Single vs Dual Supply Voltage Amplifiers

[shortcircuit]

mint state box of Tungsram PNP resistors

those might be worth big buck$

[nsubulysses]

@jmaf, if you have a big stash of germanium transistors, they can be worth big buck$. Pack 'em up and send 'em to me.

@TomCarlos, keeping with the KISS principle (Keep It Simple Stupid, or some say Keep It Stupid Simple) I lean towards the single rail design philosophy. It's more like a tube amp and doesn't rely so much on negative feedback. Much of the bad transistor sound people complain about comes from dual rail designs.

So if single supply design is more like a tube amp and more simple, did it eventually just get ditched in an attempt for more power? say using single supply 75V vs +/-75 for twice as much signal swing and more power ?

If that's the case, and maybe the transistors of the 60s/70s were limiitng how much voltage you could use for a single supply. Aren't there output transistors that can handle 200V or so now? Or maybe more?

If this is possible, it makes me think it could be fun to try to make a very loud bass amp with modern output transistors and a 200V (or whatever ) single supply design.

If Sunn concert series is 35V for the output section and it does 120-130W clean into 4 ohms it seems like it would not be that hard to make a 300W-ish bass amp with 100-200V supply ?

you got my mind going

[mhuss]

I worked on a number of power amps in the 80s that were (1) split supply and (2) had all (identical) NPN outputs. See Peavey CS400, CS800, etc. These were DC-coupled outputs, but had protection circuits that would blow fuses when outputs failed -- and it was always at least two, I used to bring outputs to gigs like spare fuses . Even though they blew up once-twice a year, we never had a damaged speaker (though to be fair, they were old JBL D140Fs!) I suspect the repeated failure was due to the design depending on closely matched outputs and we never could afford to replace all 10 at a time...

[nickb]

Having owned an Acoustic Control Corp 150 amp and having worked on a few, and seeing posts here on MEF, I was curious to read up on differences between Single vs Dual Supply amps. The ACC approach in the 70's seem to go with a single +75 Vdc, while other amps used the dual +35, -35 approach. I was wondering if there are differences in Ripple, circuitry limitations, design considerations, etc.

Anyone have any links to suggested reading material?

Thanks!!

I think you may be looking at things the wrong way. It's the amplifier topology that matters.

For example you can have the common push-pull with direct coupled output, or single supply with a capacitive coupling. These can use complementary or quasi complementary designs. But you can also have dual rail designs with flying rails , single supply with transformer coupling, single supply DC coupled bridged and many others too, I'm sure.

[nickb]

I am also in the South Hemisphere, in fact sharing a long frontier with Brazil, and what you say is impossible.
In fact, the workhorse power transistor from late 60īs to the 80's, and which I used by the thousands (like *everybody* else) was 2N3055.

Its PNP mate was MJ2955 , notice that originally it was not even "2N" because it was a single manufacturer specialty, and was both expensive and very hard to find ... in Southamerica that is

Check ALL Acoustic, early Peavey, Fender, Ampeg, etc. amplifiers, ALL used NPN transistors ... because those were the cheap available ones; PNPs were as rare as henīs teeth.

Brazil did not manufacture its own transistors and got them from the World Market.

Even IBRAPE which was the Brazilian branch of Philips, assembled theirs using European supplied dies.

Not to take sides at all, but I do recall my first experiences with red spot /white spot transistors in glass envelopes. These were all PNP. So, I'm not entirely surprised to see PNP designs in an old book. IIRC it was only when silicon came along that it switched to NPN predominance.

[nevetslab]

Back in the mid-70's and for the next 10 years or so, you tended to find mostly quasi-comp output stages, all using NPN xstrs. The higher power amps tended toward 2N3773's, even 2N6259's, and the lower powered ones using 2N3055's or their higher-voltage screened parts that we'd get from RCA or Motorola with in-house numbers. The PNP's just didn't have the same Safe Operating Area characteristics, so you tended to not build the output stages with PNP's.

In the early 80's, BGW had made a purchase for around 10 million pairs NEC complimentary parts....4-leaded 2SB706 PNP's and 2SD746 NPN's. We had taken a few hundred samples to Motorola to have them tested, and being impressed enough, went ahead with the deal. For the next many years, we made a lot of product using the NPN parts in quasi-comp designs, and a number of products using complimentary parts (BGW 750D/E). As time moved on, and the pile of 2SD746's was rapidly dwindling, we had to make a change. Turned the schematic upside down and began using the PNP's, until we finally switched over to the MJL21193/MJL21194s and other faster parts (Sanken 2SC3858 NPN, 2SA1494 PNP) in full complimentary circuits.

Sure, there were a lot of products out there fully complimentary, but....the earlier PNP's just weren't as strong, and they'd fail when pushed hard.

[loudthud]

So if single supply design is more like a tube amp and more simple, did it eventually just get ditched in an attempt for more power? say using single supply 75V vs +/-75 for twice as much signal swing and more power ?

It's rare to see a single rail Guitar Amp design above 100W @ 4 Ohms.

The earliest power amp designs used a driver transformer. You either used PNP germanium or NPN silicon output transistors. Thomas-Vox amps were made that way. When silicon transistors started to dominate the market, the problem was breakdown Voltage. There were no big silicon PNP transistors and very few PNP that could be used as drivers. That's all there was so you had the so called Quasi-Complementary output stage.

Once improvements in breakdown Voltage were made available, the problem of Second Breakdown became (and still is) a limiting factor. A transistor may be able to dissipate 150W and low Voltage, but at higher Voltages, current must be restricted below a dissipation of 150W by perhaps 50% or hot spots will cause the transistor to fail.

If that's the case, and maybe the transistors of the 60s/70s were limiting how much voltage you could use for a single supply. Aren't there output transistors that can handle 200V or so now? Or maybe more?

Lots of complementary pairs at 200, even 250V. A guitar amp is pure second breakdown torture. You either have to use an obscene number of transistors in parallel or go to a series arrangement (to limit each transistor's Voltage) or use bridge drive (see comment about Sunn Concert amp below). The most rugged On Semi examples are good for rated power up to 80V. If you are slamin' the rails with a distorted guitar, you'll be crossing over into the Second Breakdown region where the transistor will eventually meltdown.

Note: MOSFETs don't suffer from second breakdown, they have other problems but don't count them out.

If this is possible, it makes me think it could be fun to try to make a very loud bass amp with modern output transistors and a 200V (or whatever ) single supply design.

Good luck. Hope you can do the math.

If Sunn concert series is 35V for the output section and it does 120-130W clean into 4 ohms it seems like it would not be that hard to make a 300W-ish bass amp with 100-200V supply ?

The Sunn Concert power amp uses JM's fav 2N3055 in a floating bridge drive configuration with a special driver transformer. This limits the Voltage to each output transistor which is a good thing. Later versions used a totem pole configuration, still with a driver transformer.