solid state vs tubes
MrT, there may be another way to do this. The Quad has a low impedance at high frequencies and a lot of transistor amps will tend to make more power due to that, part of the reason why they sound bright on the ESLs and with less bass impact than most tube amps. OTOH, your Maggies have a flat impedance curve, but also need some power, but delivered into a 4 ohm load. With any SS amp made, the more current you demand of the amp, the more you get non-linearities caused by capacitive elements in the junctions of the output devices. IOW even though they may drive 4 ohms just fine, most transistor amps will sound better on 8 or 16 all other things being equal. So there is a single solution that might be worth pursuing, the ZERO. If you use a transistor amp, this may take away some of the glare, although you will also loose power. But with a tube amp it may not only have the amp sounding more relaxed, but may give you greater power (although only slightly so) as well. Some of the driving issues with tubes and Maggies are actually the speaker cables. The ZERO will allow the cable and the amp to see 16 ohms, and then a very short set of cables can be used on the 4 ohms side, allowing you to keep your cable losses to a minimum. Something to think about. |
Magfan, here's a nice nutshell explanation that touches on this subject. That there are capacitive elements of semiconductors should not come as a surprise to anyone- most tuners these days do not rely on variable capacitors to tune the station, instead they use variactor diodes, which take advantage of this capacitive element. Literally by changing the voltage (and thus forward current) on the device the capacitance changes and is useful as a tuning cap in the RF front end of the tuner. Here's the link for the MOSFET capacitances (you can find many others on this site): http://www.diyaudio.com/forums/tubes-valves/155106-ss-followers-drive-power-tubes-3.html#post2316318 |
Aren't 4 ohm loads more problematic for tube amps in general? Isn't that part of teh [sic] reason why you advocate 8 ohm or higher speakers for use with your tube amps? This is true. The fact of the matter is that **all** amplifier technologies sound better (better= smoother, greater detail, often more authority, and in tubes less lower-ordered distortions that contribute to the 'tube sound') when driving higher impedances. IME 8 ohms is OK (better than 4) but where the 'magic' happens is with 16 ohms. So I usually put it this way: If *sound quality* is your goal, 16 ohms will sound better than 4 or 8 ohms, all other things being equal. If **sound pressure** is your goal, then 4 ohms (or possibly less) is preferred if you have a transistor amp that can handle the load. The point is that in high end audio, we are more concerned about getting it to sound real rather than anything else. Quite often (as in Unsound's case per his last post), things might be sounding *good enough* so that one is satisfied, but that is rarely the same as saying that the sound is as good as it could be. Its whether one is interested in making that step in the interest of better sound. I say this because quite often a speaker design could be a 16 ohm design or 8, rather than 4, without seriously altering the design. It does not have to be more expensive. Its really up to the speaker manufacturers, but I think most of them don't want to raise the impedance because they fear loss of 'voltage sensitivity'. Have you noticed that there is speaker efficiency, 1 watt/1 meter, and there is also sensitivity, 2.83 volts/1 meter? The fact of the matter is 16 ohm speakers might be less sensitive, but they will not be less efficient. If this seems hard to get your brain around, consider the idea that the sensitivity spec was created to cover for the fact that many speakers are not as efficient as they are made out to be. |
Unsound, Paul Speltz who makes the ZEROs has a letter from Steve McCormick, in which Steve asserts that his amps, which have no problem with 4 ohms at all, sound better driving 4 ohms through the ZEROs. Upon hearing about this (which was about 2 years ago) the next time I was at CES I asked about this subject with several of my friends in the industry who make transistor amps. I was surprised that there was a consensus amongst them, that I can paraphrase (this one stated almost verbatim from Edge's Steven Norber) 'just because it is comfortable driving four ohms does not mean it is sounding its best', this specifically in relation to 16 ohms. One thing that you may not be considering is the role of the speaker cables, which is critical with 4 ohms but not so at 16 ohms. For example if you do the math, you find that no matter how high the damping factor of the amp, there are no speaker cables that will allow the amp to express anything more than 250 into 4 ohms. In short, in high end audio there is no compelling case for 4 ohms. Its bad for tubes, bad for transistors. How this might relate to MrT's inquiry is that his job would be easier were he to use a higher impedance speaker. BTW, I do concede that in my comments 'all other things being equal' is a serious caveat. They never are. Because of that, it took a long time to figure out how important this issue actually is. Finally, I would like to point out to MrT an issue that must be considered: speakers that are designed to work with transistors may not work with tubes, and vice versa. This is an old conversation, that of equipment matching, but touches on a larger subject: http://www.atma-sphere.com/papers/paradigm_paper2.html |
Unsound, in this case Mr. Norber was telling me, as did the other manufacturers, that indeed their amps do sound better into higher impedances, despite making less power. The Mac autoformers are used to load the transistors at a lower impedance if I recall right, quite the opposite of what we are talking about. I have to tell you, I was quite surprised to find that the ZEROs have a benefit to transistor amps in the same way that they do for tubes, although when it was pointed out that this has to do with the behavior of the output devices when more current is put through them, it makes perfect sense. With regards to speaker cables, it appears to be simple DC resistance. |
MrT, there was a time when planars ruled the roost when it comes to resolution. That time has passed- there are a good number of cone systems now that easily compete with the best planars. However, my comments were more the point that if you had a 16 ohm planar, any amp that you have tried so far would sound better on such a speaker. IMO, one of the most musical transistor amps that is also reasonably priced are the zero-feedback Pass amplifiers, particularly the First-Watt amps. The Ayre is another good choice. So far the best I have heard, better than most tube amps, is the Ridley Audio amplifier. Beyond that, most transistor amps that employ large amounts of feedback are going to sound a lot more like each other than they will like anything else, including tubes. I should point out though that the 3 amps I listed also are known for making heat, as the simple fact of the matter is that class A operation is part of how these amps manage to sound right. In fact, the Ridley employs a heater to raise and regulate the heat of the output devices. So other than occasional tube replacement, IMO you might as well have a tube amp, if quality sound like real music is your goal... |
Minorl, not to put to fine a point on it but the idea of designing an amplifier to pass a signal without distortion is being heavily glossed over in your arguments, which, other than that, I find myself in agreement with. The problem lies with the fact that our ears, regardless of the individual, use a set of common rules that govern our perceptions. Now really, that is not a problem except that the bench tests that are commonly used are for the most part not devised with these perceptual rules in mind. So as a result, a common experience is that an amplifier that measures well does not sound good, because **in the attempt to measure well, the human perceptual rules are being violated**. So it is not true to say that the amplifier that has the 'lowest distortion' will be in fact the amplifier that actually *has* the lowest distortion; the two can be quite different! Often 'higher distortion' amplifiers (as measured on the bench) are lower distortion when subject to the reality of our ears. In short, the bench measurements come off as an example of the Emperor's New Clothes. Our ears OTOH, are the real thing. Now it is a simple fact that tubes more closely obey these perceptual rules than transistors do, that is why it is so hard to find the transistor amps that really sound like music. However, and I point to Nelsen Pass as an excellent example, when you find such an amplifier it will be because the designer is also looking at how our ears work. |
Mapman, I am **not** talking about clipping, although it is true that in clipping, tube amps generally make less odd ordered harmonics than transistors do. What I am talking about is the fact that when you use feedback to control distortion, the price of the feedback is distortion of the loudness cues (5th, 7th and 9th harmonics). Not by much, but as I mentioned before, 1/100th of a percent is audible. It seems crazy to think that the ear is that sensitive, but after all, it has to be sensitive to *something* and this is the mechanism that we use to determine the volume of a sound! So when you see amps with very low THD, its likely that this mechanism has been interfered with. The result: the SS/tubes controversy since SS amps usually use a lot more feedback than tubes do (which can be built with no feedback at all)- IOW tubes more closely mimic the rules of human hearing, and so make much less of these particular harmonics, even though they often make more of the lower harmonics (which the ear finds less objectionable, though they are often credited with the 'tube sound'). Its not so much that I prefer tubes, what I really prefer is amps that don't violate human hearing rules. So that is an amp without feedback, more than it is tube or transistor. Now, without feedback, it might be better to have either a higher impedance in the speaker and/or a flatter impedance curve, but if that is the price to be paid it seems small if it is the difference between something sounding like a stereo and something sounding like real music. |
Soft clipping means that dynamic range is restricted to some extent as I understand. That means loudness cues are reduced relatively, all other things aside, as well, right? What I was talking about is the types of distortions made by tubes and transistors, VS how the ear hears. It is true that the soft clipping of tubes means you will get less of the odd-ordered harmonics, but this has nothing to do with dynamic range. That is a completely different issue! Minorl, I do take exception to one comment you made about "well engineered/designed tube or solid state amps", which is the issue of the human ear. If the equipment is "well engineered" to look good on paper, how "well engineered" is it to obey the rules of human hearing? In my book, it is thus not well engineered to the task that it was built for, since the bench specs have very little to do with how human hearing rules. Have you seen this link? IMO this is one of the more serious stumbling blocks MrT has to face in his quest, and why I commented earlier on the particular SS amps that I did. http://www.atma-sphere.com/papers/paradigm_paper2.html |
If you perceive the peaks as less loud because that is how people hear, that would seem to infer that the dynamic range is affected, at least as perceived, since the peaks do not seem as loud anymore, just as those nasty loudness cues due to negative feedback that you are focused on make you perceive things as louder as you describe. Mapman, that is true but is not the same as real dynamic range. I don't think it is a good idea to promote a practice of intentionally distorting the loudness cues in order to make something sound louder than it really is. A lot of amps do that though, and is why they sound loud and shouty, as in: "Turn that !@#$%^ down!!" IOW, if your system has the ability to sound loud, the loudness cues are being distorted. It is possible to build a stereo that does not seem to play loudly, so that even 105 or 110 db is comfortable. I'm a fan of being able to play the system to life-like levels, that is not possible if it also messes with the loudness cues! IMO/IME in about 90% of the conversations that audiophiles have about dynamics, the word 'distortion' can safely be substituted without changing the meaning of the conversation. It is the reason I avoid using 'dynamics' except in for this context. |
Mapman, I agree, the problem is this: distortion, in the form of exaggerated 5th, 7th and 9th harmonics (compared to the original sound) **masquerades** as 'dynamics'. A simple way to tell that a system is lacking this distortion is that you can't tell how loud it is, until you find that you are shouting to talk to someone sitting right beside you. Oh, and FWIW, soft clipping does not reduce the loudness cues. Its just more likely to not distort them as much. However clipping in any form is to be avoided, so we are talking about how the loudness cues may or may not get distorted when the system is not clipping at all. |
