Tubes and solid state (not a flaming thread)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This thread is not to be ment for debates like "one is better than the other". It is ment to answer some questions to guys like me with low/moderate experience.

I have listened to a few SS and to a few tube amps. In total, I have prefered the tubes. I'm an ear guy - I like what sounds the most realistic and musical, not what digits and specs show. If yesterday I've been to a live orchestra and today I have to listen to audio gear, I'll say - "Wow, the violins there are very close to what I heard", for example.

So for now, I prefer the sound of tubes. But I don't want to think that SS is not ment for audio. I haven't listened to many SS amps, but I never found one that I liked.

To me the explanation could be - are the SS amps much harder to design properly? Or do the SS amps need different kind of speakers to sound properly?

I have listened to a pair of Fe103En BL horns and a 6B4G SE at home. I fell in love with it - the music comes around me and makes me swim in it.

When I plugged a DIY class A SE FET amp in these same horns.. There was good detail, alright. But the music was dead. Dynamics were poor. I lost this 3D around me. I was now able to hear where the music really did come from. Everything sounded synthetic, plastic.. just dead.

The tube SE gave to the music power, energy, life!

Even with this, I don't blame the transistors. Maybe they are harder to be made musical? Or they just sound this way? Or they are ment for a less dynamic music? OR this SE FET design is crap?

So IMHO, why do I have chosen to work with tubes for now?

-I am building only proven things and I have a personal proof that a tube SE amp gives me what I want.
-Tubes seem to need very few components in order to work. This is music and DIY friendly.
-I find there is more freedom when using tubes (encore DIY friendly).
-There is something very attractive in them and I can't really describe what it is. They look really cool, tube amps design seems visually very alive.
-They are simple and usually run from the first time.

I am curious to build the Jean Hiraga Le monstre. It is known to be a great sounding SS amp.
 
This question SS vs TUBES was also of great interest for me and my friends, and finally we discovered that LOW POWER SS SE designs can outperform good tube designs in all aspects, important for audiophiles. Special care to active and passive parts is needed, as well as some tricks in schematics.

Typical SS designs usually are not aimed at winning over tubes, they are to provide power and enough drive, and to be proud about their THD (marketing needs).

The SS SE design you referred to simply do not observe some important design tricks (I did a lot of such a looser gear also).
 
Member
Joined 2009
Paid Member
I like my 8~10 W 300B SET amp. It mesures worse than the 100 W sand amp I have next to it, but I like the mid tone and the sound stage I get with the tube amp.

I also like the low component count of tube amps. It has a simplistic coolness to it.

I'd be curious if it has to do with the lack of global negative feedback in the tube amp.

~Tom
 
I've been curious to know some of the tricks to build a good sounding SE, but maybe the most of them are tough to get known.

In this aspect I think tubes are more Hi end friendly, they are easier and there is more info on how to make them sounding nice.
 
From what I have been told, yes. The less feedback, the better.
It sounds livelier and less sterile then an accurate amp with high amounts of feedback.

This conclusion is absolutely correct in majority of cases.
The GNFB would work perfectly if amp would be composed of infinitely fast stages, with zero input capacitances. Achieving of maximum speed from especially output and driver stages of SS amps are rarely a target during design period.
Those who tend to use fastest possible stages, they come to perfect, live and involving sound even with GNFB used (Soulution).
I personally have come to using of UHF transistors at output stages. Being used at short signal path schematics, plus some more design concepts (like constant current consumed from PS under signal), they provide everything what one expects from tube amps, plus tight and well articulated bass is achieved.
 
Last edited:
Alvis said:
From what I have been told, yes. The less feedback, the better.
It sounds livelier and less sterile then an accurate amp with high amounts of feedback.
Or you prefer 'lively' to 'accurate'? Correctly-applied feedback (which includes the correct amount of feedback) is indispensable. Your CD player is full of servo loops. Most 'zero feedback' circuits are full of followers (which all have 100% feedback). Triodes work by using feedback.

This thread was supposed to avoid going down this well-worn route, but people trot out the usual myths so this leaves the rest of us with a problem: leave the error to stand and let it gain momentum, or seek to correct it?
 
I have an explanation to this fact, a friend musician and DIY told me, and here I want to share his point of view, which can be wrong.

Tubes essentially respond to the Child´s law, in which a current in a thermoionic diode is proportional to the voltage across the electrodes at the 3/2 power. And, any tube deriving from it, also has this property.

This means, that when a pure sine signal, entered the device is amplified, the plate current has a DC term, a signal of the amplified signal, and essentially all even harmonics. If, by example the signal is of 440Hz, then a signal of 440, 880, 1760, etc will be at the anode, and succeed that all of them are frequencies that exist in all music instruments, corresponding to a tone of upper octaves. So, they "listen" well to the ears.

Transistors, has an exponential law (I = Is * e at (qv/kT)power) where Q mean electron charge, T = temperature, and k = Boltzmann constant. It means that the output signals will have all (odd and even) harmonic, plus a fundamental, and a DC term in the Fourier´s series. In other words, a pure sine signal of 440Hz will let you a 880, 1320, 1760Hz, some (the 1320 and the 5th, etc) of which doesn´t exist in an instrument, and they "dislikes" own ears.

He also told me, that for musician, a 5% THD of tube sound is preferable to a .01% SS because the last generates sounds like an untuned instrument (Piano gitar, etc). I´m not musician, so I can´t listen any difference.

What do you think about it?
 
Not true, although often said. Both 3/2 law and exponential contain all powers in their Taylor series, so you get odds and evens from both. Even if this were not true the moment you apply feedback you get all powers.

The main difference between them is that the exponential has all coefficients positive so it rapidly gets big and expands signals, while 3/2 tends to compress signals and has alternating positive and negative coefficients. This difference may have some audible effect.

I suspect that some of the difference between BJT and valves is the scale of input signals which counts as 'large' (where distortion becomes significant), when compared with normal signal levels. For a BJT 'large' means anything over a few mV (because of the size of the thermal voltage kT/q). For a valve 'large' means anything over a few hundred mV to a couple of volts, depending on which valve. So typical line level signals are very large to a BJT but only just about large to a valve.
 
Also, must be have into account the velocity that the terms of the Fourier series decay as a function of the order of the harmonic, the "a sub x" terms of the power coefficients, and the sign as you told.

I don´t agree with your last thesis. If you input a proportional amount of variable signal to the biased device (tube/sand), in example, a signal of say, 5% they would produce similar amounts of thd, and it doesn´t happens.
 
Osvaldo de Banfield said:
I don´t agree with your last thesis. If you input a proportional amount of variable signal to the biased device (tube/sand), in example, a signal of say, 5% they would produce similar amounts of thd, and it doesn´t happens.
Not sure what you mean by 5% signal. A 10mV signal would drive a BJT into significant distortion. A rule of thumb for BJT: peak signal in mV = 2nd harmonic % (e.g. 5mV peak signal gives 5% 2nd). 10mV into a valve would give almost no distortion at all. That is why you can't build a line stage with BJT without lots of feedback - you have to get the signal across base-emitter down to mV or sub-mV levels to get low distortion.
 
Not sure what you mean by 5% signal. A 10mV signal would drive a BJT into significant distortion. A rule of thumb for BJT: peak signal in mV = 2nd harmonic % (e.g. 5mV peak signal gives 5% 2nd). 10mV into a valve would give almost no distortion at all. That is why you can't build a line stage with BJT without lots of feedback - you have to get the signal across base-emitter down to mV or sub-mV levels to get low distortion.

It´s totally true. I agree this point. I mean 5% of the bias: a 5% amplitude (35mV) signal of .7V for a Si transistor, and 1V for a -20V biased tube.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.