Why tube amp sound better than solid state?

[Chuck H]

In fact, the list of players that have used transistor amps is longer than that. That doesn't mean they used them much. Or for long. Or for recording. What it means is that some transistor amp maker sent them one for free so long as that maker could say "So and so uses our amp!" Just pick a great recording from one of each of those artists and then research the gear used to make it... Tubes. Well, except BB and Willy, both of whom I love but do have crappy tone.

In fact, if you can find definitive evidence that a particular recording from that list of artists was made with a transistor amp I'll wager it's a crappy tone.

Re amping a smaller amp doesn't count either. Nor do effects pedals. They're effects. They sound different from basic tone intentionally.

[teemuk]

In fact, if you can find definitive evidence that a particular recording from that list of artists was made with a transistor amp I'll wager it's a crappy tone.

Yeah, of course it is. Whatever you say...

[Chuck H]

I have to make a concession. Rev. Billy Gibbons recorded much of Eliminator using valve preamps and SS power amps. And of course Tom Scholz used SS preamps and tube power amps (or did, when anyone was listening to his playing). I can't say I'm a big fan of either sound but they sure work for the players using them, so...

P.S. Don't roll those baby blues at me. I'll ask you on a date

[walkman]

Whoever says can we make that tube amp more like a transistor amp, at least in the guitar world. Better is subjective of course but marketing plays a big part in forming guitarists views. Vintage is in and tubes are more vintage than transistors so of course they are better. All the digital modelers out there model mainly tube amps not solid state except for the jazz chorus, as if that was the only SS amp ever made.

Reverse the question. What are tube amps bad at. Clean volume especially at bass frequencies , not as a rule but compared to modern P.A amps. Reliability can be a problem but then so too with SS. Weight , what else...

What about hybrids like musicman amps?

So midi controlled multi channel amps are out simple tube amps are in.
Speakers also make soo much difference to.

Suppose the question could be which tube amps are better than which transistor amps

[Chuck H]

Whoever says can we make that tube amp more like a transistor amp

LOL

Funny you should say that. I am one to play devils advocate. That's part of the reason I've been a pundit for tubes on this thread. But some of the qualities of transistor distortion are desireable sometimes. And I have been penciling a design that has a pair of cold clipping stages in series (padded in between) using some voltage regulation and fixed bias on that circuit. Just to create a more transistor like distortion character. The fact that it's all tube makes it novel. I expect to love it for really crunchy metal type tones.

[JoeM]

In fact, if you can find definitive evidence that a particular recording from that list of artists was made with a transistor amp I'll wager it's a crappy tone.

What style music are you referring to? One of my favorite jazz guitarists uses a 300 or 400 watt Walter Woods. I love my SS amps for clean too.

[Chuck H]

OK... Here we go...

For clean and really hard clipping it matters less. It's from 'just starting to clip' up to 'still some evidence that it's a guitar' that tubes make their mojo known. I've already been persecuted now for loving my old op amp preamp Bandit 65.

From what I've heard any bass amp would make a good jazz amp. And most bass amps are transistors. Why anyone would want to play bass on a guitar I've never quite been able to wrap my head around

JK. Different strokes and all

Peace

[JoeM]

I wouldn't want to have to choose only tube or only SS if I could only have one type of amp. I just want to point out that a generalization that tube amps sound better isn't always the case. Though the best sounding clean I've ever heard from one of my guitars was playing it through a Twin Reverb with JBLs (D120s), but I dont believe I could even lift that amp!

[Chuck H]

A common complaint for any of the "Twin" amps. I remember at the 2009 NAMM show there was a guy in the booth across from ours selling very specialized jazz guitars. He was very proud of his little 5W single ended amp. I don't even remember the name on it. But it was a very specialized jazz amp. IIRC it was lighter than any Twin.

[Alan0354]

Ewww... Chuck is an opamp lover!

The most advanced "analog modelling" work I know of in the public domain is this:
Main page

The various "Fetzer Valve" based circuits at runoffgroove.com are also worth studying.
Articles and Schematics

I have my own "12AX7 simulator" based on a LND150 MOSFET, and I've heard some surprisingly positive emails from people who just tried subbing it for a tube in random locations.
LND150 tube emulator attempt.

It was inspired by KMG's work and also the "Trioderizer"
The Trioderizer - a solid state triode

KMG's circuit is more accurate, but it needs several supply rails. I sacrificed some accuracy for simplicity, on the grounds that it might be more likely to catch on if it could be tried out easily by dropping into an existing amp.

I also have a transformer-driven transistor power stage which sounds somewhat tubey, and you doubtless all have seen before.

I looked through a lot of the circuits in Runoffgrove, most use JFET as tube and put the very similar circuit like the tube amp. Looks to me they use source by pass cap ( common source ) to make the JFET distort similar to tube. That's all they really do. The problem is with common source without degeneration, gain is not predictable and a trim pot is need for every JFET.

If JFET have similar characteristics of triode, why then almost all the SS amps use op-amp as preamp circuit?
Why don't people use bipolar transistors in the preamp?

[teemuk]

If JFET have similar characteristics of triode, why then almost all the SS amps use op-amp as preamp circuit?

Well, for starters, FETs alone aren’t really too much triode like. Their characteristic curves of operation resemble more those of pentode tubes. They do have certain tube like attributes though, like high input impedance, basically voltage drive, and they can be biased pretty much like a generic triode tube. This makes simple subbing them to circuits using usual tube architectures somewhat plain and simple. This is why they have become somewhat popular in circuits built by amateur pedal builder; even an idiot can take a tube circuit, replace triodes with tubes, drop the supply voltage to about 9V (for battery powering) and replace the plate resistor with a trimmer.

Except that the resulting circuit behaves nothing like a similar stage built with triode tubes and high voltage power supplies. …And the operating characteristics of each stage change drastically in accordance to how their trimmers are set.

Like you mentioned already, the FETs themselves are also irky devices in comparison to generic triode tubes: they have a huge parameter spread so they require a delicate biasing means that fine tubes each, individual device. A trimmer resistor is not the best option for this except for its sheer simplicity; past that point it gets to measuring and handpicking devices and building complex stuff like constant current source biasing circuits. This you may find from commercial applications that use FET gain stages but it is kinda expensive so you have real benefits of using FETs weighing in the cup.

Which brings us to the issue that it isn’t rocket science to make OpAmp –based circuit as “tubey” as any FET circuit out there.

Furthermore, FETs alone are expensive, they may be somewhat rare and a continuous manufacturing of parts isn’t assured. One can pick an OpAmp at lesser price, it will always have a substitute in case manufacturers decide to obsolete a certain model and it will have no considerable spread of parameters. It will almost always behave consistently in as simple circuits as plain audio amplifiers.

I’m not even going to go into other options of novel device choices like compandors, OTAs or inverters configured as amplifiers.

So, why bother with FETs?

Why don't people use bipolar transistors in the preamp?

Sort of for similar reasons. Today there are really very few valid reason to use those instead of OpAmps again. And besides, people do use them; For example, Peavey’s TransTube preamps are pretty much based on bipolar transistors. If that’s not a considerable commercial use of bipolar transistors in guitar amps I don’t know what is. In addition to that I commonly see them in fuzz pedals and other specialty circuits. As plain gain stages they have less troublesome and more consistent options today.

[J M Fahey]

Agree and add:

The problem is with common source without degeneration, gain is not predictable and a trim pot is need for every JFET.

This alone is a deal killer in any commercially produced and sold amp.
Even worse, trimmers only ensure that , if you wish so, clipping can be symmetrical.
But gain varies wildly (easily 3:1) .... compound that with cascaded stages: gains are multiplied, not added, so you can easily , in a relatively simple 3 stage preamp, have gain vary 27:1
How can you sell that ?
Of course, you can add another 3 trimmers, one after each stage, to compensate gain differences.
Ok, maybe you can get by with two.

If JFET have similar characteristics of triode, why then almost all the SS amps use op-amp as preamp circuit?

Because Op Amps are *killer* design blocks, inexpensive, repeatable, you can do anything with them *if you do your homework*.
Well, OK, you can´t feed them 300V DC and have them put out 100V RMS.
Not needed, anyway.
You can scale all your waveforms to fit between +/-15V.

Why don't people use bipolar transistors in the preamp?

Ah !! but they did !! For ages !!
From the 60's until today (and *most* power stages are still discrete), although Op Amps started to be massively used in the 80's.

Not forgetting that any Op Amp means a lot of bipolars inside a little case, with, sometimes, a couple Fets thrown in, usually at the input stage.

[teemuk]

Well, OK, you can´t feed them 300V DC and have them put out 100V RMS.
Not needed, anyway.
You can scale all your waveforms to fit between +/-15V.

Oh, but you can: just add a bit of external circuitry to bootstrap the supply rails and you'll flying pas the typical +/-18V rail limits.

Anyway, enough with offtopic on that part.

[Alan0354]

The main difference of op-amp is that it has very low distortion compare with a tube in open loop configuration. I looked through a lot of SS preamp circuits from different brands, I don't see any compensation for the distortion characteristics of tubes. By distortion, I don't mean the over drive distortion. I meant the non linear transfer curve of the tube. Seems like you get closer to the tube with transistors. I don't agree that transistors are more expensive and more difficult to design. I did both kind of design for years and I feel transistors has advantages that opamp don't. Noise is one of the example. There are a lot of limitation in opamps, like power consumption, noise and frequency response. These characteristics are almost mutually exclusive.

I believe the non linear of the tube give the more the characteristics and that the opamp cannot do.........at least in the circuits I read from Fender, Marshall and Peavey. They are just standard op-amp design. If I miss any of the opamp design that compensate for tube non linear behavior, please point me to some of the design.

[Steve Conner]

You are right, many SS circuits like the Marshall Lead 12 etc. don't have any emulation of the tube's transfer curve.

There are other circuits that do, but they usually use diodes or BJTs which both have an exponential law, or FETs which have roughly a square law. Neither is quite the same as the two-thirds power law of a vacuum tube, but I'm not sure how much the difference matters in practice.

When using transistors, a bit of negative feedback from an emitter or source resistor messes up the math and produces a whole spray of harmonics. To my ear, this works well with FETs, and makes them sound more tubey. Without the degeneration, the distortion is almost pure second harmonic, and that sounds a bit sweet and bland.

I remember one guy who owned a Vox AC30, and using measurements from his amp, he modelled every tube stage using op-amps for the gain and zeners for the non-linear elements. The result worked well enough that Korg bought the design from him, and he had to take the schematics down.

Teemuk has written an entire book on these circuits. If you want to get into the subject in this kind of depth, it's worth reading. I picked up a lot of useful information from it.

http://www.thatraymond.com/downloads...ttala_v1.0.pdf

The relevant bit to this discussion is around page 268.