Pros and cons of valve amps

[Firebottle]

Not sure the MOSFETS sound like valves but I believe they perform in a similar way.

The MOSFETS do indeed operate in a very similar way to a vacuum tube but the real benefit is that they are a low impedance device (as opposed to valves being quite high impedance devices), hence the ability to drive speakers directly.

When a MOSFET is used in a power 'buffer' mode, i. e. without providing any gain, it has virtually no sonic signature.
Use a valve for the gain stage(s) and you do have the best of both worlds, IMHO.

[Arkless Electronics]

A bit of a generalisation there Jez don't you think?

Many valve amps have no trouble dipping as low as sub 10 Hz (not that you want that when playing vinyl) and valve amps using sufficient global negative feedback to get the output impedance down don't suffer from any lack of bass grip. The only "issue" here is whether the designer, when using negative feedback has a) used enough of it and b) implemented appropriate and competent time compensation. Some actually don't! I suspect that's why you see so many using 5 or 10dB GNF without compensation....makes you wonder if a lot of the circuit copy-cat merchants know how to implement compensation properly or at all.

The Radford STA 15 and STA25 are a point in question. They manage to deliver sub 0.18 Ohms output impedance so damping factor is superb (ok it's not sub 0.05 Ohms but that's almost irrelevant given most passive crossover losses), and there's no lack of bass or bass control. I've heard one of Anthony's Copper amps and that had no lack of control either, nor did one of Nick's when I listened in on a room that he was exhibiting in.

Perhaps it's more accurate to say that poorly engineered valve amps have little bass grip, and even more accurate to suggest that this only applied to valve amps that aren't primarily intended to drive high efficiency horns, or indeed driving high efficiency horns where SE makes a lot of sense.

Some SE amps can have remarkably low distortion through much of their bandwidth as the signal level is reduced (so are low in distortion) typically having assymetric transfer characteristics, so that any distortion encountered when over driven tends to be even order. PP designs tend to cancel out much of the even order harmonic distortion as they run back to back, having symmetric transfer characteristics. If over-driven they can sound harsher than SE equivalents as a result. It's true to say that used WITHIN their efficient operating bands, a well engineered SE valve amp is LOW in distortion, which blows any myth about all SE's having high distortion out of the water. Horses for courses really. You don't pick an SE amp to drive inefficient 'speaker designs or those with high back EMF and large cone inertia/poor damping.

I've heard countless systems using AN amps to drive inefficient speakers and in every case, the sound has been obviously high in distortion and low in fidelity, a point in case.

A final thought is that few valve amp makers pay sufficient attention to S/N. TRON is one that does (in fact, it's one of their specialities, engineering very high S/N ratios so that when used with high efficiency speakers, the "inky blacks" remain "inky black"!). This is important where driving anything much more sensitive than 94 or 95 dB/1w/1m.

Absolutely not!!!

SS amps have a huge advantage over valve amps in the bass and this is 100% measurable and repeatable. If you read my previous post again you will see that I said valve amps can have very good grip and punch in the bass... However it will never technically come any where near a good SS design. SS amps can be an order of magnitude better here!! And you are trying to "cheat" by using a Radford as the example Radford valve amps are technically pretty much the best ever made, yes even after all these years. They are hardly typical!

Power bandwidth (not small signal bandwidth) for a valve amp (in all discussion here I am referring to power amps BTW) is lousy. A typical 100W valve amp will deliver only around 10-20W at 20Hz! And that will usually be for several % distortion.

Direct coupled SS amps (most are) can easily be designed to give full power down to DC! Most are deliberately rolled of at a few Hz to avoid amplifying offset voltages from a preceding pre amp but typically a SS amp will deliver full output even at 5Hz and with no increase in distortion whatsoever. Crown DC300A is an example of a SS amp which is allowed to go down flat to DC.

Output impedance. For a typical valve amp using NFB (negative feedback) about 0.6R. For a typical SS amp about 0.05R or less.

It's completely game, set and match to SS at the bottom end! This doesn't mean of course that a valve amp can't be sufficiently good in the bass to subjectively seem impressive. They can. Really analyse it though, get the power up and the frequency going down, and it's no contest. SS wins and by a huge margin of x10 or more.

The reasons for all this and many other typical shortcomings with valves is not, per se, the valves themselves. It is firstly the presence of an output transformer and secondly (much less important and can be got round, sometimes, with difficulty) the fact that valve only come in "NPN". For the non technical what I mean here is that transistors are readily available in "mirror image" polarities NPN and PNP. Valves are not. Ask any audio electronic engineer what he would give his eye tooth for and most will say "PNP valves!"... including me!

Radford valve amps measure so good because that's what they were designed to do. No way in a million years were they in any way designed by ear.
Their superiority comes from the fact that everything about them was optimised to allow as much NFB as possible to be applied... exactly the opposite of what many audiophiles today want!
The huge (for a valve amp) 36dB of NFB used in some Radfords was enabled by two things:

1/ Truly superb output transformers designed by Arthur Radford.

2/ A wide bandwidth triode/pentode long tailed pair phase splitter designed by Arthur Bailey (who later became well known for "inventing" the transmission line speaker... yes , yes that's way it's in "")

Both of these things minimise phase shift and hence allow more NFB to be applied before instability.

Life's too short for me to attempt to explain Nyquist Stability Criteria to lay folks but take it as read that you can only apply so much NFB to an amp before it goes unstable. The phase compensation Paul mentions is part of maximising the NFB and allaying the point of instability. There is no such thing as time compensation in an amp BTW. Also applying only 5-10 dB of feedback would usually not need any compensation and is a very bad idea anyway as such small amounts of NFB can actually make things worse and result in much higher odd order distortion... all for reasons technically well beyond the scope of a post on AOS. For those with enough interest look up works by Scroggie, Baxandall, Atwood and Terman.

Single ended amps are intrinsically much higher in distortion than push pull amps due to the asymmetrical transfer characteristics they possess.
This applies to ALL amplifiers no matter whether valve, SS, whatever... Again life's too short for all the caveats so assume the push pull amp is class A to ensure we are comparing apples with apples. ALL single ended amps are class A. They wouldn't work otherwise! (as there will no doubt be a clever twat yes they can work in class C and E but this is only applicable to radio frequency amplifiers ie driving an aerial on the end of a transmitter)

If we take a small enough section of a curve it can be taken as being a straight line, even in the case of the asymmetrical curve of a SE amp. This is why small signal amplifiers (for this context pre amps phono stages etc) are very often SE and yet still manage very low distortion and work great.
Take an SET, let's say it is quite a powerful one at 20WPC. It's saving grace can be that its distortion V power will follow a monotonic function and so IF we are only using 1W with suitably 103dB/W type efficient speakers then the distortion at this low power level can be only 0.1% and so all tickety boo etc. This often of course exactly how they are used in practice and hence with the right speakers you can get away with it. Take the same amp and increase the output to 10W and distortion will no doubt be in the 5% region and completely unacceptable. Increase it to 18W, just a couple below its maximum rating, and you'll probably see something like 10%+ distortion. 10% is the maximum considered acceptable for "Tannoy systems" in railway stations etc as any more and the intelligibility of speech suffers!

Take the same big triodes and power supply and make them into a push pull amp, still using zero feedback, and not only will power go up but with good design it is possible to get no more than 1% distortion at 20W. A few commercial valve amps, some Audion,s come to mind but there are others, take this approach and it is vastly technically superior to using the same valves in an SET amp!!

I just wish I understood fuses and lengths of 3 core mains flex as well....

[farflungstar]

My Cary's have NFB that can be switched off completely or applied gradually. The sound is so much better without NFB it's a no brainer. Bass gains a little better 'clarify' but highs are f*cked, strangled, mangled compared to zero applied. This may be poor design (Dennis Had reputedly said he only included the function so people could hear how bad feedback was so maybe didn't use best practice lol).

Adey

In perpetual pursuit.

[Arkless Electronics]

You could get the best of both worlds with a Croft.

I vehemently disagree but for political (with a small p) reasons all I'll say is, to misquote Crocodile Dundee "call that a hybrid power amp? THIS is a hybrid power amp" http://theartofsound.net/forum/showt...n-the-pipeline

The main boards were also completed and tested (I didn't add photos) and in fact so much was completed that I could actually finish this in a day or so if I put my mind to it.... "why did you abandon it after spending about 5 weeks working on virtually nothing else Jez?" I hear you say... Because it would be very expensive, probably £4K + even bought from me, and so no dealer margin, and I suddenly thought to myself "Jez, You're only really doing this as technical tour de force.. to show off really... you've got more chance of shagging Rachel Riley then ever selling one so get the fuck back to doing stuff to pay the rent"

It was completed and tested to the point where I know it will work ie one channel going (both are built) with the housekeeping circuitry on a breadboard and a bench power supply in temporary use for testing...

Edit. I never listened to it even in mono form as it was, in this test form, too big, unwieldy and fragile to move from the work bench...

[Arkless Electronics]

I use the Croft 7 which is the best of both worlds being an OTL amp as well no problems with nasty transformer noise. The input valve and MOSFETS coupled with an OTL design work very well in my system. Not sure the MOSFETS sound like valves but I believe they perform in a similar way.
Very transparent sounding and marvellous with vocals and acoustic instruments.

To follow on from what Alan said it should be borne in mind that they are far more like pentode valves than triodes Comparisons with valves though only hold good to a certain extent... many aspects are most un valve like and in many ways more similar to a bipolar transistor (that's a "normal" transistor to you lot)

Mosfets are my usual weapon of choice for power amps though.... but not always

[Arkless Electronics]

My Cary's have NFB that can be switched off completely or applied gradually. The sound is so much better without NFB it's a no brainer. Bass gains a little better 'clarify' but highs are f*cked, strangled, mangled compared to zero applied. This may be poor design (Dennis Had reputedly said he only included the function so people could hear how bad feedback was so maybe didn't use best practice lol).

Adey

In perpetual pursuit.

See my remarks above about small amounts of NFB. For the technically minded I commend this excellent paper by Bruno Putzeys http://www.edn.com/design/consumer/4...ng-as-too-much

There are superb amps with loads of feedback out there just as there are with little or non BTW. So many ifs, buts and caveats come into it that it is not possible to say that NFB is a bad thing. See also "do all apples have pips" by Reeves and Mortimer

[Arkless Electronics]

Bet that lots killed the thread Unlike if it was on mumbo jumbo about magic beans in which case it would go on for 30 pages...

[farflungstar]

I don't think it will have killed the thread - FB is an important topic. Though I take your point about 30 page topics on magic beans.. but magic does exist... It's just unprovable hence its called magic.
Adey

In perpetual pursuit.

[Macca]

Vic says:

'I can tell you if an orange has got pips just by looking at it. I discovered this at the age of five and thought I’d got it made. Unfortunately, it’s not a marketable skill as such. These days they mark up oranges as seeded or seedless, so any fool can do it. '

[walpurgis]

Bet that lots killed the thread Unlike if it was on mumbo jumbo about magic beans in which case it would go on for 30 pages...

Just start a thread on cable directionality Jez. That'll keep 'em going for a week or two.