|
|
Output classes A and AB
Andre Jute wrote: On Oct 24, 6:57 am, flipper wrote: On Wed, 24 Oct 2007 13:44:13 +0100, Eeyore wrote: Patrick Turner wrote: Eeyore wrote: Multi-grid wrote: maxhifi wrote: Stating that "the amplifier is Class A until XXX watts", really, is telling you how hot the tubes are biased relative to the two extremes of pure class A (full dissapation), and pure class (cut off).- Hide quoted text - Class A has nothing to do with dissipation either. Just because some marketing group noticed that class A means something good, does not make it right either. Just because it seems to make sense is no reason to *******ize the definition. Find some other way to describe it. The definition of Class A is very simple. It requires that the output device(s) never cease conducting under any signal condition. It means slightly more than this because tubes don't cut off as sharply as other devices. I see what you're saying but I do believe that the definition is unchanged. Obviously avoiding any region of significant non-linearity is preferable but that in its own right doesn't change the definition. People are talking semantically past each other with some speaking of the definition of the 'amplifier' class while others are speaking of it's behavior under a restricted set of operating conditions; and it is useful to observe that under an appropriately restricted set of conditions the output tubes conduct 360 degrees as in 'Class A' operation. "Useful", indeed, Flipper. Thanks. But I would go further and say that the signal and dissipation restriction is part of the definition, as you go on to imply: For example, if maximum power, or efficiency, were the primary concern then one might bias more to the 'B' side of the equation while if fidelity were the primary concern one might bias more to the 'A'. Precisely. Both Class A operation and Class B operation are inherent in the nomenclature and definition and their relative importance is clearly intended to be in the designer's discretion. Historically, the original purpose of Class AB was to annihilate the second harmonic which before made up such a very large part of the THD, while still allowing beam tubes and pentodes to give much larger power than available before. The "invention" of Class AB as a hi- fidelity amp is what spurred part of a Olsen's work on perception; before it wasn't known that odd harmonics are proportionately much more disturbing than even harmonics. It seems to me that AB amps with largish parts of their output in Class A is a relatively modern trend, possibly related to ever less-sensitive speakers. He he, No, The largish % of class A power in an "AB" amp is a very very rare trend these days. I cannot think of a single PP amp which has a large % of class A power. ARC and CJ and most others are locked into a stupid desperate war of watts, where they try to extract a maxima of class AB watts and to hell with class A, and reliability. Who is going to buy a class AB amp with 2 x 6550 per channel that puts out a "parsimonious 50 watts" when the guy down the road is making a similar amp which puts out 75 watts? Hell, may as well try for 100 watts. The Chinese meanwhile are dumping absolute crap onto the market which falls apart or burns out sooner rather than later and distracts buyers from spending on ARC and CJ who of course have prices which are 15dB higher than the Chinese amps. To try to snare sales though, the big US brands try to extract more power, its like they can't bare to be seen to make 40 watt amps with a pair of 6550 and then be seen to charge the 15dB more. What the big makers do, ARC, CJ and many others around the globe is just raise the B+, and lower idle current, and then lower the RLa-a and presto, you have a raw PA amp around which you run about 12dB max of GNFB, and call it hi-fi, but your lucky if there is 5 watts of class A before the amp lurches into class AB for the majority of the power ability of 75 watts which BTW will NEVER get used. Amp manufacture has, like 1,001 other products become a marketting exercize, not any thing else. Certianly not an exercize in providing maximal fidelity with a minimum of NFB. One might also observe that's likely why it's called Class AB and not Class insert unique letter. There is in fact a class between Class A and Class AB with a unique description: "Limiting Class A1", which is set up so that the crossover happens when one valve just reaches current cutoff and the other simultaneously reaches zero bias. It makes for an amazingly smooth sound but it is a bitch to set up and keep tuned if you want to keep your circuits simple. I was therefore rather interested in what Patrick said elsewhere in this thread (in more general sense rather than specifically about LImitiing Class A1) about within 10 per cent of conduction angle being imperceptible... Limited class A is the correct term, not limiting, which can mean something else. All PP class AB amps are capable of producing class A power only and never moving into class AB. And I meant what I said about defining the class A. Class A in a PP amp is where the anode current is *always* at leasst above 10% of the idle current. So where you have 50mA of idle current in each of the PP tubes, and an Ia swing of +/- 45mA, each tube is working in PURE class A, and thus the pair are ALSO said to be working in pure class A. Suppose you have a pair of tubes, fixed bias, B+ = +500V, and a swing of +/- 350Vpk on each tube. There is 700Vpk a-a, = 500Vrms. Let us say the Ia change at each anode = + or - 45mApk, then the RL seen at each anode while working in pure class A = 350V / 0.045A = 7.78k. Effectively, the OPT places these loads in series, and you have a load RL a-a = 2 x 7.78k, or 15.6k. With 500vrms across 15.6k, you get 16 watts of pure class A. The idle pda for each tube = 500V x 0.05A = 25 watts, so for 2 tubes its 50 watts. Efficiency maximum = 16 / 50 = 32%. We would be describing approximately the outcome with a pair of pure class A KT88 in triode. So what happens when we use a 5k load on the same amp? The same idle current flows, and the same range of Ia variation 0f +/- 45mA will define the class A **current** swing, ( where the **current** wave THD 5% and mainly all 2H ). So the class A load on each tube = 1/2 x 5k = 2.5k, so the maximum class A V swing at each anode = 2,500ohms x 0.045A pk = 112.5Vpk = 225pk from anode to anode, or 159vrms across 5k, giving 5 watts of class A. But the load value allows for a much larger increase in Ia than the 50mA of maximum decrease in Ia. This also means that once the Ia travels below 10% of the idle value, the gm of the tube cutting off has diminished to such a low value the other tube turning on harder is providing virtually all the Ichange x Vchange across the available load, and is the only device coupled through only 1/2 the OPT primary to the load, so the RL seen by this tube turning on hard has reduced to 1/2 its class A load, or 1/4 of the nominal RL a-a, and in this case its 1.25k. The load is the same as that for a class B amp. Load lines will describe it all far better at my website. But you would find that triodes in AB with Ea = 500V, and load of 5ka-a can make a peak Emin = 220V, so swing = 500-220 = 280V, and at this point peak Ia = 220mA approx on each output tube, and this operation is at the limit of operation without being hindered by grid current. So pk a-a sw = 560V pk = 396vrms, and this means you get 31 watts of class AB power into 5k. 5 watts of pure class A is possible, then the operation ***gradually*** changes from impure class A to AB, where one tube cuts off gradually, and the other tube turning on reaches a peak current several times the idle value. if the current waves in each tube are examined, with a 5k load the waves are seen to have about 5% mainly 2H when making 5 watts into 5k, but at 31 watts, the current THD becomes over 20% with lots of harmonics. Most don't reach our ears because of the complementary action of the two tubes. Pentodes and beam tetrode amps have less gradual change than triodes when passing from class A to class AB, and in fact generate far more dirty sounding "switching" harmonics higher than 2&3H in what is called the "switching zone" or "crossover region". McIntosh became renowned for producing 50 watts from a pair of 6L6 running them in low bias current class AB, and applying a total huge amount of local and global NFB to get the Rout and THD/IMD low. ARC use 16 x 6550 in their reference 600 amps to make 600 watts. This means there is 75 watts coming from each pair of PP output 6550. Not much class A power though. I have just completely re-engineered and re-wired an ARC VT100, made in 1996. It has 4 x 6550 per channel. I found that when there was 8 ohms connected to the 0-8 outlet, you'd get about 125watts of AB power, but very little class A because the tubes were being pushed hard into class AB1. 8 ohms was the load where maximum PO is available. I found that using an 8 ohm load connected to the 0-4ohm outlet gave less maximum PO, about 75 watts but a much higher % of class A po before class AB PO begins. The speakers I have to test this amp for any remaining bugs have 8 ohm woofers for 20Hz to 250Hz, and a pair of 6 ohm SEAS midranges in series for 250Hz to 3.2kHz and a dome 6 ohm SEAS tweeter for above 3.2kHz. The speaker Z is thus well above 8 ohms in the main power band for music, and when cranked loud, the amp's new green-red LEDs which indicate Idc at the cathodes do not change from the correct pale green to red. There is plenty of class A power AND GREAT NATURAL FIDELITY available. Some folks would insist in using "4" ohm speakers with dips in the main power band to 2 ohms, and connect to the 0-4 ohm outlet **which really should be labelled 0-8**. The 0-8 outlet should be re-laelled 0-16 ohms. The use of speakers with Z = average 3 ohms instead of 10 ohms like I have means that the THD/IMD will be approximately 3 times as high for the same PO as with 10ohms, class A power is reduced to near nothing, and the damping factor is hopeless. Class AB PP amps have a varying Rout which is lowest while the amp works in class A, but which doubles when in class AB at extremes of wave points. This unfixed Rout translates to compression, massive 3H, and lots more IMD than while working in class A. Anyway, the quad of 6550 while working in class A with a 10ohm load connected across the mis-labelled 0-4 ohm outlet do sound VERY well. Those wanting a schematic of what have done may ask as I have a .gif available. Its much simpler than the original, and I won't beak ARC rules by handing out free copies of their abominable concoctious junk. The reformed ARC VT100 can make 23Vrms into 8 ohms at its 0-4 ohm outlet, = 66 watts, or about 24Vrms into 10 ohms = 57 watts, which is completely plenty! The 0-8 ohm outlet should ***never be used*** unless you have genuine 16+ ohm speakers, or perhaps ESL where the Z is high below 1kHhz down to 100Hz, and you simply need a large voltage drive. One may always have a high Z midrange ESL across the 0-8 ohm outlet, and a lowZ woofer across the 0-4 ohm outlet. The other config available is the 4-8 connection. This amount of never-spoken-about-section-of-winding is equal to 0.293 x the whole secondary winding turns, which is a theoretical match for 16 ohms, but should have been for 32 ohms. The 4-8 connection gives a match to 1.37 ohms, if we considered the two labelled 4 and 8 as right. In fact, the VT100 would give superlative fidelity into 2.7 ohms if a speaker of 2.7+ ohms was connected between the 8 and 4 output posts. I have set up the VT100 so it has separate bias adjust pots for each tube, and Ia+Ig2 at idle = 39mA measured at the cathode. VT100 is a UL amp with OPT with ct speaker secondary which is optimistically a match for 16 ohms, and each end applies some local CFB to the output tubes. I have Ea at +430V, so Pd at idle for each 6550 = 17 watts only. One can only barely keep a hand on the mesh cover over the tubes even with such a low Pd at idle. I sure don't need to run the 6550 any hotter than they are now. The amp may be played very loud, and music does not unbias the amp badly. The total Pd = 67 watts, and if UL class A efficiency max = 45%, then maximum possible class A from this amp = 30 watts, AND THIS IS PLENTY!!! I have also fixed an ARC Reference One preamp to use with the VT100. One of its EI-6922 had gone noisy. Its complicated circuit has been left alone except for lifting the OV rail from the case and re-connecting via 22 ohms. The VT100 had a true horror for a PSU and after fitting a CLC B+ filter and re-locating earth paths, I finally got hum&noises to be less than 1mV with preamp gain at max with open cd input. The ARC Reference one preamp is a fully balanced design except that the balanced output isn't from a floating secondary of a transformer. It should be, but ain't. Gain max is 12dB, and because there are 8 x 1/2 6922 triodes used in each channel, with an easily possible open loop gain of 1,000 at least, I would suppose that the level of global NFB used around the totally balanced LTP type stages amounts to about 40dB at least. Its bandwidth = 1Hz to 680kHz, and noise is low, and it is a superbly measuring preamp. But I would get similar sound using 2 triodes instead of 8 in a line stage, and have no need for balanced at all. The ARC preamp does seem to be sonically neutral as one would expect from a totally pure class A amp with a shirtload and bootful of NFB. I am amused at those who would say that changeing from say Sovtek 6DJ8/6922 will change the sound from say Siemans NOS 6DJ8. I would always suggest that the high amount of NFB must blind the listener to sonic variations, since any artifact or sonic signature is cancelled away by the NFB action. For a lot more about class A and AB in power amps, go to the educational/diy pages at my website. http://www.turneraudio.com.au Patrick Turner. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
Output classes A and AB
John Byrns wrote: In article .com, Andre Jute wrote: Historically, the original purpose of Class AB was to annihilate the second harmonic which before made up such a very large part of the THD, while still allowing beam tubes and pentodes to give much larger power than available before. Andre, why was Class AB necessary to annihilate the second harmonic, didn't Push Pull operation already annihilate the second harmonic irrespective of the class of operation? Class AB was found to be more efficient and cheaper per watt to produce, and give better sonic performance than the SE amps of 1935. Once anyone needed more than 3 watts, PP was considered, and the bonus was to banish high THD/IMD of the SE amps of the day. The SE radio amps using a lone 6V6 in beam tetrode mode with no NFB were often used, and had mainly 2H, but many OTHER H, and sounded lousy over a whisperish level. HOWEVER, in the misplaced zeal to banish 2H and settle for the clean PP sound, especially when you had a pair of output 2A3, designers would labour away to force an input triode to feed and IST to drive the PP outputs with a two phase secondary. SUN amps are a classic example. Typical primary voltage needed at the 1:2 IST = 50Vrms. So the bloomin input driver tube was making lots of 2H and there was NO net betterment in the sonics except that because the power ceiling was slightly higher. More thoughtful PP amps were designed in the late 40s by leak, Quad, Radford etc, where ALL distortion was considered bad, and where the driver stages were designed to produce far less THD/IMD than the output stage. The Williamson is a classic example. It can be used with 300B in the output, and NFB needn't be used, and THD/IMD will remain low enough, and the 28 watts AB1 will be enough for most folks even now with insensitive speakers. With sensitive speakers of the 50s, the 28 watts of AB triode power was the equivalent of having 112 watts today on average. But very very fine SE amps can be built, and the 2H is low, along with other H, and not much NFB need be used. Patrick Turner. The "invention" of Class AB as a hi- fidelity amp is what spurred part of a Olsen's work on perception; before it wasn't known that odd harmonics are proportionately much more disturbing than even harmonics. It seems to me that AB amps with largish parts of their output in Class A is a relatively modern trend, possibly related to ever less-sensitive speakers. I'm not sure I would agree with that, class AB amps were common even in the days of efficient speakers, I don't see it as "a relatively modern trend", if anything is a modern trend, I would think it is the return to "pure" class A amplifiers on the part of many audiophiles. Regards, John Byrns -- Surf my web pages at, http://fmamradios.com/ |
Output classes A and AB
Multi-grid wrote: On Oct 24, 6:29 pm, Andre Jute wrote: Historically, the original purpose of Class AB was to annihilate the second harmonic which before made up such a very large part of the THD, while still allowing beam tubes and pentodes to give much larger power than available before. . - Show quoted text - Well Andre, you've stepped in it this time...:) AB operation cannot effectively cancell *ANYTHING*. Cancelation of even order harmonics occurs in amps working in class AB during that part of the wave forms which are in class A, ie, the bits either side of the zero crossing. But once each tube moves into cut off, nothing is cancelled. The two non linear current waves in the tubes of the AB pair are summed, and the VOLTAGE total is substantially linear, with a small fraction of the THD of each tube's current wave. Its magic, but it works for most ppl. Odd sums anyway. Depends on the phase of the harmonics relative to fundemental. Its possble to have the 3H of the driver stage acting like compression, and the 3H of the low bias current output stage with 3H acting like expansion, and then there is the transformer 3H etc..... Nothing is general, generic, or simply explained for all occasions, he, he. Not second HD( or the even of any order ) because each phase is biased where the characteristics are changing too rapidly with plate current. This is the rest of the Class A definition that is implied, the finals are biased so that the change in characteristic for the opposing phases approximately cancells( and thus the even HD ). It is why the AB amp can't be labled A while both phases of the finals are conducting( or that that single definition is not enough to describe Class A ). Ah, but all AB amps start off with at least a bit of pure class A before beginning to work in AB. Even transistor amps with miniscule bias currents, and acting very close to pure class B amps. There is always some little bit of class A power where the I swing +/- is less than the idle current. That's class A. Its not the most wonderful class A, but its still class A. The x power in A, and XXX power in AB is just serving notice that marketing had its way with the ad copy. AB was an obvious means of minimizing cross-over distortion and maximizing power. Indeed it was, and still is. It works just as well for directly heated triodes with no NFB as it does for pentodes like the KT88 running a lot of it...:) cheers, Douglas Purists don't like impurity like the Pope don't read Penthouse. Class AB triodes are NOT much used because the 3H is considerable, and for a given % of 3H, its 9/4 times worse than the same % of 2H. Some say its 27/8 times worse. But its never the THD itself that bothers, because music is mainly all harmonic which are related except for some notes which ain't related and the transients; drum beats etcs. Its the resulting IMD that really grates, and the sum and difference frequencies produced when a zillion F are in the amp at the same time dynamically adds up to unwanted grunge as background noise better abandoned if possible, except abandoning 2H, 3H etc is not so easy as dropping off an unwanted child at the orphanage. You gotta cancel it out, not so easy, or feedback it, or prevent it happening in the first place by using all class A to cover the whole dynamic range. So many purists will go for an SE triode operating right in the middle of its most linear curve, then use only a small part of the curve, which is nearly straight, so the music sounds fine. Horns and SET were made for each other. I would say that the IMD products produced by an amp with low 2H is probably nowhere near as objectionable than the same amount of IMD products made by a PP amp. A PP amp which is into AB transition during average levels isn't any better sounding than the SET... Patrick Turner. |
Output classes A and AB
Multi-grid wrote: That's a bunch of nonsense and drivel, if the cancellation were simply the result of "the finals are biased so that the change in characteristic for the opposing phases approximately cancels", then the amount of odd order distortion could also be changed by PP connection. The fact is that the even order cancellation is not dependent on the biasing of the tubes, beyond the requirement that the two tubes are identical and are both biased the same. So then John, when one is cut off, what is cancelling? Look at the current wave in each tube. In AB or B the current wave has SEVERE distortion spectra. But the voltage across the WHOLE PP OPT primary is largely free of the horrendous THD in each tubes current wave. So its THD reduction by complementary action. Its like two men sawing a log with a long bush saw, with one man at each end either pulling or pushing the saw all the way on each stroke. This is class A, and the differences in applied forces by each man tend to cancel and a linear saw action results. In class AB, each man pulls the saw about 1/2 way across the stroke then lets go, and the other guy grabs his saw handle and pulls the saw back the other way. Each man only mainly pulls the saw in turn, and applied force is jerky, and frankly, a difficult way to work; the Union will be down soon to have a go at the boss who told the men to saw the log that way. But in electronics, we can switch things on and off with absolute ease, and there is no Triode Union to make a boss's life a misery. and on this: "The fact is that the even order cancellation is not dependent on the biasing of the tubes, beyond the requirement that the two tubes are identical and are both biased the same. And how are two tubes 'the same' when one is in the traditional AB bias piont? Further, as one starts cutting off( not far from its idle point ), how is this remotely like what is going on in the other phase? I'll give you a hint.....nah, you'll get it eventually. Do try and keep your answer out of the Nonsense&Drivel category. cheers, Douglas We all get it in the end.... Check your High School Trigonometry book to understand why the even harmonics cancel while the odd ones don't, it's a simple bit of math that doesn't depend on bias, only that the two sides of the PP circuit are identical. No kidding. Where would one get these magical identical tubes? Getting a class A amp to cancel its 2nd HD is not a trivial exercise. It mainly IS a trivial exercize and was invented about 2 days after the first triode was made. Any two tubes are never identical, but are often within 90% equal to each other. Therfore the 2H cancelation in pure class A PP is so substantial that 90% of the 2H of either tube is cancelled away leaving far far less than if the two tubes were used in SE & parallel for the same class A power. I suggest you try it on your AB amps, just for kicks. cheers, Douglas Trying to make the driver stage make distortions that will cancel the output stage's PP distortions is extremely difficult to achieve, and nobody sets out to do it. I tried it, I failed. Trouble is that PP output tube distortions change in amplitude and perhaps phase and spectra with the dynamic changes in loads connected. Speakers are not lile a simple one value resistance. One can manage to more easily exploit 2H cancelling phenomena in SE amps between SE driver and SE output tube. PP amp makers all sensibly try to make the driver linear, and output stage linear, and then not have to apply too much NFB to cancel the little amount of mainly 3H distortion.. Some think this is BS, and build a class B thing with lots of THD/IMD without any NFB. Then they apply lots more NFB than the fella using mainly all class A in his output stage. The unruly class B amp with a shirtload of NFB in the output stage McIntosh, emitter follower etc, and a bootfull of global NFB has become the natural choice of profit hungry amp makers. If you hunger for best music, you must spend more for the inefficiency costs of class A. The staus quo among all amps is 95% AB with rough working AB / near B devices switching with lots of NFB. 5% are either mainly class A, or AB tube amps, or the new PWM amps with whatever correction facilities thay can muster. In 20 years time, all the SS AB amps will have become extinct like dinosaurs. The PWM amps can sound as good as generic low bias SS amps and be far more efficient and far cheaper and lighter/smaller. Tubes will remain if the Greenhouse Police are not too zealous, and they will hold sway amoung a small minority of listeners actually willing to put their money on the sound, rather like there will always be someone who likes sailing yachts around the bay, rather than drive around the bay in a stink boat. Patrick Turner. |
Output classes A and AB
Eeyore wrote: Andre Jute wrote: Poopie Stevenson aka Eeyore wrote: John Byrns wrote: Andre Jute wrote: Historically, the original purpose of Class AB was to annihilate the second harmonic which before made up such a very large part of the THD, while still allowing beam tubes and pentodes to give much larger power than available before. Andre, why was Class AB necessary to annihilate the second harmonic, didn't Push Pull operation already annihilate the second harmonic irrespective of the class of operation? Yes, you are right. It does. The SOLE purpose of AB is to produce larger power outputs using the same tubes. Now this poor dumb slow learner Poopie Stevenson, an embarrassment to everyone who comes into contact with him, claims to know more than the RDH4! Yo, Poopie, in the RDH4 (Newnes 1997) on p 545 we find this nugget of authoritative information by F. Langford-Smith B.Sc. B.E. himself: "Class AB operation indicates overbiased conditions, and is used only in push-pull to balance out the even harmonics." It's the push-pull that cancels those harmonics YOU UTTER CRETIN ! A Class **A** push-pull output stage will do that too. AB operation has NOTHING to do with distortion cancellation. But it does. There is partial cancelation of 2H currents up until cut off in one device, and in each wave of signal voltage. However, despite the very non linear **currents** in each PP device when in AB, the net voltage & current when summed in the load is substantially linear. The equivalent circuit of a PP pair is that of two non linear current generators in series with each end of the RLa-a. The summed current in Rla-a is linear. But you can also have PP action and class B where the devices are in parallel and working on a common load. There us more than one way to set up devices in PP... Patrick Turner. Your problem is that you don't understand what you'r reading so you quote out of context as a result of your utter IGNORANCE. Graham |
Output classes A and AB
Patrick Turner wrote: Eeyore wrote: Andre Jute wrote: Poopie Stevenson aka Eeyore wrote: John Byrns wrote: Andre Jute wrote: Historically, the original purpose of Class AB was to annihilate the second harmonic which before made up such a very large part of the THD, while still allowing beam tubes and pentodes to give much larger power than available before. Andre, why was Class AB necessary to annihilate the second harmonic, didn't Push Pull operation already annihilate the second harmonic irrespective of the class of operation? Yes, you are right. It does. The SOLE purpose of AB is to produce larger power outputs using the same tubes. Now this poor dumb slow learner Poopie Stevenson, an embarrassment to everyone who comes into contact with him, claims to know more than the RDH4! Yo, Poopie, in the RDH4 (Newnes 1997) on p 545 we find this nugget of authoritative information by F. Langford-Smith B.Sc. B.E. himself: "Class AB operation indicates overbiased conditions, and is used only in push-pull to balance out the even harmonics." It's the push-pull that cancels those harmonics YOU UTTER CRETIN ! A Class **A** push-pull output stage will do that too. AB operation has NOTHING to do with distortion cancellation. But it does. ONLY because AB working is by design push-pull. The same thing happens in long-tailed pairs. the distortion cancellation is NOTHING whatever to do with AB operation. Graham |
Output classes A and AB
In article ,
Patrick Turner wrote: Cancelation of even order harmonics occurs in amps working in class AB during that part of the wave forms which are in class A, ie, the bits either side of the zero crossing. But once each tube moves into cut off, nothing is cancelled. Patrick, I'm surprised to hear you say this. What are you trying to tell us, that the even order harmonics are only cancelled during those parts of the cycle when both tubes are conducting, but that the even order distortion components reappear during those parts of the cycle when only one tube is conducting? If you actually believe that you should go back to the books and study the theory of harmonic distortion more carefully. I hope you didn't get this notion from the RDH4, I haven't read the RDH4's harmonic distortion explanation, but if this is what it says I have just lost any respect I had for the book. In a perfectly balanced PP amplifier the even order harmonic distortion is completely cancelled even when the tubes are cut off for parts of the cycle. It sounds like you have become one of Multi-grid's sock-puppets. The two non linear current waves in the tubes of the AB pair are summed, and the VOLTAGE total is substantially linear, with a small fraction of the THD of each tube's current wave. And that small fraction is very small indeed, approaching zero to be precise, for the even harmonics. Its magic, but it works for most ppl. It's not magic, it's just math. Regards, John Byrns -- Surf my web pages at, http://fmamradios.com/ |
Output classes A and AB
John Byrns wrote: Patrick Turner wrote: Cancelation of even order harmonics occurs in amps working in class AB during that part of the wave forms which are in class A, ie, the bits either side of the zero crossing. But once each tube moves into cut off, nothing is cancelled. Patrick, I'm surprised to hear you say this. What are you trying to tell us, that the even order harmonics are only cancelled during those parts of the cycle when both tubes are conducting, but that the even order distortion components reappear during those parts of the cycle when only one tube is conducting? If you actually believe that you should go back to the books and study the theory of harmonic distortion more carefully. I hope you didn't get this notion from the RDH4, I haven't read the RDH4's harmonic distortion explanation, but if this is what it says I have just lost any respect I had for the book. In a perfectly balanced PP amplifier the even order harmonic distortion is completely cancelled even when the tubes are cut off for parts of the cycle. I'd love to know how that happens. There's no cancellation of ANYTHING once one side has ceased conducting ! Graham |
Output classes A and AB
On Thu, 25 Oct 2007 15:40:55 +0100, Eeyore
wrote: John Byrns wrote: Patrick Turner wrote: Cancelation of even order harmonics occurs in amps working in class AB during that part of the wave forms which are in class A, ie, the bits either side of the zero crossing. But once each tube moves into cut off, nothing is cancelled. Patrick, I'm surprised to hear you say this. What are you trying to tell us, that the even order harmonics are only cancelled during those parts of the cycle when both tubes are conducting, but that the even order distortion components reappear during those parts of the cycle when only one tube is conducting? If you actually believe that you should go back to the books and study the theory of harmonic distortion more carefully. I hope you didn't get this notion from the RDH4, I haven't read the RDH4's harmonic distortion explanation, but if this is what it says I have just lost any respect I had for the book. In a perfectly balanced PP amplifier the even order harmonic distortion is completely cancelled even when the tubes are cut off for parts of the cycle. I'd love to know how that happens. There's no cancellation of ANYTHING once one side has ceased conducting ! Graham Because if you add an even harmonic to a signal, you have to make it asymmetrical. You always get a peak coinciding with a trough on one half cycle, followed by a peak coinciding with a peak on the next. If you modify the signal to remove any asymmetry, you must by definition remove the even harmonics. d -- Pearce Consulting http://www.pearce.uk.com |
| All times are GMT. The time now is 10:32 AM. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
SEO by vBSEO 3.0.0
Copyright ©2004-2006 AudioBanter.co.uk