Ultralinear vs. Triode vs. SET

A lot of comments here that fit my experience. I own, and like, SET amps. Yes, there are things that they do that really cannot be matched by other designs (realistic decay of notes, truly enveloping soundstage).

But, the difference between the SET topology and pushpull is perhaps not as big a difference as that between tube types. I agree with the poster who noted how much a 300b in SET and pushpull share common characteristics. I have heard quite a few 300b, 45 and 2a3 pushpull and SET amps. I own, and like the sound of my pushpull 45 amp, even though it doesn't do all that a good SET can do. The amps tend to sound most like the tube type, with pushpull topology sounding "tighter" and seeming to not have as much "bloom" and not delivering all of the harmonic richness of the SET.

On the power limitation/speaker compatibility issue, I would agree that, if one happens to like a speaker that demands more power, then the priority is to find a compatible amp, and if that means foregoing certain topologies, that is just one of MANY compromises one must make when building a system (to me the priority is speaker choice first and foremost, then finding a compatible amp--not the other way around). But, I find that most people overestimate how much power they really need and the range of speaker choices for SET amplification is really not as narrow as some would believe. I've heard several 845 and 211 SETs that would work with almost anything, provided the room is not too large and one does not require extremely high playback levels. If one finds that much more power is needed, I would look at OTL amps first, then solid state amps (I tend to dislike higher powered pentode tube amps).

On the subject of OTLs, it may well be the case that a good OTL would be a first choice, regardless of power requirement. I love the speed, liveliness and "directness" of OTL. Within their power limitations, SETs can also have that quality, but, it is really hard to beat an OTL in those regards. To me, the tradeoff is that the sound is a touch "raw" or "rough."

One more thing, there is a pretty wide range of capabilities of SET amps. Although the SET topology is quite simple, SET amps are actually quite costly to do correctly because of the demand placed on the output transformer. In order to handle the DC current in the primary, the transformer must be big, air gapped, etc. SET transformer design requires even more complicated balancing of tradeoffs than pushpull transformer design. The better SET amps tend to be quite expensive.

Ralph makes an excellent point in the inherent superiority of an amplifier employing triode tubes to one operating tetrode/pentode tubes in triode.

However, Ralph's statement, "The result is an output section that that has nearly triode linearity, and nearly pentode efficiency" undersells pentode operation, while overselling ultralinear.

Things seem to fall right inline with simple arithmetic interpolation - connecting the dots. Two tubes in pentode push-pull operation can produce 60 - 75 watts, the classic 30 - 40 watts in UL that we all have grown accustomed to, and 15 - 25 watts in triode (not muddying the waters here with Class A, Class A/B, fixed bias, cathode bias, etc.). Without providing analogous data for linearity, the fact that an abundance of current day high-end audio amplifiers provide the means to switch between UL (often mistakenly calling it pentode) and triode operation makes the point that again, the distance between the ultralinear and triode is not insignificant. As it always seems in life, we just don't get something for nothing - the old "there's no free lunch" adage.

Continuing on the pentode - ultralinear - triode vein, what almost always turns out to be the case when one is talking about pentode in this day and age, they mean ultralinear. That's why there are so very few two tube per channel high-end audio amplifiers putting out 60 - 75 watts today. Thus, very few of us have actual experience with them.

Although the tube linearity of ultralinear operation may slot in between triode and pentode, there definitely is, to my ears, something else at work. Ralph has long argued the destructive sonic effects of feedback, and points out that ultralinear operation introduces feedback into the equation. That might explain why I hear the argument of the sonic superiority of ultralinear over pentode oten fall apart. For those who have heard a true pentode amplifier, it can actually be a revelation; there is an openness, clarity, and immediacy that jumps out at you. Maybe that is why some of the older designs (some of the famous Scott amps) which employed true pentode operation continue to be respected, sought after, and even prized despite all of the extraneous circuitry onboard that the past generation or two of high-end audio have stripped off.

I'm not advocating pentode as the be all and end all, but it's more than obvious that the pendulum has swung far too much in the other direction. I believe the hobby would be better served if there were more pentode amplifiers available to us. In no way do I think there is anything insidious at work, just a kind of "because that's the way grandma did it" scenario. It's one of those things where the argument is accepted as truth without another word because it's been such a long time since there was any challenge to it. Obviously, I'm standing up and challenging it here...

My problem with what I consider the better (45, 2A3, 300B, etc.) SET amplification is that there simply is not enough power to truly do justice to the music. You can get sound, you can get volume. You can even a lot of volume. However, in my opinion, if you really want to power a real world, full range loudspeaker the way they were meant to be, these amplifiers are not a very good way of doing so.

Many feel the higher power SET candidates such as the 211 and 845 are the answer to that. In my experience, you definitely do lose that magic with these tubes. Though we're still dealing with SET, it really is a different kind of sound.

As has been mentioned, push pull triode amplifiers are another (often excellent )option. Though again, you do take a step back from that sought after SET quality.

So far not mentioned here is the parallel SET (PSE) topology. In my experience, you lose little, if any, of that SET purity, while effectively increasing amplifier power in the expected ratio (two tubes provide double the power of one). Still, I would argue, not to high enough levels to adequately power a full range loudspeaker, but for at least a lot of audiophiles, you're close enough to have that conversation.

Another route for someone who is drawn to the 45, 2A3, 300B type SET would be powered subwoofers, a loudspeaker that features onboard amplification of the low frequency drivers, or rolling one's own with separate amplifiers, an external crossover device, and loudspeakers that allow such an implementation.

Trelja,

I do agree with a lot of what you said above. The transmitter tube SETs (845, 211) do not have as much of the favorable qualities of the low-powered triode tubes--like almost everything else, it is a matter of picking tradeoffs--but I don't find them completely wanting.

My primary amp is a parallel SET (two 2a3s per channel). I like this amp a lot, particularly when the 2a3s are EML meshplates. A friend, who designs amplifiers, said that he personally could not get that topology to work; one tube in the pair inevitably ended up hogging the current and doing most of the work even when the tubes started out as a matched pair. I have also heard other claims that some "purity" is lost and the sound becomes a bit muddled when more than one output device is used (the same claim is made for pushpull and even transistor amps). Also, while output is higher in parallel (I would guess primary impedance is also lower) we are still talking about pretty low output. My amp is rated at something like 6 watts from the parallel 2a3s. Am I losing something because it is a parallel SET? I don't know, but I like the amp.

I have heard a number of nice sounding pushpull pentode amps that put out 40-60 watts using two tubes per channel. For the money, small pushpull amps running EL84s sound pretty good to me (punchy and fun sounding). What has been most disappointing to me are the really big, expensive pushpull amps with many output tubes, particularly if those output tubes are 6550s. These sound brittle and dry to me (MUCH more so than decent solid state amps). Those that can be operated in triode (screen tied to the plate) do tend to sound better that way, but, I still don't really like them that much.

What two-tube per channel P/P true pentode amps are there in current production?

Pubul57,

Off the top of my head, there is the Audiospace AS-6m (KT88, ultralinear/triode switchable), Primal Luna Prologue 5 integrated amp (KT88 ultralinear), Cayin A-88 integrated amp (KT88 ultralinear/triode switchable), at least one Ayon integrated amp(KT88?), Rogue Atlas (EL-34 ultralinear), Quicksilver Audio Mini-Mite and Mid Mono (EL 34, KT88, KT90, 6550, etc).