Sonic atributes ultralinear vs tetrode vs pentode
The document also has comments about OTLs that I don't agree with, as it comes from the 1950s. But the UL stuff it has in it, while obviously biased (if you will pardon the pun) is representative of the general technical/engineering understanding of ultralinear operation- its easy enough to find plenty more articles like this one. I just clicked on one of the higher-page hits on Google. I can find some of the stuff on the UL curves and that bow I mentioned if you are interested. Full credit and abundant praise must be given to the persistent courage of the Japanese in regard to the triode tube amplification renaissance. Many of us, AtmaSphere included, reap the benefits. For the most part the Japanese seemed to justify our stance on triodes, which we have used since the late 1970s. Back then, people thought triodes were 'quaint' (I suppose many still do). These days they have a lot more acceptance. |
Atmasphere, it's interesting you'd recommend a document that makes as many statements against so much of what you stand for as it does. For example, UL operation clearly more linear than triode versus your statement made in the post previous to providing the link, "How it misses triode linearity is that a triode will typically have a nearly straight portion of its curve where a UL will always have a bit of a 'bow' in its curve." Figures 3a and 3c would show you have things exactly backwards. At any rate, please correct me if I'm mistaken, but this is a commercial White Paper (IOW, as we all know, advertisement), put forth by the company that held the patent on Ultra Linear operation, and sought to make money on the invention via licensing fees paid for by those who wish to employ the topology. In no way do I mean to imply that's not a decent and honest path toward generating income. But do not lose perspective on the intent, which the literature is completely up-front with, "The information offered here will give the reader a positive picture of Ultra-Linear operation, and assist him in choosing an amplifier containing the genuine Ultra-Linear circuit." So, going in, we're not dealing with anything on the level of a peer-reviewed industry position paper. Thus, that UL looks as attractive as it does against both triode and tetrode/pentode should not surprise. A discerning reader should feel the obligation to not accept things strictly as presented in the paper. Indeed, there are issues that immediately jump out at us: 1) The wpc data listed in Table 1 lie far out of the realm of what 60 years of experience with UL have proved possible, and provide insufficient background on how such numbers were generated 2) Returning to Figure 3a reveals an overall sloppiness of presentation; the triode schematic shows the screen grids (which obviously do not exist on true triode tubes, though the text of Figure 3b does reference "the same tetrode tube"), connected to the plates (yes, triode operation for a tetrode/pentode vacuum tube), with the accompanying curve traces labeled (true triode) 2A3/6B4 output tubes 3) Forgetting #2 for a moment, the text for Figure 3a, 3b, and 3c mentions tetrodes exclusively. No can say if that was intentional or an oversight; again, sloppy presentation. But obviously, if the inherent deficiencies of tetrodes did not exist, there would be no need to develop the pentode Through such advertisements and the pressure to keep up with the competition, as I previously mentioned, designers were literally forced in to producing UL amplifiers regardless of whether they felt it the right decision or not. Obviously, the latter part of that implies many engineers (and I'm incredibly lucky and thankful to have met a bunch over the past several years) understood things were not as convincing as we were led to believe. Regardless of truth or fiction the UL "standard" for hifi tube amplifier operation represented, triode (and tetrode/pentode) operation fell completely out of favor. In the age where the race for watts was preeminent, this hit triode all the more hard than tetrode/pentode, which could at least stand on the typical data of higher wattage and survive outside of the world of hifi. After all, what good would an obviously antiquated amplifier design that makes less than half the power and generally double or triple the distortion be to anyone? Full credit and abundant praise must be given to the persistent courage of the Japanese in regard to the triode tube amplification renaissance. Many of us, AtmaSphere included, reap the benefits. Perhaps, tube amplifier designers will also reconsider tetrode/pentode operation in the same vein. Again, I am an obvious fan of UL, but the path to the ultimate can only come out of the position of curiosity, honesty, and impartiality. |
Pete Millett has a nice website with a lot of tube-related information. Table 1 at the end of the document at this link (a quick read) tells the story: http://www.pmillett.com/file_downloads/ultralinear.pdf |
The rub comes when the statement is made that we can meet that measure AND "approach triode linearity" at the same time. Impossible. Possible: "approach triode linearity" = linearity curve lacking the 'kink' of pentodes, looking far more like a triode curve than that of a tetrode or pentode. How it misses triode linearity is that a triode will typically have a nearly straight portion of its curve where a UL will always have a bit of a 'bow' in its curve. But at the same time we have power output that is easily double that of a triode circuit. Its a nice compromise... Of course, if you don't want compromise you will have to seek an amp with real triodes. |
Al, as always, I appreciate your perspective. Thank you. I do not consider 0 wpc an endpoint because the only time a tube amplifier, regardless of triode, UL, or pentode connection, would (practically) meet that figure is when it's shut off. And, of course, at that point, all three would be at 0 wpc. No, the lowest we're going to see in this example is the triode's 20 wpc, so that figure represents our lower boundary. At any rate, I do not disagree whatsoever that a UL connected amplifier can make "most" of the power of pentode - "most" adhering to your strict (and, I believe accurate) interpretation of the word. In fact, it can go far beyond that, and achieve the "great majority" and "nearly all" meanings as well. That's no trouble at all; it's been done countless times. The rub comes when the statement is made that we can meet that measure AND "approach triode linearity" at the same time. Impossible. It's akin to both ends of a playground see saw being up at the same time. Obviously, one side has to be up, the other one down, or both can be level with one another. In other words, as I've stressed, there is no free lunch. To believe anything else is fantasy. Of course, in the strictest sense, the term "approach" can be taken to mean, even to the slightest degree, and by that I mean, 0.000001% of the way, let alone half the way there. However, in the context it's been presented, I've taken this as "most" (IOW > 50% of the distance between the two endpoints). Perhaps, that is wrong, and I've thus far happily waited to be corrected there. But I believe we need to be extraordinarily careful not to present a technology proposed in 1951, and adopted as the de facto standard not longer after, as possessing any unexpected or even mystical advantages - enumerated in the preceding paragraph or otherwise. When it comes to UL, most tube amplifier manufacturers of the modern era have been there and done that. Indeed, even tube amplifier designers who didn't believe in it were literally forced (in such entities, engineers and scientists work for the business area, not the other way around) to build such designs by the sales and marketing folks of the time. Surprising as it may be in this thread, you might tell from my correspondence with Mike that I'm not at all opposed to UL operation. In fact, of the eight commercial high-end audio tube amplifiers I own, all but one feature UL. And that last one does not only because it uses the 6AS7G triode tube. As I said, UL has ruled the day in high-end audio since the 1950s. If I were to raise any criticism, it would have nothing at all to do with that which has thus far been discussed here in this thread; power and linearity. But I don't believe we're ready for such conversation until we level set what UL is and is not. |
A quick search reveals that dictionaries are inconsistent in defining "most." The relevant definition that is indicated in many of them is "a majority of," or some equivalent words. Some, however, indicate "a great majority" or "nearly all." Even if we grant the contention by Joe (Trelja) that "0 is not one of the endpoints," which is debatable IMO, but if we accept the "majority of" definition, Ralph's statement comes within an infinitesimally small increment of being correct (for example, 40.000000001 encompasses a majority of the difference between 20 and 60). In which case claiming that Ralph is wrong because 40 does not encompass a majority of that difference would seem to be splitting hairs, as he indicated. On the other hand, using the "great majority of" or "nearly all" definitions, Ralph's statement would be incorrect, certainly if the lower endpoint is considered to be 20, and probably even if 0 is considered to be that endpoint. The bottom line, IMO: It's listener dependent :-) Best regards, -- Al |
Atmasphere, regarding the triode/UL/pentode (20/40/60 wpc) relationship, you've now repeated what I've put forth, not once, but multiple times. This calls to mind the exchanges of Bugs Bunny and Yosemite Sam. That might be funny except that you still have yet to get your head around the fact that the number 40 lies halfway (and not "most" or 2/3 of the way) between 20 and 60. In your words, "We're not talking rocket science here." Agreed, it's only fourth grade math. The good (or bad) thing about this is that no matter how many times you get it wrong, the correct answer never changes. Until you understand how the number 40 relates to 20 and 60, your understanding of how ultralinear relates to triode and pentode operation will likewise forever remain fatally flawed. |
Trelja, IM(H)O, you need to step back a couple of paces and look consider how most audiophiles would see this: The fact that a triode makes 20-30% of pentode power means to most anyone that if you have a pentode amp of 60 watts, if it was then wired in triode, it would make maybe 20 watts. Similarly, if I say that UL makes most of the power of pentode, while approaching triode linearity, its not going to be skin off of anyone's back to find out that the result is about 40 watts, which is 'most' of 60 watts, since 40 watts is more than half of the power. We're not talking rocket science here; its pretty common to trade off between power and linearity. |
Trelja, choosing the midpoint on a scale may be simple math, but it's not simple dimensional analysis. Since wpc is a ratio scale with a meaningful zero, the proportions that Ralph referred to apply in this case. If this is confusing to you, you might ask someone who has had a beginning statistics course. Audiofeil, thanks for the careful editing of my comment. You may want to take more credit for your creative work in the future. |
>>12-18-11: Mcphersn I'm enough of a boob to be ignorant of whatever has pulled your chain, but I don't think that you are making much of a contribution by making it unpleasant for Ralph to respond to this forum.<< I think Ralph does a fine job promoting his ideas and products in the forums. Trelja won't scare him away. IMO |
Thank you, Mike. Jolidas are fixed (adjustable) bias. Some of the Cary offerings are cathode bias like the AES Superamp, and are considered to fall more on the lush and full-bodied side of things; some of their other products can be more forward. Mcpherson, "would you try to argue that 10000 wpc does not retain "most" of the power of 10001 vs. 9999? After all, 10000 is exactly halfway between 9999 and 10001." I'm not really sure how we do such a good job at falling so flat on our faces here. You've answered the question in stating that 10000 lies exactly between the two other values provided. However, in the attempt at wit, you also seem to suffer the failure to understand the lower endpoint you chose as having the same relevance as the upper. Again, this is simple mathematics While the two endpoints of triode and pentode are distinct from one another in an obvious sonic way, those you chose in this example would be impossible to distinguish from one another by anything outside of a machine. In practical matters, the question is more of perception, and our ability (or inability) to resolve the differences, but the fact remains that in both cases, we are dealing with the midpoint. |
"What first needs to be recognized is that 0 is not one of the endpoints for a comparison between triode (20 wpc for the example at hand) and pentode (60 wpc). UL (40 wpc) is exactly midway (IOW, 50% not 66.67%) between the two. As such, the word "most" can never be used in such a context." Speaking of "fatally flawed" analysis, would you try to argue that 10000 wpc does not retain "most" of the power of 10001 vs. 9999? After all, 10000 is exactly halfway between 9999 and 10001. I've only been following this dialogue between you and Ralph for a while, but it seems to me that your zeal to show him up outweighs your interest in contributing to this forum. I could just dismiss you as a troll, but you actually seem to know quite a bit when you are not trying to bait him. I'm enough of a noob to be ignorant of whatever has pulled your chain, but I don't think that you are making much of a contribution by making it unpleasant for Ralph to respond to this forum. |
Trelja, Thanks for the responses'. Never having opportunity to hear a Pentode amp I feared that might be how it would sound. Definitely not what I am looking for. Honestly I have heard a Eastern Electric 520 amp that does Pentode/UL. I did not like that amp in Pentode, so I listened to it mostly in UL which did not sound like UL in other amps that I have heard, so I did not really want to take that experience as "true to like". Jolida,Cary, and other amps of that ilk did not sound like that in UL, they sounded "Hard" Thanks Mike |
Mike, so often, it's the loudspeakers that are the culprit in an amplifier coming across as hard. However, that's an argument for another time and place, so let's continue to focus on the amplifier. In today's world, there are scant few examples of true tetrode/pentode amplification. You have to look long and hard to find them. What makes things all the more confusing is so many amps labeled "pentode" are actually ultralinear. As I said, UL rules the day, and has since the 1950s. That said, given the case you laid out in initiating this thread, I think you would would find more happiness with UL than pentode operation, the latter sounds more forward, open, and to use your term, hard. Pentode would steer you more in the direction you're trying to get away from. So, I think UL might be getting a bad rap if you think it sounds hard. Here are three characteristics that can "soften" an amplifier: 1) Triode operation 2) Cathode (or self), as opposed to fixed, bias. To put this in a more understandable way, though it doesn't hold up all the time, look for an amp where the tubes are "plug and play", and need no adjustment of bias on the part of the user. The label "fixed bias" is counter-intuitive to most folks 3) Tube rectification I wouldn't exclude a product based on having having only one or two of these features, as I can name some amps with only one of these three that sound quite warm and lush. But again, generally, these three take things toward where you mentioned wanting to end up. |
http://forum.audiogon.com/cgi-bin/fr.pl?aamps&1323578084&openflup&4&4#4 Despite the distraction, this still holds true. |
Atmasphere, "If we take 60 watts as being 100% of power, 40 watts is then 2/3 of that, or 'most of the power' as I stated above. I don't see that being significantly different from your position." Your understanding is fatally flawed. I'm not sure whether the issue lies on the side of mathematics, logic, or language. This is not a question of splitting hairs, but rather a fundamental difference in statistical analysis. What first needs to be recognized is that 0 is not one of the endpoints for a comparison between triode (20 wpc for the example at hand) and pentode (60 wpc). UL (40 wpc) is exactly midway (IOW, 50% not 66.67%) between the two. As such, the word "most" can never be used in such a context. Again, how we got into this circle was your claim, "Ultralinear allows the tubes to approach triode linearity, while retaining most of the power that Tetrodes (or Pentodes) otherwise make." If you simply took the position that you are personally fond of this particular compromise, I'd understand and accept it. However, steadfastly refusing to concede the point regarding the word "most" in your argument is analogous to mixing black and white paint (or, red and yellow M&Ms?) in a 50 - 50 ratio, then attempting to claim it has more than 50% of one of characteristics of black (or red), while at the same time possessing more than 50% of one of the characteristics of white (or yellow). I'm not sure how else I can make the point, but a great many of the classical scientific laws dating back to the ancient Greeks have proven you can't get more out of something in this world than you put into it. |
OK- here's my text: Ultralinear allows the tubes to approach triode linearity, while retaining most of the power that Tetrodes (or Pentodes) otherwise make. here's your assertion: If you compare an amplifier running in the three modes, if triode put out 20 wpc, UL would do 40, and pentode 60. Those figures work out in the absolute linear, proportional, and expected manner If we take 60 watts as being 100% of power, 40 watts is then 2/3 of that, or 'most of the power' as I stated above. I don't see that being significantly different from your position. Some hairs are in danger of getting really thin here. |
Atmasphere, you've done nothing to address my refutation of your quote regarding Ultralinear operation giving "most of" what both triode and pentode offer. As I initially said, the subject at hand represents yet another case of simply splitting the difference between two endpoints, as opposed to getting more out of something than what you put in. The world just doesn't work any way other than that. If you compare an amplifier running in the three modes, if triode put out 20 wpc, UL would do 40, and pentode 60. Those figures work out in the absolute linear, proportional, and expected manner. In other words, again, your assertion, "while retaining most of the power that Tetrodes (or Pentodes) otherwise make" is untrue. "Ultralinear allows the tubes to approach triode linearity" would not be an incorrect statement in and of itself. But in the context of your position inferring it goes beyond the proportional and expected distance, it is. "Rule the day" means sales/volume... |
Trelja, obviously we see things very differently :) I've done a lot of work over the years with ultralinear amps (quite a lot of those were Dynaco amps)! There is no way I would say they rule the day, I would put ultralinear as a distant second to real triodes. With regards to your prior paragraphs, I think we are pretty much saying the same thing. One way I use to deal with the linearity issues that allow me to get the amp to work better is to cross-couple the cathode circuits around the output of the transformer. In this way I get greater linearity, with lower output impedance, and that is what has allowed me to run zero feedback to the driver section. IMO that is how you get the amplifier to be low distortion and listenable at the same time; if you use feedback the amplifier will pick up a harder edge. |
Atmasphere, "Ultralinear allows the tubes to approach triode linearity, while retaining most of the power that Tetrodes (or Pentodes) otherwise make." Unfortunately, in audio like life, there is no free lunch. In reality, ultralinear splits the difference between triode and tetrode/pentode, as opposed to offering "most" of both. Historically, folks have described it as offering the best of both worlds. As it's the textbook definition of a compromise, I'm currently of the mind that like so many other compromises, it's also offering the worst of both worlds. At any rate, how many high-end audio amplifiers actually feature true tetrode/pentode operation? From what I see, the number these days is very small. Ultralinear rules the day (past 50 years?), with triode coming in a distant second. |
Atmasphere, I really appreciate someone with your knowledge and experience answering the thread. However, the sound characteristics of Ultra-linear you describe, I unfortunately have never heard. The sound always seems compressed and hard, smaller sound stage and just not very appealling when compared to the triode mode that most amps offer in tandem with the ultra-linear mode. Thanks |
You won't hear so much difference between Tetrode and Pentode, but you will hear a difference between Ultralinear and the others. Ultralinear allows the tubes to approach triode linearity, while retaining most of the power that Tetrodes (or Pentodes) otherwise make. So Ultralinear circuits make less distortion and tend to have a lower output impedance. If the designer is crafty and takes advantage of this, the result will be an amplifier that is smoother and more detailed, as removal of distortion usually has that effect. Ultralinear amps need less loop feedback in order to work, and in some cases don't need any at all. Loop feedback adds brightness to almost any amplifier, so if you can run without it you can have a more relaxed presentation. However, not all designers will do that, so YMMV. |
I know what triode mode is, its a pentode or beam tetrode "strapped" in triode. There are a lot of amps that use pentodes or beam tetrodes that operate in different modes. I know what an amp "strapped in triode sounds like and Ultralinear, I have an amp that functions in tetrode mode only. Maybe I just should have asked about a amp that operates in Pentode, but I thought a comparison would be more helpful and easier to put into words. Thanks Mike |
Do you mean "triode" vs. pentode? It would help to have more info in re: what speakers, size of room etc. Ultralinear uses taps on the output transformer to help broaden the frequency response (bandwidth). Triode/ Pentode refer to the number of elements inside the tube that perform the amplification. |
