Power/Volume difference Tube vs Solid State

[Krwkka]

look, tube watts are bigger than solid state watts.

Aha! I knew that!

[Steve Conner]

I always found that the sound produced by the tube amps always floated comfortably above the other instruments and into my ears.

I like that! It nicely sums up what's good about tube sound. It has a kind of alive, floaty quality, that somehow helps the guitarist to hear himself comfortably but without complaints of it being too loud. Well, unless you're playing a Marshall stack and you WANT it too loud.

The small amp that I posted pictures of above is now together and working, you can read more about it here: scopeblog » Ninja Corvette Hybrid and get sound clips, schematics etc. I am pondering a name change to the "Corvettonator".

Technical note: The transistor output stage isn't theoretically quite as good as a tube one, by the standards set out in this thread. The reason being, tubes clip by running out of current, but the speaker voltage is free to keep moving after this happens, even to the extent of the "on" plate going negative. Any global NFB is broken, and the speaker is free to continue flapping, hurtling under its own momentum or whatever, until the power amp gets back into its linear region.

But the solid-state amps are usually designed to clip by running out of voltage, at which point they stick to the rail and the speaker cone gets damped. I think this is part of the explanation for the bigger tube watts. Loudthud has also mentioned it.

The solution is just to make the transistor power amp bigger, and the rails higher than necessary, so the tube part always clips first, but I'm not convinced it's a major issue. It might be interesting to build a SS output stage with diodes that allowed the speaker voltage to soar outside of the rails, and see if this made it sound "louder" when overdriven.

[woodyc]

Excuse me if somebody said this already but... tube amps generally have way less feedback, which means they have less control over speaker/cab resonances. This makes a big difference in output at certain (mostly low) frequencies.

[Steve Conner]

Yes, that is absolutely true. Newer solid-state designs use tricks to boost the output impedance. It works great to liven the amp up and make it sound more tubey. The earliest example I know of is the Polytone Mini-brute, but the Marshall Valvestate made it famous.

The solid-state amp I described above addresses the issue by having no NFB at all. The regular kinds of SS power amp circuits are unusable without NFB, and it took a lot of head scratching to make it work.

[teemuk]

Yes, that is absolutely true. Newer solid-state designs use tricks to boost the output impedance. It works great to liven the amp up and make it sound more tubey. The earliest example I know of is the Polytone Mini-brute, but the Marshall Valvestate made it famous.

The earliest examples I'm aware of are Triumph's Silicon 100 amps from 1965. The theory was actually known even earlier since it was used in "reverse-mode" in tube amps to reduce the output impedance. Bogen used the scheme in their tube PA amps, later it even landed to some tube guitar amps such as the Seymour Duncan Convertibles. Someone clever just figured that if positive current feedback is turned negative you can increase the output impedance - which served well in mimicking tubes.

The whole circuit was a big marketing point for Randall, already in the mid 1970's. Barcus-Berry used it too around the same time but called it "load correction". And of course companies like Peavey and Crate were also using it, at least since the early or mid 1980's. Starting from Rivera-period in 1981 Fender sold a whole line of solid-state amps that used the circuit and eventually they even devised another "power rating" to market such amps. "CIP" or "Current Impulse Power" was conveniently higher than RMS power rating because bigger numbers look sexier and sell better. It's a big number for output power, it means it must be loud, right? Wrong. All this way ahead of Valvestates and all. ...And of course we can't pass this topic without mentioning the 1985 "Carver Challenge" in which several golden ear audiophiles failed to distinct Carver's current feedback SS amp from the $$$ tube amps.

It's rather revealing that in early 1980's Dynacord patented a solid-state tube emulating power amp, and merely passed the current feedback scheme it utilized as a mere mention, as it was quite well known prior-art even back then. What was more important patent-wise was the diode clipping scheme wrapped around the PA's feedback loop. The guitar amp's gain control actually overdrive the the power amp itself and the clean channel simply switched the clipping diodes away from the loop. Way ahead of time technology back when companies like Peavey were merely introducing diode clipping preamp circuits. Though, that's another story. Germans were way ahead in SS guitar amp technology back then. Way ahead.

Today I take CFB as a standard since you more often see a guitar amp with it than one without. Today the manufacturers hardly even bother to market that they use the scheme or figure out fancy names for it. It's already so common that each of their competitors surely uses it as well. No point stating the obvious.

The solid-state amp I described above addresses the issue by having no NFB at all. The regular kinds of SS power amp circuits are unusable without NFB, and it took a lot of head scratching to make it work.

Actually omitting feedback is pretty regular as well, at least in musical instrument amp. Most hybrid circuits with tube voltage amp + MOSFET output tend to work that way, or if they do include a feedback loop it's not used for excessive levels of negative feedback but rather serves a purpose of e.g. implementing tone controls or levelling out the most extreme peaks in response. Sort of like in some tube amps that also come with feedback loops and have more damping due to that. The entire MOSValve line from TubeWorks was mostly based on driving MOSFET output stages open loop. Open loop schemes or low levels of NFB are quite typical stuff for solid-state instrument amps too, mainly for the reason of their poor damping and also because things such as common source MOSFET outputs or transformer coupled output stages can actually clip somewhat softly when the feedback is not linearizing the entire operation area up to rail clipping. All such amps really need is a DC servo.

[Steve Conner]

Teemu is right of course. But I think my design is better than the classic SRPP driving a bunch of MOSFETs, for the following reasons-

It uses BJTs which are better than MOSFETs for the reasons described by Douglas Self.

It has inherent current limiting.

It drives the transistors as current sources, so it has a high output impedance, unlike the MOSFET circuit which is still a kind of follower. NFB, presence controls etc. can all be added in the same way as a regular tube amp.

The interstage transformer adds whatever mojo transformers add, but doesn't have to handle the full output power, only about one tenth of it. (My prototype uses a Valve Jr OT)

It doesn't need a DC servo.