Most "high biased" >100w amps are around 5w Class-A, Unless you go Gryphon Antillion which has 3 X switchable bias up to 100W!!! Class-A and is 150w A/B The Class-A heat factor "goes up exponentially" the more Class-B you have. So two amps both with 5w of Class-A each. One is 50w-a/B the other 200w-a/B. The 200watter will be 4 times hotter "at idle" than the 50watter even though they both only have 5w class-A. Hope that explains it?? Cheers George |
I have experimented with different bias currents in my Class A/B Integrated Amp and found that higher values provided significant improvements in sound quality, especially at low volume. Heat buildup becomes a limiting factor. The "Leaving Class A" article by Nelson Pass may be of interest. https://www.passlabs.com/technical_article/leaving-class-a/ |
Is there a general industry standard for the amount of Class A power in Class A/B amplifiers?For instance SimAudio has always touted that they run Class A for the first 5 watts.No. If the amp is only 1/4 or 1/2 watt before it goes into B operation its still considered class AB. Some amps out there have what is marketed as 'enriched A' operation where they have 20% or so of full power in the A region, but that is only marketing; such amps are still considered class AB. There is no such thing as 'A/AB' unless there is a switch to differentiate the two. |
| Post removed |
No. Because there is no standard on what type of output stage to use. The bias current is set to give the best performance for the type of output stage in terms of linearity and distortion, while keeping an eye on the power dissipation (heat sinking is very expensive). There really is no such class as Class AB, it is used to describe the mode of operation when a Class A amp hits a low impedance and one of the pairs of output transistors stops conducting. Almost all amps are Class B, where they are biased just above the transistor Vbe so as to keep the transistors conducting in that dead band from -0.7 to +0.7 volts to avoid crossover distortion. Increasing the bias beyond that point (but less than Class A) may provide benefit by reducing crossover distortion further but you pay a price with gm doubling distortion. How much above that minimum Vbe setting do manufacturers bias is done to maximize performance. For example, a Darlington output stage requires about 100ma of bias and a complimentary feedback pair output stage requires about a tenth of that for best performance. So it is not possible to standardize something like this. |
My Coda CSIB integrated gives you 3 power choices,I choose the lowest power. my integrated 1st watts in pure class A says 18 watts ,it’s closer to 20 ,then 150,300, 600 wpc , and incredible 120 amps short term available for control ,and Huge potted 3kva transformer , and pure class-A preamplifier section made in USA , I know of nothing under $10k that is even remotely close in many areas ,a true High end bargain.stereo times was spot on in their review ,after buying it I could not agree more . |
Some interesting insight. @gs5556"There really is no such class as Class AB, it is used to describe the mode of operation when a Class A amp hits a low impedance and one of the pairs of output transistors stops conducting"..Would this mean that by keeping the load impedance stable & highish it would keep the amplifier operating in Class A longer? |
| Post removed |
Hi Douglas, I will definitely read the review and I keep hearing amazing things about Gan powered amps. I am waiting for a reasonably priced HT multi channel amp with Gan. There are other factors of amp sound, like the tube input stage of my BHK300 amps. Bascomb/PS Audio just couldn’t get anything with a solid state input section to sound as good as a tube input section. One theory of why tubes sound so good (aside from the space between components) is that there are more electrons to flow and you seemed to indicate in the review that Gan has more electrons to flow (if I’m reading it correctly). I don’t plan to ever give up my BHK’s but I could definitely see Gan based devices being in my Arsenal of gear. |
There is no standard. The amount of class A can vary from less than a watt to a couple of watts and sometimes more. Note that the class A rating is usually for a solid state amp at 8 ohms. It's less as the impedance drops. Also note that most class A amps are really very rich class AB amps where the rating is just the class A rating and they produce more power but in AB. As an example PASS makes 30 watt amps that are rated at 30 watts class A but often produce more than twice as many watts in AB so almost all the time they are in A but on peaks they go to AB. As implied previously most class A amps that really are class A into 8 ohms are not as the load impedance drops. The only one I'm reasonably sure was is the old Mark Levinson ML2 which was rated at 25 watts class A and was still class A at 100 watts into 2 ohms. |
2nd there is no standard. In most amps in fact bias is adjustable. What you may not realize is that its simply a trade off of heat vs "class-A-ness". I honestly find very little difference between a truly high biased AB amp and class-A, and i can simply turn the POT. I designed and manufactured a very high bias amp in the mid 1990s - ran class-A about 40% of its rated power (which, for heat reasons and others, was accordingly modest). The prototype remains my "daily driver" today. Sometimes 2 in monoblock mode, but usually I'm swapping things too often to go to that level of complexity. Given that music has a peak:average ratio of about 10, an amp that is 50W and is 5W class-A would only be out of class-A operation a small % of the time, and those times would be major crescendos when subtlety is more of less completely masked anyway. G |
Fascinating thread here. OP didn’t know how loaded the question was. The Pass Labs designer’s write up is excellent. I don’t know where this video is at the moment, but an interview with amp designers had John Curl: the question was, “how much bias do you use?” ”As Much as I can.” - and back to the discussion of the balancing act between too much heat & distortion and SQ/ cost. |
Yeah I got bored over at the other place, much more fun over here, for breaks giving **** to the snake oilers and voodooist like fusers, etc etc. That sort of snake oil stuff is kept under control over there, so there no real dog fight action. But I must admit it’s starting to become chronic here, as too many can see 10000’s% of profit selling re-badged 10cent fuses for >$150+ and having no overheads doing it! Cheers George |
It’s possible. You can get fairly close to full Class A performance with A-B, but may be limited by a slight lack of foundation to the sound. When people say class A is smooth, sweet, or well-rounded, those are subjective opinions. The conduction angle of class A is 360 degrees. Power hungry but with maximum performance. |
There really is no such class as Class AB, it is used to describe the mode of operation when a Class A amp hits a low impedance and one of the pairs of output transistors stops conducting.This statement is false. Class AB is defined as any push-pull amplifier which is class A at lower power (which is done to prevent artifacts at the zero crossing point) and class B at higher power. The *amount* of class A power is undefined. Types of amplifier class: Class A1, A2 and A3 (the latter recently develeped) Class AB (which includes class AB1 and AB2)Class B Class C (only used in RF power amplifiers) Class D (doesn’t refer to digital; ’D’ was simply the next letter- the first commercial class D amps were made in the 1960s) Class E (switching RF amplifier) Class F (harmonically tuned RF amplifier) Class G (modified A or AB amp; IIRC first used by Hitachi) and on and on... |
Might be helfpul to understand how A and B differ, and that transistors are not perfectly linear. It takes a little current before they wake up and can act in a linear fashion. This is what causes the notch distortion during the crossing of 0V. In Class A, both legs start at full on. Since they are balanced this ressults in 0V at the output, and maximum power draw through what are essentially to closed switches. In Class B both are off, again, achieving 0V at the output and minimum power draw through essentially open switches. The bias, or quiescent current, in an AB amplifier keeps both legs on enough to avoid the dreaded notch distortion without consuming 100% of the power in a true class A amplifier. The difference is that eventually as the output voltage shifts far enough away from 0Vthat one leg will shut mostly off, switching to B. When the voltage goes back towards 0V that leg now turns on, with some inherent distortion as they attempt to reach equilibrium. Personally I find the arguments for/against Class A on technical terms alone lack overwhelming evidence. I've heard Class A I liked and Class A I really did not. Same for A/B amplifiers. So while it's fun to look at how engineers have take on the same challenge via a variety of methods I always find my ears better at determining what I am going to want to listen to than amplifier classes. |
Personally I find the arguments for/against Class A on technical terms alone lack overwhelming evidence.@erik_squires The arguments are twofold for which there is a lot of evidence if not proof. The first is that you can bias the device in its most linear region. Given that most amps tube or solid state can't really be given the feedback they really need, this is a useful means of allegedly not having to have so much feedback to build a linear amplifier because distortion is lower to start with. The second issue is that you get more accurate even ordered harmonic cancellation when both output devices are operating in the A region. With the even orders cancelled, there is less overall distortion and so the odd orders tend to be at a lower value (this particular aspect is more useful if the amplifier employs the push-pull/differential operation from input to output). This again is helping for the fact that its likely that insufficient feedback will be applied to the circuit. |
The first is that you can bias the device in its most linear region.This is correct, and why the "better" higher power Class-A’s use up to 24 sometimes 48 even 64 complimentary outputs as in the Threshold Stasis SA/12E monoblocks, so as to only use the "most linear part" of each transistors curve. https://ibb.co/d7BKwzL The SA/12e, being fully class-A from input through to and including output stage, draws its maximum power from the line at idle—about 1000W for the pair.https://www.stereophile.com/content/threshold-stasis-sa12e-power-amplifier-page-2 Cheers George |
Hi @atmasphere Your technical explanations are, as usual, spot on, but since I failed to explain what I meant by by "proof" I am afraid you made a counter argument to a point I did not mean to make. What I meant to say that my own ears have yet to discern a preference of Class A over Class A/B or even D which is even moderately consistent. The behavior of class A circuits vs. A/B in technical terms is more or less well understood, as you show in your rebuttal to my post, but not in the realm of personal experience. To me, there is no proof class A is consistently or even broadly better than any other class of amplifier when each class is well executed. Yes, some stellar Class A amplifiers exist. Do they make me go "wow this is what Class A means?" No too many other middle of the road or poor sounding Class A amps out there. At best, I could be convinced that class A amps are an acquired taste which some develop and then never grow tired of. Fair enough! It's a happy man who finds what makes them happy early in life! Best, Erik |
Of course Class-A is better if made correctly. Once a low bias A/B amp leaves it’s low Class-A bias you get cross-over distortion. "NO" crossover distortion is better to listen to in anyone’s language, it’s one of the worst distortions there is, if it were second harmonic distortion then you could forgive it https://ibb.co/VpFSzSz Do yourself a favor (here’s your proof) and listen to a Gryphon Antillion with switchable bias on the run while listening low, med, high (100w), if you can’t hear the difference then give it away and take up something else. https://gryphon-audio.dk/shop/power-amplifiers/antileon-evo-stereo/ . |
| Post removed |
believing that Class A was the SHIT! ... But I came to find that slew rate & wide band frequency response under punishing conditions is a far better resume You can't state that unless it's the same amp your comparing the different Class-A bias with. You also need to listen to the same amp (Gryphon Antillion) as I stated above with all those parameters you mention, but that is switchable in Class-A biasing low med and high, and then state it doesn't matter much, if it's in high class-A bias compared to low or even medium. Cheers George |
What I meant to say that my own ears have yet to discern a preference of Class A over Class A/B or even D which is even moderately consistent.:) This is an entirely different issue! If the amp has feedback, you'll probably not be able to tell anything about class A or AB. If the amp is zero feedback you will. Most amps that are class A or AB and run feedback simply don't run enough, so you'll get artifacts on that account (usually brightness and harshness caused by distortion). Pretty hard to tell which is which when that issue is overshadowing things. |
Well, @atmasphere read my previous post and responded appropriately while @georgehifi seemed to have ignored the point I was making, which is about audible benefits and sticks to cherry picking statistics which he feels must prove his point. As I’ve mentioned like, forever: a measurement is not audible nor a preference. It’s just a number. To be "better" to me it has to sound better. Otherwise it’s either equivalent or worse. We can argue that 0.01% THD is better than 0.03% THD, given equal harmonic distribution of the distortion, but you can’t prove one is audibly better since I assure you I cannot hear it. And so, I have yet to hear evidence that any amplifier class is inherently superior all by itself. Even @atmasphere starts to get into feedback vs. non-feedback which occludes the entire discussion over amplifier classes. I might induce he’s saying feedback is more important than amplifier class. :) I am going to state, categorically and forever that brand names are more important than amplifier class and that everyone who argues otherwise is a muggle. |
If the amp has feedback, you’ll probably not be able to tell anything about class A or AB. @georgehifi seemed to have ignored the point🤦♂️ NO! this is bad information and definitely not correct I have built many Class-A’s (even a massive water cooled one), I and friends could tell if they were in low bias or high bias, with feedback applied, using global or local, and using matched complimentary outputs. And if you don’t believe me ask Nelson Pass one of the foremost Guru’s of Class-A amplification that neither of you could hold a candle to, he’ll set both of you straight. |
Coda is a very old school Company Threshold Nelson Pass - his engineers when Nelson decided to go Solo became Coda they offer 3 power variants like the excellent CSIB integrated amp the preamp is in pure classA ,Mosfet,and FET the separate Amplifier section the first 18 watts are in pure class-A ,then 150,300,600WPC,2 other power choices the higher in power the lower in pure classA, with largest potted low noise power supply I have seen at under $15k a 3kva transformer and over 120 amps ,that’s enormous power- control on demand and 40 matched Bipolar resistors on the outputs |
