Hybrid systems (valve and solid-state)

With most hybrid systems I see people combining a valve pre-amplifier with a solid-state power-amplifier, but how about combining a solid-state preamp with a valve poweramp? Is that a good approach too?

The reason that approach is of interest to me is because I like to upgrade to a valve system, but I have been unable to find an integrated valve amp or valve preamp with tone controls and I absolutely must have tone controls. I have many poor sounding records which force me to adjust the bass and treble, and without tone controls those records would become unlistenable. So to get the convenience of tone controls plus the valve sound, I would have to get a solid-state preamp that has tone controls, together with a valve poweramp.

And if someone could explain why designers of valve amplifiers never add tone controls, I would be very happy.


Edited by Analog Kid - 17 Nov 2008 at 6:54am

If there is one thing I have learnt the hard way, it is the saying "Thou shalt not plow with an ox and an ass together" (actually this is a biblical verse found in Deuteronomy 22:10)

Radar never really worked until the advent of solid state transistors because valves were unable to handle the current as well as the high frequencies required. This sends a warning in my mind, but that warning doesn't get through to the non-technical hi-fi buyer - he/she is not party to the wealth of knowledge the proper design engineer has.

So why valves? Nostalgia is a great marketing tool. It doesn't take much brainwashing after that. Brainwashing is about repetition - the more people that jump on the bandwagon the more effective that brainwashing becomes. The desire for valves has little engineering basis - those who understand audio engineering properly (and there are precious few left) know that the hype surrounding valves is mainly groundless. For example: The Davidian Sect convinced a group of otherwise sound minded young people that their leader was the only right person on the planet. However, there is a book few people read nor understand (the Bible) which points out the folly of their ways. It was far easier for those people to accept what that Davidian Sect leader was saying than it was for them to find out the truth. I can remember what the Feds had to do to release those kids from their prison: it involved the death of a federal agent right there in front of the millions watching the "live" broadcast, and then it probably took many days, months or years to bring the kids "down" off their "trip". Today another much larger sect exists warned of by a revelation! An "agent" died to bring us the message of the penalty, and again most are still "tripping" but a minority have come "down" off the "trip". So it is with valves.

A case in point is the Novo headphone amplifier. It is solid state. It uses 6 transistors per channel in a discrete (EG. not integrated circuit) implementation. It does not have "the transistor sound", and Miguel who is an ardent valve fan put it on par with his Singlepower valve headphone amplifier costing 9 times the price!

It is not the technology but the recipe! Almost anybody can make something that doesn't work.

Tone Controls

In the answer I gave yesterday to your question about the Era Gold and Reflex I included a plot which I reproduce here again...



Look at the green line between 100Hz and 10kHz - it has two valleys - these are the equalised phase. The red line above is the unequalised phase of the input. RIAA EQ is simply fixed-tone-controls. If you remember, I explained about phase margin? A tone control introduces phase differences when it is set to anything but zero (the mid point of the control knob). Here you see the 10kHz phase "dip" is some 70 degrees out, leaving only 110 degrees (180-70) of phase margin.

At full boost or cut, tone controls introduce a near 90 degree phase difference, leaving only just over 90 degrees of phase margin. The "loss-less" tone control configuration includes the EQ in the negative feedback loop just like our phono stages do.

However, instead of a non-inverting stage, "loss-less" tone controls use an inverting stage (no other way) and the preceding or driving stage has bearing on the circuit's stability. Unless it affords a very low impedance source to the tone control stage, the entire thing runs out of phase margin. It oscillates! It is very hard to accomplish a low impedance source with valves economically - a simple cathode-follower (just like a solid state emitter-follower) can only afford low impedance to the positive swing of the signal, and on the negative going swing you lose the phase margin. With solid state it is far easier to implement a cost-effective low impedance source on both swings.

The only alternative with valves, and the one that used to work in the era of the valve (which is not today) is the passive tone control. As it exists outside an amplifier stage (between two amplifier stages) phase margin isn't an issue. However, all passive EQ results in "insertion-loss" and requires up to a tenfold additional gain to recover the signal. Ten times is 20dB, and with 20dB additional gain the result is a 20dB loss in signal to noise ratio. That was acceptable back then, but today when the valve stage has to measure up to the same expectations as digital sources, 20dB of additional noise is out of the question.

The above should explain why today, tone controls don't feature in valve gear (or if they do, in very few offerings).

They can be successfully implemented in solid state however, and as many have encouraged me to make a preamp with tone controls, this is something in my design book.

One final note: Those wishing to deceive the customer cannot obviously do it by the almost pure engineering sciences. They need to garner support. The press have a reputation of being able to brainwash. I saw it take shape in the mid 70's to mid 80's. As a result, today, it is easier to set up shop cloning than it is for an engineer to set up shop designing things properly. Therefore, it has taken me 10 years to reach the point I am at now - because the cloners have enjoyed better sales through what I can only call "their wickedness". If the customer is going to be so easily led, then how can I offer more of my best?

Hang on there - I didn't say any technology is superior.

I nether said your own personal judgment was of no worth.

I simply answered your question about tone controls truthfully.

You can believe who you want.

If there is any bias in audio it is for the valve with solid state taking a considerable amount of stick. So what do I do? Lay down and take it?

The secret to success is by sticking a valve in it - doesn't matter what it is - stick a valve in it and it will sell - especially if it's stuck out the top like a phalic symbol!!!!!!!!!!!

But slave over a solid state design until it beats the pants off any valve design and you're still ignored. That's the way it is for me and my colleagues here! Yes, after 10 years of slogging our guts out a majority of people still don't want to belive we're any good - the valve is automatically better obviously? Why? Because it's a valve - no other reason!

I used to work with valve gear and it sounded stacks better than solid-state, but that was in solid-state's infancy - it is now 35 years later.

All we ask for is a level playing field but it's now got that slanted in favour of the valve we may as well jump ship.

And, as for butter, it's better than eating hydrogenated vegetable fat! But hey, margarine is hyped-up too!

Buy what your ears tell you - not what your eyes see or what those who tickle your ears tell you. It's hard isn't it? And don't blame me for trying to make engineering understandable - at least we try unlike the Doctor I just saw today!

Nice to hear somebody admit he only does it for marketing reasons - it kind of proves what I set out to say...

Valves have more gain only because of the higher voltage rails - rule of thumb (for the want of a better description) gives 20 x the anode voltage drop for a valve whereas it's 40 x the collector voltage drop for bipolar transistors, so we see the transistor actually gives more gain, but limited to it's smaller voltage rails I guess his argument wins...

...but if you utilise a current source a single transistor stage overtakes the valve by a mile - gains in excess of 8,000 on relatively low voltage supplies.

However, the valve is a high impedance voltage device where a bipolar transistor is a low impedance current device. 100Mhz would be great, but let's measure it into a load should we?

As for a warm glow, that can be arranged using LEDs - they never burn out either.