more reliable amp: tube or solid state class A

i got to reading this thread:

i require no convincing that class A sounds better than AB or D or whatever else, but the efficiency is terrible, with the efficiency losses being reflected as heat.

and heat, as we know, causes thermal breakdown. this is a matter of engineering: the hotter a component runs, the shorter the mean time between failure. simple stuff.

but here's the question: if we took 2 equally hot-running amps, one tube and one SS, over the long haul, what would be more reliable? the tube amp, or the SS one?

i'm thinking the tube amp, solely b/c the tube is the hottest part, and its failure is accomodated for in the design (you simply plug in another tube). a hot running SS amp will eventually burn out resistors / transistors, and joe audiophile will be forced to send that to the factory for replacement.

(i am going to do some HVAC work on my room, and if i can keep in cool in mid July, i will be moving to the winner of this argument)

I'd bet on solid state class A.
I agree with the solid-state Class A amp also. Not because of the circuitry, which can be equally robust with both types of amps, but tubes are unreliable, especially output tube types that have limited lives and can become noisy, etc. If the heatsinking is done correctly on SS amps, they will take the brunt of the heat and any damage is localized to the sinks. If you really want to feel heat, try some OTL tube amps. I had some Fourier Panthere monoblocks that used 8 6C33CB output tubes with cooling fans--these things drew 5-6 amps at idle and doubled as space heaters. The unfortunate effects were a lot of internal heat damage, especially tube sockets.
I've yet to hear any evidence that class A is any less reliable. I've only heard opinion and presumption. I've had many A/B and A amps over the years and the only one that burnt out any transistors (FETs) was A/B. The Plinius amps I have now have been constant since October in class A. Cheaper than running the furnace.

Let me know when a tube amp comes with a 20 year warranty. On the other hand, it's easy to change a light bulb.

There are many A/B amps that are excellent that you will be limiting yourself from. If you're speakers are a demanding load, A/B might be the better choice for the dollar.
all kinds of Mark Levinson, Threshold, Krell, etc. class A amps running for years and years and years? Sound pretty reliable to me. My Forte 1a lasted 13 years before I went to another amp.
It would depend on your definition of 'more reliable'.
I would have to say the class A SS though.
A well designed solid state amp will always be more reliable than a tube amp. Properly designed means that the output transistors are operated well below their maximum ratings, and will therefore last a long time. Properly designed also means that the heat is directed away from the rest of the amp. On the other hand, tubes will always burn out after some time. I am firmly convinced that a tube amp will sound better but if reliability is the criteria then SS wins.
A fifth support for SS. Maybe the right word is "maintenance" rather than "reliable" in this case. The tube will need some maintenance -- i.e. first of all, changing the output tubes every 2-5k hrs.

The SS (and tube) requires changing the lytics -- every ~6-10yrs (more like the latter).
If you want an idea of what one renowned amp designer/builder thinks about Class A, the heat it generates, the effect on the amp's circuitry, etc, I would suggest a read of an owner's manual of an early Pass Labs Aleph amplifier (available, for example, here). I would never have thought an owner's manual entertaining but it is. One telling excerpt...

The amplifier does not require any maintenance. While the design is conservative, this is a hard running amplifier, as single ended Class A operation is the least efficient operating mode. In fifteen years the electrolytic power supply capacitors will get old. Depending on usage, you will begin to have semiconductor and other failures between 10 and 50 years after date of manufacture. Later, the sun will cool to a white dwarf, and after that the universe will experience heat death.
If I understand your question correctly, as an EE, my vote goes to the tube amp. Tubes are the hot parts of the amp and are outside the "box" in most cases. Heatsinks will invariably heat the amp and the capacitors inside it which will not be good in the long run. The capacitors are the parts that hate heat - not so much the transistors or tubes.
I think if you look at a 20 year old used amp the answer is in the demand for the product. How many solid state amps are valued for their sound many years after they were built. It is known that solid state amps will loose their sonic impact over time but tube amps just need a replacement of their tubes and you have a new robust amp. I have been running tubes for many years along with solid state amps and have never had a repair on any of my amps. But after 4 years on of use on my tubes and then a repacement of all my tubes it is amazing, I have a new amp. Is my solid state amp loosing sound as the transistors age, I do not know until I overhaul the amp.
Amp reliability is determined by the level of conservative design, and this applies to tube and SS. The reason that SS class A amps tend to be more reliable than tube class A amps is that a SS amp will blow up immediately if it is not designed with adequate margin. With tubes, you can get away with marginal design for a few months before failure. Designers of SS amps HAVE to build in more safety margin. Class A tubes amps can be built to very high reliability levels, but they have to be designed very conservatively. I do not know of any class A tube amps in this category. For example, a 6L6 tube is rated at a plate dissipation of around 30-35 watts. They are typically run at this level. If I build a class A amp out of these tubes, I will run them at a dissipation of no more than 10-12 watts. They will run many years under these conditions. The amp manufacturer needs to use six of these tubes (per channel) to make a 30-35 watt amp reliable in a tetrode mode transformer or ZOTL coupled amplifier. More are required in an ultralinear or triode mode amplifier. Many more would be needed for a traditional OTL amp, assuming that these tubes were appropriate for that type of amplifier (they are not).

On the other hand, it is of questionable merit to demand a class A tube amp, as tubes behave much better than SS devices in a class AB application. SS devices have much more abrupt cutoff characteristics than tubes, and as a result, they inject high-order harmonics in the crossover region. Tubes have a very soft turn-off and as a result, they don't inject these harmonics to the same degree. The gain in going from class AB to class A in the SS amp is much greater than it is in the tube amp. An fact, the class A tube amp makes little sense, unless it is a very low power amp. The idea of making a hot amp in general does not make sense to me in unless it is strictly a winter amp and you can supplement the work that your furnace has to do. To use that amp when you need to apply air conditioning to cool it is stupid. How does the noise of the blower affect your room noise level and your attempt to achieve good sound? We all also need to think more responsibly when it comes to our energy use. We have limited resources and we may be doing damage to our climate with excessive energy use.
"i'm thinking the tube amp, solely b/c the tube is the hottest part, and its failure is accomodated for in the design (you simply plug in another tube). a hot running SS amp will eventually burn out resistors / transistors, and joe audiophile will be forced to send that to the factory for replacement."

You are correct. In addition, some solid-state amps use output transistors that go out of production, so if you lose a transistor, you're in trouble. Caps are no big deal for either SS or tube designs, as they are relatively cheap and easy to replace (you can easily get 20 to 30 years out of them on a very high quality tube amp, even designs biased in Class A). If reliability is your concern and you are looking down the road, buy a tube amp that uses the same output tubes used in guitar amps (KT-66's, EL-34's, 6550's), as the Marshalls and Fenders of the world sell half a million tube guitar amps every year and you'll always be able to find replacement tubes.

In any event, as you noted, a broken transistor amp has to be opened up. A tube amp just requires re-tubing and re-biasing, and once that is done, you've basically got a new amp.
regardless of class or bias, buy what works best with your speakers.
We've been making class A amps for nearly thirty years. In fact, other than a few prototypes, that's been the only class that we have indulged in.

So I can tell you this, and it is not a matter of opinion but is simple fact. It does not matter whether it is tube or solid state, what *does* matter is how conservatively the amplifier is designed to survive operating class A day in and day out.

Tubes of course have a service life that is different from transistors and if set up properly, will be just as reliable in class A as any other class of operation. The same can be said of transistors as well. Its all in the design- nothing with the device.
I've owned two amps in the last 7 years. A Krell 200FPB and VTL 300 Deluxe Mono Blocks. Both were returned for repairs.
Over the long haul? Do you really expect to be using which ever you decide on for more than a few years? If you are seriously considering replacing your McIntosh amps already, it seems unlikely. It always amazes me when someone using this site says they've been using the same equipment for more than two years or so. I wonder what the turnover rate really is.
Anyhoo, solid state gets my vote. Less hassle in the short run and probably in the long run also.
+++ It always amazes me when someone using this site says they've been using the same equipment for more than two years or so. +++

What an strange comment. I have equipment that I will listen to till I go to the big sound-room in the sky. Incidentally, that includes a pair of Meadowlarks; you seem to be hanging onto a pair also.

BTW, Meadowlarks were designed to be used with vacuum tube amplifiers.

Hi Ralph-
You are technically correct according to most published definitions of class A, which usually define class A as being an amplifier in which output devices conduct for the entire cycle. However, I should point out that the definition of class A is a very loose one and can easily be met without meeting the true intent of class A. One of the most important attributes of the class A amplifier is that the output devices are biased so that power drawn from the power supply is constant regardless of signal. This attribute is an important one if one is to obtain the full sonic benefits of class A.

Now let's look at your 60 watt amp. I believe you use four type 6AS7 dual triodes for the push and four more for the pull. Each triode is rated for 13 watts dissipation, and you have eight of these in parallel for the push. I believe you are using about 160 volts for each power supply, so 160 volts is applied across each of eight triodes. In order to not exceed the maximum tube ratings, your idle current can not exceed 0.65 A for the combined bank of tubes. Now, if one were to assume linear amplifying devices such that one group of triodes turned off at the same rate that the other turned on, an idle current of 1.95 A is required to meet the peak speaker current requirement of 3.9 A into 8 ohms. This would be 120 watts peak power into 8 ohms (peak amps squared times load resistance) or 60 watts RMS into 8 ohms. Therefore, you are operating at a bias current of only one third of what would be required to maintain a constant power flow from the power supply.

Now how is it possible to meet the commonly published definition of class A with your amplifier? The reason it is possible, is because the definition does not address device or circuit linearity, all it says is that no output device can be turned off for the entire cycle.

The 6AS7 was never designed as an audio amplifier tube but rather a series pass tube for voltage regulators. The 6AS7 is a very non-linear tube and is difficult to actually turn off completely. When these tubes are used in a push-pull application, such as your amplifier, one bank of tubes turns on much more than the other turns off, so much so that it never turns off. Voila! class A!

To me, your amp meets the intent of class AB, and this is good. In fact, the point of my previous post was to criticize the idea that class A was required, especially for tube amps. Your amp is an excellent demonstration of why tubes can work so well in class AB in that they can have a very soft or gentle turn-off, and in the extreme, no turn-off at all (meeting the loose definition of class A).

I should point out that there are a number of SS amp manufacturers that use tailored bias schemes to cause the output transistors to not turn off, and they like to call their amps class A. I have seen some of these at shows, and you put your hand on them and they are barely warm. The problem with the definition of class A is that it is so loose that these amps indeed meet the definition. So...buyer be where. These amps certainly in general do not sound as good as a true class A SS amps because they still have abrupt discontinuities in the current through the output transistors, and the power supply current is highly modulated by the audio signal. Furthermore, if the user of my Berning ZH270 wants class A operation, it can be obtained with no changes to the amplifier; all one has to do is to find a set of poorly made output tubes that don't turn off completely to meet the definition of class A.

In summary, the point of this post is that the buyer should pay less attention to amp names, specs and classifications and listen with an open mind. In particular, the definition of class A is so loose that anything can be made to meet it, and I feel that the soft turn-off characteristics of tubes makes these devices ideal for practical amplifiers that do not have to be excessively wasteful of energy.
Hi David, Your comment that the 6AS7 is not that linear is not really true. The curve has a lot in common with a 300b and also 2A3s; in that regard the 6AS7G is in good company. On that point the idea that the amp is class A due to the non-linearity of the tube does not hold up. BTW if you study the RCA documents that they released on the tube back in the early 50s, RCA called the tube an audio amplifier *and* VR (Voltage Regulation) tube. FWIW for best VR operation linearity in the pass element (in this case the 6AS7G) tube is important!

While power supply stability is important, I don't think I would say that is why you do class A. IMO, you do it because of the increased linearity of the device. Particularly in push-pull, very low zero feedback distortion levels are possible. Achieving low distortion without feedback allows for lower amounts of higher-ordered harmonics, which are loudness cues for the human ear. Sort of a have your cake and eat it too thing.

Our power supply voltages are not as high as you describe, and it is a fact that the tubes operate class A2, which is something that I am happy to point out should anyone ask; like, right now :)

In A2 there is substantial grid current during part of the waveform, similar to class AB2. Our driver circuit is designed to handle the current, similar to the way Fisher did their class A2 amp back in the 50s. The 6AS7G is remarkably linear in the A2 window, like a lot of power triodes are (we built a 300b OTL once just to see if there was any advantage; the 300b has a similar A2 window BTW).

If the bias and B+ points were a bit different (IOW if the tubes went into cutoff before clipping; right now they cut off only *after* the amp clips), the amp would be class AB2, not A2.

I am if the opinion that class AB is harder to design for and get right since the driver waveforms have to be larger amplitude and the driver power supply has to be more stable. If you are trying to set things up without feedback, which is what I would do, class AB gives you less opportunity for distortion cancellation so you really have to have your ducks in a row to make it work.
Mr Berning and Mr Karsten (two of my favorite designers... FWIW) thanks for getting into the discussion. I have questions about some of the same issues and your input is helpful.

That's one of the (few) good things about AudiogoN, getting first-hand, good information from people who know what they're talking about...



This is why I come back here so often. Geesh, I'm not even a tube guy, but to get a chance to listen in on a discsussion between Mr. Berning and Mr. Karsten is a real treat indeed! Thank you gentlemen.
And amid the warm glow (shameless tube analogy) creating by listening in on a "conversation" btwn 2 audio greats, lets not lose sight of the most impt point.
In summary, the point of this post is that the buyer should pay less attention to amp names, specs and classifications and listen with an open mind.

There are very well executed designs that utilize tubes and solid state, Class A and AB, OTL, ZOTL and transformer coupled. ANY designer makes a series of decisions and compromises to meet their design objectives. Buy/use what sounds good, is well built, and well supported!! Mssrs. Karsten and Berning make products that meet all three criteria.
Yo Pauly, "strange comment?" As I read through these threads it seems very common for respondents to mention the many components coming and going through their systems. Some Agon members seem to change components as often as they change their socks.
For a while I was one of those. My Meadowlarks have been with me for about three years. They've been with me longer than any other piece. Although I believe the other components currently in play will be here for a couple of years.