame self-rip-offs. The original Ampeg SVT uses six large tubes working from 600V-700V and makes an honest 300 Watts clean power.

There was a short fad: the Fender 300 was similar, and Marshall attempted 400 Watts on similar plumbing (didn't survive stage abuse as well as the Ampeg or Fender). 

Output stage efficiency is over 50% so the total DC sine-test demand is closer to 500W nominal. Of course speaker impedance varies, guitar/bass is often overdriven, so it should support more than 500W DC load without crap-out. 700V 1A is not a bad goal.

If you size this up with 50/60Hz iron, and put a handle on it, you understand why these amps don't go to more gigs.

> tap into the HV side of the switcher and regulate down

Regulation is NOT needed. These things ate raw transformed-up wall power and it never hurt them. (Considering the way wall-voltage has risen in the last 30 years, and the occasional way-out wall outlet, it might be nice to cap the nominal 650V B+ at 700V.)

"tapping the HV side" is equivalent to a Hot Chassis amplifier and is NOT safe. We need galvanic isolation for life-safety.

The way my small mind sees it: the generic PC power supply takes 115V/230VAC and rectifies to 330VDC. A half-bridge slaps a winding. Two diodes and two caps could recover 328VDC, non-isolated. Or the winding is CT gets 330V peak per side, and an identical winding (without CT) to a doubler would give 660VDC. Which is what we want. But I think the CT driver is rare.

You can buy a "1,000 Watt" PC power supply for less than a SVT transformer. If the switcher transformer could be re-wound (AFAIK, a major pain), we'd remove most of the secondaries and wind a new secondary similar to the existing primary, or twice as many turns one gauge smaller wire. Ground one end, a doubler spews +/-328V; ground the "-328V" instead you get +328V and +656V DC, which is sweet. Toss a few turns on to get a nominal 4V winding to tickle the original 5V sense. Adjust so that it is normally un-regulated, but will cut-back if "656V" soars much higher.

You could even wind for heaters, and they can eat supersonic 6.3V just as easily as 60Hz or DC. But there is near 100W of heaters, our total load is 800-900 Watts, our winding skills are dubious, we may be pushing our luck with a "1000W" rated PC power supply (the makers know you don't have 1,000W inside a PC case except maybe for a few seconds at spin-up).

The other thing I see: fast-recovery diodes don't come in HIGH voltage. The doubler will need 650PIV diodes, plus enough safety margin to survive a gig in downtown Lodi or out in a field on a generator. And if normal-abuse means ~~1A DC, doesn't a doubler need "2A" diodes, with ample margin above the nominal stress? I don't know how fast or how fat we need, but I suspect this is not a common part.

> http://www.qscaudio.com/products/products.htm

Yes, at the ~~100V-200V level, the diode voltage rating is less of a problem. (But must be HUGE current rating in some of those amplifiers!)

 

[Gus]

I found a ham radio web page that someone made a high voltage switcher but it looked so dangerous I will not post the link.  I posted the links to QSC because of the power of the amps and the use of a switching supply but as PRR noted the voltage ratings needed at a few amps in a SVT might make sourcing parts a problem.

Maybe you can cut a bit of mass with using a switching reg for just the heaters and leave the B+ etc to a smaller custom made guitar amp transformer without heater windings?

  What is the power needed for the heaters?  I have only worked on one SVT and I opened it up on the floor because of how heavy they are so I understand wanting to reduce the mass.

 

[mcs]

The other thing I see: fast-recovery diodes don't come in HIGH voltage. The doubler will need 650PIV diodes, plus enough safety margin to survive a gig in downtown Lodi or out in a field on a generator.

How about these?

http://www.vishay.com/diodes/rectifiers/ultrafast-recovery-hexfred/

1200V and 6-16A.

Best regards,

Mikkel C. Simonsen

 

[Svart]

Hexfreds work great, so do the the SiC diodes

 

[bcarso]

> Oh wow I didn't know we needed that much..

Sorry; I thot you knew what "SVT" spelled.

Just what DOES SVT spell??  Evidently none of these:  http://www.acronymfinder.com/SVT.html

 

[AnalogPackrat]

Super Valve Technology.  NOT Sandy Valve Technology. 

 

[PRR]

> Just what DOES SVT spell??  Evidently none of these:

See small text: "We have 102 other definitions for SVT in our Acronym Attic"

Service Vacuum Thermoplastic
Super Vacuum Tube
Superior Vacuum Technology

"SVT" is a line-cord, intended for service (not fixed installation), extra-flexible (will last in household vacuum-cleaner service), and Thermoplastic instead of ordinary rubber (SV).

"SVT" seems to be spelled different everywhere you look:
"Super Vacuum Tube"
http://en.wikipedia.org/wiki/Ampeg#Super_Vacuum_Tube
"Super Valve Technology"
http://en.wikipedia.org/wiki/Ampeg_SVT
"300-watt all-tube phantasmagoria they called the Super Vacuum Tube— or SVT, to save on vowels."
http://www.ampeg.com/history.html

Even the weight is uncertain:
"At 85 lbs, the Ampeg SVT provided 300 watts of RMS power"
http://en.wikipedia.org/wiki/Ampeg#Super_Vacuum_Tube
"Unveiled at the 1969 NAMM show in Chicago, the SVT head alone weighed 95 lbs and contained fourteen tubes, six of which were massive 6146 power tubes. To heat all those tubes, massive transformers with magnetic fields powerful enough to cause genetic mutations were necessary." http://www.ampeg.com/history.html

I think Ampeg picked the TLA first, and all interpretations are after-the-fact. The vacuum tube technology is neither super nor superior, just hefty. The only odd thing in gitar-amp tech is the cathode followers to smack the power tube grids somewhat past grid current; but Macintosh did that in hi-fi and Bogen tried every which way to push grid-current.



> How about these? http://www.vishay.com ...

Hmmmm.... too many zeros for my brain. 50 nanosecond recovery time means they will absolutely suck at... 10MHz? So rates like 20KHz-40KHz are not a particular drag? The 1200V 6A part is $2 in-stock at major distributors.

With just 1,200V rating, has to be a Doubler. But winding issues suggest this also.

> What is the power needed for the heaters?

You've seen one; you know Franks, you could guesstimate.

Six 6550 is 6*6.3V*1.6A= 61 Watts. Eight or nine little bottles is 17 more watts. Say 80 Watts heater.

Power output is 300W rated. Suggested condition for 100W/pair is 600Vp, 100-270ma, 60W-162W in plate circuit, and another 10W in G2. Times three pair is 516W. Maybe 10W in little bottles, 525W. Add >25% for stage-fudge, indeed 650W-700W in B+ power.

> Maybe you can cut a bit of mass with using a switching reg for just the heaters and leave the B+ etc to a smaller custom made guitar amp transformer without heater windings?

Heaters: 80W
HV power: ~~600W

Even if you can get a 12V 7A switcher with zero mass, 680W to 600W is not going to save much weight.

They did do it about as good as can be done. Plate current means cathode and heat. Plate voltage does not need more heat, up to ~~600V where ion bombardment kills oxide cathodes. Down at 110V B+, heater power tends to exceed plate input power. Up at 600V, heater power can be far less than plate input power.

How do we re-wind a PC power supply transformer for a ~~300V secondary?