tubes and analog


I just "upgraded" from a Mac SS integrated to a Prima luna dialogue 2 tube amp. The reason that I changed amps was that i assumed that the tube amp would be a better match for my Zu Druid speakers. The amp change was a big improvment for listening through my CDP....but not so when listening to my Rega P9. I had to switch to my spare SS phono stage (Graham slee) to get it to sound right. I was using a tube phono (AES) with my Mac. In Short, my tube amp with SS phono stage is not really an upgrade from my Mac with Tube phono stage. My question is.....should i consider a further upgrade to a better tube phono pre or is it simply that a change from SS to Tube amp is more "pronounced" in digital playback?
csmithbarc
Hi Opus111, of course I do have a 'vested' interest, but I also stand behind what I say, IOW I try to practice what I preach FWIW. I hope that's OK :)

Rauliruegas, thanks for your explanation! No worries :)

I think Hagtech stated the matter better than I could. And I agree with him that if you are going to make a solid state phono section that does detail retrieval correctly, it will have to be some sort of zero feedback topology using FET style transistors. They are the only devices (I know of) that have linearity like that of triodes.

I for one would love to not have to work with tubes. High voltages, filament circuits, the production of tubes and the like all make transistors desirable. But so far I have yet in the 45 some odd years since they really started to appear to see them actually bring home the bacon. If they did, there would not be more tube equipment manufacturers here in the US then there was 50 years ago! The market has spoken very clearly to that.
Dear Hagtech: First my apology to Csmithbarc for this post but IMHO it is extremely important to speak about:

+++++ " The big advantage tubes have over solid state is that they are *far* more linear. " +++++

Yes, I agree with that in normal designs: in our design we develop a totally new mathematics theorem to make the transistor full linear, so that advantage dissapear.

+++++ " They also overload in a much more sonically benign fashion, tending towards compression rather than clipping. " +++++

I agree too in normal designs and in the past SS ones ( even on those times M.Levinson and Mcinthos introduce in their electronics designs a " soft clipping " stage. ). Today almost all SS designs take care about designing with very high overload levels so the clipping subject it is not an issue.

+++++ " Most solid state amplification employs feedback ... " +++++

No, more and more SS designs come with NON feedback design, as a fact the non-feedback design is the SS rule today.

+++++ " But as Atmasphere points out, the micro details and very small signal information is better recovered via tubes. This is not just opinion, but a technical limitation of topology. The exception would be an open-loop class A gain stage " +++++

It is untrue that there is a technical limitation topology, it is not: what could exist is a not so good design but there are a lot of right SS designs out there.
As a fact you state that there is an exeption: " The exception would be an open-loop class A gain stage ... " well this is one of the exeptions.

As you can see there is no single advantage from the tubes against SS, what exist is different designs ( good and bad ) in both technology sides.

You already know all the tube technology limitations, like you say almost all work with coupling caps or coupled transformers, the tubes are harmonic generators and the problem is that that harmonics does not exist in the original signal, the impedance problems are bigger too when a tube amp try to handle the " electrical speaker impedance ": almost all the tube electronics are high output impedance that when is connected to cables, audio devices or speakers change the frequency response: I can go on speaking about the tube limitations but this is not the subject. I believe that exist very good designs ( within its own technology limitations ) on both sides and we the customers have the choice.

Regards and enjoy the music.
Raul.
>>we develop a totally new mathematics theorem to make the transistor full linear<<

What kind of fluff is that? I'm not familiar with any of your designs, but coming up with some new equations on paper does not change the laws of physics. Active devices follow certain behaviors and have been well established. On one end of the scale is the triode, the most inherently linear amplification device used in audio. At the other end is the bipolar transistor (ok, maybe an IGBT is worse). In the middle are JFETs, pentodes, MOSFETs.

The bipolar transistor has a nearly perfect logarithmic transconductance. Hard to beat in that respect. But exponential is not proportional (linear). No theorem is going to change that.

If you have, however, invented some new circuit topology that does what you claim, then I suggest you patent it immediately. Silicon Valley is going to pound your door down.

>>more SS designs come with NON feedback design<<

Well, I think this is a little exaggerated. Most SS designs on the market employ opamps. Only a handful use open loop discrete gain stages. Perhaps this number is growing, but I hardly think it is dominant.

>>It is untrue that there is a technical limitation topology<<

What I didn't make clear here was the basic difference between class A and class B stages. As far as I know virtually all opamps run class AB. The output signal is driven in one direction by one device (transistor), and in the other direction by a second device. They keep handing off the signal. Only during a small crossover window do they both conduct. Making this window larger can help. Making the window full scale turns the amplifier back into class A. No device ever turns off.

In contrast, a class A stage is driven by a single device. It is always on. There is no cross over or handoff. This is the difference between a 2A3 SET and your typical SS amplifier. Why is it people will live with a flea-powered 2A3 when they can have 50 watts out of an LM3875 power opamp? Because of the micro detail. The SET is running full bore at idle. It excels at small signals and the distortion becomes vanishingly small. By comparison, the class AB amplifier excels at large signals. But it's distortion rises as the signal gets smaller. Exactly the opposite of the 2A3. That is an example of topology.

There are a lot of other topologies I can get into, but that example illustrates the point well.

>>tubes are harmonic generators<<

Yes. So is feedback. And the harmonics generated by feedback are far more insidious. They might be quite small in relative amplitude, but multiply with each pass through the amplifier, generating a lot of non-integer harmonics. That leads to the cold, sterile, and sometimes fatiguing sound of many high feedback amplifier stages.

SS stages that do not use feedback are also harmonic generators. Heck, every amplifier is to some degree. The question is, what sort of harmonics do you want to live with? I'll take a triode any day.

jh
In fact the odd-ordered harmonic generation that is a hallmark of transistors is a feature that runs counter to the rules of human hearing. We all hear the same way in this regard, and is not something that we can change.

The odd-ordered harmonic content we are talking about is used the human ear as loudness/harshness cues and occurs in vanishingly small amounts- hard to measure with conventional test equipment. But our ears are very sensitive to this type of distortion- finely tuned you might say.

Tubes do not make this type of distortion. So- in a phono stage, if you *add* this distortion, it will exist in the system right to the loudspeaker. So the phono section is the place where you really want to use tubes.
Dear Jh: +++++ " If you have, however, invented some new circuit topology that does what you claim, then I suggest you patent it immediately. " +++++

We are in this process.

+++++ " This is the difference between a 2A3 SET and your typical SS amplifier " +++++

Maybe the problem is that there is no more ( almost ) that " typical SS " topology. Today almost all preamps goes non-feedback, pure class A and discrete stages. Btw, we are using bipolars.

+++++ " Yes. So is feedback. " +++++

Like I told you almost the rule today are non-feedback designs. I think that you need some update about.

You certainly know every single tube technology limitations: very high output impedance, tubes are a noisy device " per nature ", tubes designs have higher distortions level than SS or hybrids one, tubes have heavy problems to reproduce accurately both frequency extremes specially the low bass where ( beteen other things ) does not have control over the woofers like the SS topology, tubes are untrusty: time to time ( very short time ) blow-up, tubes are not performance consistent: almost every single day sound different ( many people don't take in count because every day are in touch with their systems ). Jh you know that we can go on speaking about but this is not the subject: you and me know that there is nothing perfect in tubes or SS electronics.

Jh, the subject is that the " very old myths " about SS designs, fortunately, dissapear. Today we have very good SS designs as we can find good tube designs. The best of all is that today the customers have several choices that in the past don't.

Regards and enjoy the music.
Raul.