diyAudio Forums Archive > Top > Amplifiers > Solid State Pages: [1] 2 3 4 5 6 7  Incredible quality amplifier by Graham, prepare your ears for it - Click HERE for Original Thread destroyer X I finished assemble and testing, and hearing for 3 hours the last Graham Maynard design...he sent me july, 8. A 100 watts over 4 ohms amplifier, class A to AB, having enormous identities with circuits of our days, and some differences that produce, all together, a very incredible amplifier. I think he will publish very soon..... do not let to assemble this one, and tell me if i am wrong. ahahahaha!...can call me Charlotte if less than 99 percent of guys completely astonished and satisfied, the ones that decide to produce this unit will have enormous satisfaction. Prepare your boots to kick a lot of amplifier.... a genuine S kicker is arriving soon. Yeah!...i already love AKSA 55, but this one have complementary qualities...better to have both. regards, Carlos destroyer X Here is the picture, sorry by the quality, and the amplifier appearance..... my God!..... do not evaluate by appearance, as this amplifier is a Diamond! regards, Carlos Graham Maynard Hi Carlos, I know you've had troubles recently, but you are happy smiling there ! I had already posted the circuit in Lumanauw's thread entitled the many faces of distortion, but here it is again. Carlos tells me he is using a single class-AB 2SA1216/2SC2922 output pair with 2SA1837/2SC4793 drivers. With a 2SC1819 video transistor for the voltage amplifier he had to fit 50pF across its base-collector legs to prevent oscillation. At least the circuit works with transistors other than those shown, though I can touch anywhere in my amplifier with a wetted finger against a screwdriver blade and there is still no oscillation. My main aim with this circuit was to make it have low first cycle distortion/ good transient capabilities by making it have a very low internal inductance. Thus dynamic loudspeaker system generated back-EMF should cause minimum amplifier-loudspeaker interface reaction, with the sound remaining neutral. Imagine the JLH four transistor class-A. Add 'helper' class-AB push-pull drive to incresase power and efficiency. Add a differential input pair for output voltage zero control, but note the 10nF capacitor that retains two stage hf stability. Add a mirror to prevent power-up thump, but note its 220nF capacitor again for two stage hf stability; ie. mirror at audio frequencies, CCS at hf. It runs with constant class-A control, and yet has decent class-AB output powering capabilities. Cheers ........... Graham. destroyer X Sounding great Graham. Thank you very much, i am very happy with this amplifier. Here is the schematic, with the option to change the 2SC1819 by a 2SC4793 or other that belongs to your own preference. Those transistors are very good, and they are used in wonderfull amplifiers. But you can change a lot of things, including the input transistors....yes, i changed for a 300 Mhz unit...but could not feel diferences. My circuit can have defects, or errors...but i guarantee that is sounding very good and working for 5 hours without any problem. regards, Carlos destroyer X The voltages measured referenced with ground are inside square lines. The voltages with arrows are VBEs The supply is simple with 15000 plus 15000uF The supply voltage without load is 37 volts plus and minus, when amplifier adjusted, my supply measured exactly 35 volts plus and minus. I am using 8 ohms speaker....also hearing with Phillips good headphones with 220 ohms in series...wonderfull sound! Power consumption is 550 miliamps in stand by mode. heatsink total of 1600 squared centimeters of air exposed area. Inside my supply there are two transformers, each one of them around 250 to 300 watts.... separated filters and rectifiers, the output is joined together...two supplies in parallell, as voltage is almost identical...have 100N bypass capacitors inside, connected from plus to ground, negative to ground and positive to negative also..... 500 volts poliester units. Sorry, the image is not so good, needed some effort to read voltages there. regards, Carlos destroyer X No good quality. But better than no image to show you. One image have the value of a thousand words. regards, Carlos destroyer X If you like to make your own boards, a hand work, here is some idea. You can sketch with parts drawn.....At same time you can see the other side copper lines. Imagine this board as something made with glass, completely transparent...so, you can see the other side lines, and the parts that will soldered over. I use to solder over the copper, this way i have not to make holes and have not to invert nothing. And sound without holes is good, also sound of parts soldered over copper are good too...no difference. As board will be inside enclosure, and i will not watch it many times, it can be ugly...no problem....but have to sound good. Because pretty board and sound bad...trash box is the place. regards, Carlos destroyer X Just some idea to you. This amplifier is easy to construct, seems confused because the way was drawn to fill one Tina small screen. But in reality it is simple and very standard in the main thopologie, some details are different only. here some vision more close. regards, Carlos destroyer X I adjust it to class AB and the unit works fine....not so good as class A. To adjust to class A have to advance bias trimpot and read more than 110 milivolts over 0.22 ohms resistors. Well, Graham informed that have to adjust 400 miliamps for the standard output, and then increase more 100 miliamps adjusting the extra NPN transistor bias. I could hear....yes HEAR, around 10 Hertz, and full output was obtained at 16 Hertz, related the superior limit, i could not test, as my computer do not produce tones with frequencies higher than 15 kilohertz. But hearing White noise and Pink noise, i could perceive it goes very high...including the harmonics to reproduce bells and wistles. This is the first time i perceive deep bass here in my small burn transistor place. regards, Carlos AKSA Graham, Carlos, I think this amplifier design is extremely clever; it uses a very innovative, hybrid approach. Congratulations! You have richened the art........ Cheers, Hugh destroyer X And if you observe, there's a Miller capacitor, despite Graham dislike it, my unit oscilated because of transistor made. I will fix, following Graham instructions....who am i to discuss with Graham? The Same output topologie you have in Aksa 55 and some clever ideas from Graham....this is all Graham design, i am just following his progresses and constructing some amplifiers. Another interesting is Symassym version 4, but MikeB disappeared when i told him that step by step he was changing his amplifier to be an AKSA....ahahahaha....i think he did not like the comment...... a long research, and changing here and there....a hard work....when i saw the output...there's the Aksa. He disappeared...may be with redish face near Andromeda Constellation. Live is beautifull, and our friends are wonderfull. Hey Hugh!...do not tell the guys!...between you and me... some secret, ....he was near to produce one Aksa...ahahahahaha! regards, Carlos mauropenasa My next amp of comparison( technical and acoustic ). I am a little worried by the Carlos enthusiasm. We hope that Graham leave still some chance of amelioration! :D Joking aparts, is a lot of interesting to build a amp so "simulated" and studied by Graham, because allows to all we of make a will in the reality his convictions. his availability to furnish her backgound ( and the schemes ) shows ( if there from it were be necessary ) her "genuine" passion for this subject. Am not much the planners prepared for himself put " in play ", but surely is the better thing... Ciao Mauro apassgear Mauro, No offence to you Mauro but could be offensive to others. Even if we know that this was not your intention. We all are, to some degree, limited with languish but reading your last post called my attention the way you refer to Graham. He, his, > refer to a male person. She, her > refers to a female person Someone had to tell you :) destroyer X I feel good when i found a good amplifier, already assembled more than 2.5K units...maybe 2.7K....and i am very tired of assemblage. This amplifier, as many Graham designs sound great, also Aksa 55 is a wonderfull amplifier and also the Symassym4 is very good. I can find more 5 or 6 good units...the rest....hummm...capiche? Your flag, seem Italy..but i am not sure uomo... va molto benne. regards, Carlos destroyer X To understand and copy other guy work, sometimes produce more work than start our own work from the beginning. But, with the intention to help, here is some images that may help friends. regards, Carlos destroyer X It is self instructive. I use to solder direct over copper, but you can transfer to the other side without too much work. The board size is 90 X 45 milimeters. A little more than 3 inches and a litle less than 2 inches...ahahaha! The TO-220 transistors and the output are connected to the board using thin twisted wires...not too long, as you may have 1 picofarad of capacitance each inche...so...keep them short. Over the board only the small transistors of the differential amplifier and CCS regards, Life is wonderfull, having amplifiers to assemble. Carlos destroyer X Graham is a wonderfull person, he will not bother related that small mistake made by Mauro....hehe, try to follow the awfull things i wrote here.... confusion of whole with hole....beach with.... my God!... forum friends need to have patience. Olá hermano, es un placer mui grande recibir el hermano que arriba sin embargo.... yo no hablo espanhol...solamente Portunhol. Hello brother, is a big pleasure the brother land here is this forum, without doubts i do not speak Spanish, only some strange language that is something between Portuguese and Spanish. Bienvenido nesta thread...welcome this thread! One transistor was missed.... a new image. Do not construct without check my board...i guarantee that will explode, as normally happens here..... understood....Destroyer Extra? Saludos Carlos destroyer X Lânquido es mucho sexy!...ahahaha! I do not now the English translation...let's see...language!...yessss!...my God! Carlos mauropenasa ooppsss... :xeye: Thanks of the correction, apassgear. Excuse me Graham... Ciaoi Mauro apassgear quote: Originally posted by destroyer X Lânquido es mucho sexy!...ahahaha! I do not now the English translation...let's see...language!...yessss!...my God! Carlos jejjeje, you are right Carlos :rolleyes: :hot: Graham Maynard Hi Hugh, What an incredible length of time it has taken me to come up with such a simple circuit. I have run it with a VN66AFD voltage amplifier (10th June page) which simulates with much higher open loop gain and thus extremely low closed loop thd. I did not make direct A-B comparisons, but did not notice any real advantage. However this driver does make the output stage become slightly inductive and increases first cycle distortion, thus I choose to stick with bipolar and be reassured that the transient response is excellent and damping control more phase coherent. I might look at this further, but then I ask myself if it will be worth it because I now have other more significant problems to solve - like source and speakers. I think Carlos feels like me here, ie. don't mess with it - it is already good, though any suggestions for improvements or variation will be viewed and/or tested with genuine interest, for this 'Diamond' as Carlos puts it, has only just been uncovered. Simulation shows that a wide range of bias conditions will still provide good results, but my intuitively chosen values give excellent real life performance. The class-A provides sufficient continuous damping until NFB assistance is provided via the class-AB output stage, and as before I have gone for 2 degree limits for phase accuracy between 20Hz and 20kHz, which is aided by the 600 ohm input.. This gives bass instruments real firmness and voices a detailed delicacy. The AB bias can be backed right off so that the circuit acts more like a 'dumper' instead of a 'helper', but I'm not interested in that. I also tried theoretically superior balanced inputs, but in real life ( and quite different to simulated results ) this impinged upon amplifier stabilty margin This was not worth the risk. I tried Susan's zero global feedback transformer/Mosfet arrangement, and it really does reproduce extremely smoothly, but for me the lack of low frequency phase coherence and hf detail put me back to expanding upon JLH class-A basics. Hi Carlos, I have never come across anyone who builds and tries as many amplifiers as you, and so quickly too !!! If you lived near Mauro, all our amplifier problems would soon be sussed out. I am honoured that you have not yet found reason to write that it is 'already in bits', 'in the bin' or 'got the 12 bore' ! I now see that you bring all wires to the pcb - this might be the cause of your oscillation. My pcb has only four small transistors. The VAS transistor is deliberately allowed to warm up on the main heatsink to increase its gain. I have only five wires going from pcb to heatsink. Two thick for power rails. One thick for output. One thin input to VAS. One thin from bootstrap to positive driver. All are kept spaced appart to prevent any unexpected cross-coupling. Hi Mauro, This amplifier (she) does well in the known to be unrealistic simulation tests I proposed, ie. examination of a first suddenly starting forward sinewave cycle at 10kHz; reverse injected damping examination, also a suddenly starting reverse injected 10kHz sinewave. Thus I expected it to cope well with real audio requirements, and it does. I do not have a clue whether this circuit will come out well in your tests, but its loudspeaker reproduction is good, so it must have something going for it. At least it is something different to add to our selection. I am taking a break from all the hypotheticals at the moment, but I noticed that reproduction comments were not related to electrostatic loudspeaker types, where the driving requirements are completely different and driver back-EMFs are not a problem. Some amplifiers are good with electrostatics but not with dynamic loudspeakers, and vice-versa; so an amplifier that is said to be good, should have the loudspeaker type tabulated too. Cheers ........... Graham. peranders Graham, I have a schematic over GM.25-8, what became of it? The amp Carlos now has built :up: is very similar. peranders quote: Originally posted by Graham Maynard I now see that you bring all wires to the pcb - this might be the cause of your oscillation. The design sure looks interesting and getting a pcb may be harder than expected because the circuits are a bit "nested", right? Not from one end to the other. It will be interesting to see where this design take us. Graham Maynard Hi peranders, The 25W-8 was not powerful enough. This is. It might look a bit nested but as Carlos said it can be opened out to look more conventional. With all the drivers and power transistors on the heatsink it becomes very easy to wire up. Cheers ........ Graham. destroyer X Also i had to fix transistos, and in this case, i had to produce some insulations too. I was very surprised that did not take too much time to finish, adjustment was fast and easy, but some care must be taken related the bias adjustment, as trimpot has not limiting resistor to avoid excess of current. When you advance control too much..current can go to 10 amps fast!... but this can be done without problem, with the inclusion of a single resistor. I continue to listen at it with pleasure, and now it is plugged in my DVD, Pearl Harbour Movie is shaking all home glasses...this amplifier go to very deep bass reproduction, and quality is very good. When the bomb penetrated the ship floor, crossing many stages and stopping in the ship's kitchen.... you can perceive, with details, the bombs delay helical mechanics spining... all voice wonderfull, but not too much emphasized in "S" letters, also rain is rain, and claps are claps, bells sounds as bells too. As all of you know, it is hard to perceive clapping different from rain, and without a helping image to inform you, some doubts can ocurr, as very few amplifiers can inform that to your brain....well, having that "resolution", we can say that the amplifier is fine. Related the amplifiers published, of course related the ones i could have or construct, as i can remember, only Graham 25W, Aksa55, Symassym from MikeB and this one....Oh!, the first JLH design have to be included, naturally. I have here one class D also...but rain are claps and claps are rain...i could not perceive difference, also Sound stage is negative...all sound appear having no deepth, separation is good and dinamics too. Tube sound is great, and those i mentioned have some "tube" alike sound. regards, Carlos destroyer X I have a lot of "brain" friends, also Hugh is competent as Graham is....but once, Hugh accepted that made a small mistake..... well, but Graham!.... everything he design runs well, of course, if you know him, as he dislike Miller capacitors, if something oscilate he will blame wires, batteries, conections, boards and everything but his amplifier..... the trick is to put a Miller capacitor if unit start to oscilate overheating the zobel filter. He is lovely and Great, but were you find love you find hate!...how can he so good this way? This time i will get him..... i found some "unstable" schematic to give a 12 bore or 12 gauge, double barrel shot!!!....ahahahaha! The circuit is hot as a hell...this way i will use small heatsink! To avoid guys discover that i am "preparing the results" i will put a fan over it. To the output TIP35 will be used To drivers will use reasonable pair...not so good 2SB1566 and 2SD2395 To Voltage amplifier, the good 2SC4793 (ahahahha!...low gain!!!) And to differential..... BC556C and BC547C The circuit has not Miller compensation capacitor...he is a destructor, and also a wonderfull thing when guys know how to use it, as need special material and deep tuning.... changing from 2 to 2 picofarads, the value, hearing sibilants sounds.... a kick in the bull! This time i will get you! Of course joking, but some true is hidden inside everything. I will construct this one, as the one that openned the thread is already deeply tested and i could not find defects...so..could not give a shot over it! But i have 26 of Graham's circuits.... at least one!...yeahhhh!...at least one i will SHOT! regards, Carlos destroyer X But this time i will get him! regards, Carlos destroyer X Will produce a very interesting effect. Booooom! Carlos destroyer X Will work a little.... already had some fun for this morning. regards, Carlos jacco vermeulen quote: Originally posted by Graham Maynard If you lived near Mauro 'Mauro' num pais tropical. :clown: (Jorge Ben will kill me) Mmm, Carlos, shouldnt the heatsink fins point the other way around :D destroyer X To have more success, the old guys (more than 70!) with genuine black hairs (still natural black) tried to change his name. Result that he disappeared in limbus...maybe dragged by an Universe Black hole, or travelling with grass to Sdrovosvária! hehe, really...not a good position, but the reason why is that my design department decides to put that thing horizontally inside an enclosure, having another one exaut fan...this is breathing air througth the fins...hehe...will melt everything!...will be Graham fault..... terrible i am! Vivo num país tropical Abençoado por Deus E bonito por natureza I live in a tropical country With God's protection. And beautifull because it's own nature Sdrovosbaria or Sdrovosvária, is something alike "nowhere land" regards, Carlos destroyer X I will draw components over it and will post it here. This will help me when constructing, and may be good to someone that is thinking in construct also. regards, Carlos destroyer X The board will be assembled here. regards, Carlos destroyer X Canavial because is the place we plant suggar cane...Cana Vial. The destiled of the liquid is called Cachaça, or Cana, or Aguardente....(water that burns). See the Rocket, will it explode? regards, Carlos destroyer X As i already explained, i use to solder over copper, but you can adapt this to create the standard board, the one with holes. I forgot the bias trimpot, it was missed in the board and i draw the connections....put some aerial trimpot, make your adjustment...wait amplifier to reach good temperature and adjust it again. Now remove trimpot, measure its resistance and replace it with a fixed resistor, if not exact value, use lower value than the one you measured, to reduce bias, in place of increase it. The bias is a little high, this represents enormous heat....79 watts of heat is difficult to manage...so, some reduce may be better to control heat, and will be worst to amplifier audio reproduction....well, there are sittuations that you have to make some choice, depending the heatsinks and fan you have. regards, Carlos destroyer X So, this trimpot must be installed some way to adjust. After that you can replace it by a fixed resistor, this trimpot is 100 ohms unit. I will assemble and test tomorrow morning, lasting some time to Graham evaluate boards and decisions i took. regards, Carlos Graham Maynard Hi Carlos, I check my e-mail fot the 10th June circuit - it said 'stabilisation not determined', so I hope you do not end up 'destroying something. I did have that one running but it was an intermediate circuit, and thus would be a backwards step from the 8th Jult one. Try 2k2 in series with VAS base. Cheers ........... Graham. destroyer X Well, i will take a look at it, if not perfect, i will change to the first amplifier published in that thread, the one is working very fine. Thank you, this way i will not bothered with this one, having any problems will produce the earlier one. regards, Carlos Greg Erskine I was having a look at the schematics of Graham's amp in posts #3, #4 and #5 and thought I do another to help me understand the circuit and give me a little practice using a schematic editor. ;) Anyway, I think it turned out pretty clear and I thought it may be of help to someone else. I used Carlos's version of the schematic (single output devices) and Graham's component selection. If someone sees an error please let me know. destroyer X This amplifier is very nice.... have wonderfull deep bass, you can inject 2 hertz and see the speaker diafragm moving. Have also impressive treble, but really natural, a think may be a little shy, as it is normal, not better than natural...our 55 produces a treble clarity that i never heard before, the effect or Hugh parts and tunning, the Australian King! But i will elect, my vote, the UK king, Mr. Graham Maynard. Aksa 55 produces more power, not beeing class A, also produces a little bit more dinamics, and produce less heat than Diamond Graham's amplifier. This one, Diamond recreated, is a very Hot , and not only related heat.....voice is very good, and i think it matches 55, beeing a wonderfull combination, installing Diamond into enormous woffers and Aksa 55 to voices and trebles, but both can reproduce all spectrum very well, having each one its own better points....a five channels amplifier will be something out of the Galaxy, if used 2 AKSA 55, 2 Diamong and 1 Symassym4 made by Michael Bittner, MikeB. MikeB is working hard very hard in Gutersloh, Germany, sleeping in his office, arbeite 18 hours each day to reach the dead line of his obligations...i am waiting his return to publish this Symassym4, because this one, the Symassym, used to subwoofer bass can go to 200 clean deep bass power and can go 1Hertz as Diamond can go (the last Graham model, he did not told me the name, this is the way i call it, so, he may appear with other name) I tried here to produce the same treble that i have with 55 amplifiers..... i used Hugh parts...but i could not. Diamond is perfect in treble, natural, do not emphasized absolutely nothing, but some sounds produced by 55 amplifier, as saxophone, marimba, bells, and some instruments are more than natural, are wonderfull.... to bass i jumped to Diamond...better to have all three i told you, as i already tested around 3.7K amplifiers and those are great!...but i cannot forget JLH basic designs, the first one is magnificent, the first King, infortunatelly dead...the problem is that it stay alone in its small power level, jumping out to any combination with other amplifiers...have to hear in small rooms and accept not to shake glasses or disturb neighboors, and this may be something very week for daily use. Greg, your amplifier will work immediatelly, without any problems.... well, i have one, oscilation!!!...but Graham already told the problem may be my wires and the transistors i used as voltage amplifier, the 2SC1819..... i stopped oscilation using a 50 picofarads capacitor from base to colector , directly soldered over the voltage amplifier leads (OH!... TERRIBLE!... will say the guys that love clean assemble), i do not care about that, beauty is not so important, others qualities are more important. My wife is beautifull as an accident, because i runned to her when i perceived what she has inside, as character qualities, was lucky to receive a pretty package to. I did not remove capacitor, because brazilian say told me: "Never change nothing in a team that is winning!" Because you can go and destroy the team quality substituting one person, as the same can happen substituting one transistor or capacitor. Symassym4: I will have to wait Bittner return alive, ...return to have normal life, as he have to finish to pay enormous money to banks and have dead line...he will born again and i will ask him to publish his schematic...he is very shy to do it. The last amplifier, a Graham diagram that was in development, not finished. I decided to construct, but Graham told that was just one experience, one step into Diamond creation, and that had some adjustments to be made. I did those adjustments, but really....aaaagh!...the amplifier has nothing to compare with diamond...i promiss to give a shot in every sh-- amplifier that i assemble, my dirty revenge, but as Graham is a very good friend, i will dismount board and construct there another P3A ( as i already made one once)....and probably will shot P3A circuit...because Rodd is not a close friend, and Graham is a close friend, so, if i dislike some board...hehe, better to be Rodd's circuit assembled there, than Graham circuit assembled over. I made the circuit, the second one published, but really not finished as Graham told..... he told that i will go backwards, alike crab walking backwards......and crab remember me a good name to this amplifier...crab!...crab!...crab!.......ahahahha! Please, DO NOT assemble that one, go to the first one, the finished and tested Diamond, a genuine product from the King of UK! God bless my Kings!.... long life to the Kings! I will show you some pictures of the crab..... it works...but is normal and standard, having nothing of special, alike thousands of amplifier you can find in web..... another grain of sand, very alike other grains of sand. regards, Carlos destroyer X So, image very adequated. Carlos destroyer X And P3A designer is not so close as Hugh and Graham. So.....hehe Carlos destroyer X And both are good amplifiers, but nothing exceptional on them...just a very good amplifiers, near to pop out of the bubble of nice amplifiers.... near to go to special amplifiers, and not incredible amplifiers. I am already reasonable old and Neurothical!... and i have not more patience with common amplifiers, as i made thousands of crabs!.... the only way i have to compensate the enormous tension and frustration with that long search of heaven!... a very frustrated search when i go to alive music to compare!, than i turn very agressive...better to be agressive with parts and boards than point all that agressivity to humans..this way guys!...shot the boards! I think will not receive a shot...let's see.....if receive a shot i will show you the result! regards, Carlos destroyer X I printed an studied hard the Blameless and all theories, and i run to construct happy to find, finally, a good amplifier. The most terrible crab i had in my home. Here is the picture of my feelings, that result in an agressive behavior related the circuit.... i Shot! the board....BOOOOOOOM! Will believe in someone that publish something with 1 percent distortion...may sound good. Carlos Greg Erskine Geez Carlos, I'm having trouble keeping up. :D Originally the thread was about Graham's "Diamond" amp. That's the one I think I did the schematic for. Then, I think you were talking about Graham's "Crab" amp, an early version on the Diamond but no good. You're pulling that apart? Next was Rod's P3A that was good but not special, so you're pulling that apart. I am right? And...the Leach amp sounded about the same as the P3A, so do you still have the Leach amp? I'd like more of your feelings of the Leach amp as I have been tempted with this one. Then, Douglas Self's Blameless "or should it be crab" amp. I guess you didn't like that one much. Arh, just got it, it has zero distortion, and you didn't like it so you want an amp with 1% distortion. ;) Thanks for the feedback on the schematic, now I have to learn how to design PCBs, then build the amp. Don't hold your breath. destroyer X In my point of view. It was produced to receive low impedance, already amplified signals from portable wonderfull modern CD players and things alike that, that have low impedance and 775 milivolts minimum over 8 ohms output. You see, that Graham do not like Miller capacitors, he do not want capacitance inside the audio chain. So, you see circuit very strange, without not too much capacitance in series with feedback line...only 22pf...also he is avoiding, in many schematics, to install the conventional 100 to 390 capacitor to ground....he is using, unless in the majority of his schematics, values lower than that...this is from Radio frequency only, to send Radio frequency to ground...but this is for the ones live near Radio Amateurs or Big Broadcasting stations...something that will work in one home and will not work in other thousand of homes. You see that he do not make any economy in electrolitic condensers.... because the use to non stabilized power supplies, some of them with small filtering inside...those capacitors guaranteed a stabilized voltage to amplifier input, without any ripple over it....also in the feedback network you will see that enormous condenser is there, ...to guarantee half cicle reproduction...this is the reason...the other one, is that Graham simulator, when you click in condensers...appear Milifarads in the place of Microfarads...so.....more fast to use, to go 1 Milifarad.... a thousand microfarads (he may be laughing now) Graham do not like this coil used in output.... he always cut it out!.... but my simulator, without coil, normaly do not works...present errors...damn simulators!.... Graham amplifier is working! The extra transistor is the real class A unit...over it all 400 miliamps and the one is heat as a hell...... the others, the normal output, common position of output when you observe, the complemetary units, are adjusted to 100 miliamps only.... so, this hi current, class A reference amplifier.... clean the whole thing!...how?... ask the King!...i do not know. I suggest you to construct....yeah, i suggest that...but take a close look and evaluate for heating of the 2R2 resistor.....if it is hot (very hot..near melt) this is a result of oscilation flowing inside and going to ground...... if this happens to you, put immediatelly the Miller cap.... oscilation will finish immediatelly...but do not put amplifier into wave form monitor...because...hehe.... sinusoidal at 80 kilohertz will be reproduced as an wonderfull triangle! There is something missed in the bias control...as you can advance it too much related the current...join base to emitter when start to adjust....dangerous to put enormous current and burn output...of course use series protective resistors. The other one, more an more carefull, as this one is the potentiometer to adjust the extra transistor VBE...you can kill it easy.... put the Bias transistor with base to emitter and them go first measuring base to emitter, i am talking about the extra transistor, put voltage there around 630 milivolts to start...check current now, in series with negative supply...beeing reasonable, let if be... fron 200 to 500 miliamps.....now, measuring over the 0.1 ohms, or even 0.22 ohms emitter resistors, try to obtain 22 milivolts over those resistors...check if both sides are equal in current. In the reality i use another adjustment.... a wrong adjustment... i made all circuit work with 400 miliamperes.... and i increase the current in the class A main extra transistor...this way.... the extra is working with 100 miliamperes...the opposite related Graham idea...but worked very fine...i will try the correct adjust latter. But you can see, the amplifier reproduce wonderfull sound, with big power ( i did not measured...but may be 35 RMS in my case), and not listenable distortion, very clear and flat...remembering JLH basic designs with much more power and deep bass reproduction......better than JLH?...yes, it is. Graham, please, take the chalk and go to the black board....oh!..now a days no more black, and nor more green..now is white with magnetic ink!.... modern things.... remember JLH guys! regards, Carlos Graham Maynard Hi Greg, That certainly looks more conventional, however;- I have found that taking the power rails to the pcb and then back out to the power drivers, gives more stable reproduction than when taking power straight to the output devices and from there to the pcb. This is why I drew the circuit with that seemingly unusual configuration. My mirror transistors are BC547C. The NPN bias sensing transistor is another 2SC3421bolted to the heatsink, though could be anything similar. 2SC5200 NPN outputs are my intended choice. I was showing exactly what was running in real life (not just a simulated suggestion), and there is no reason why a wide range of transistors should not work very satisfactorily. My Zobel resistor is 4x 10 ohm, and none of my resistors are inductive wirewound. Unless constructors use larger output devices like Carlos with his single output pair, I would not recommend driving unknown 4 ohm loudspeakers flat out unless plastic TO-3 sized transistors are run in parallel. I have smaller unindicated 'C's in parallel with the larger value ones shown, though I cannot say at the moment whether they actually make any audible difference in reproduction; would depend on electrolytic capacitor quality, which might deteriorate with time. I always use star earthing because I found that common mode injection or rf interference pick-up or even circuit instability has been caused by just a few inches of wire linking a signal ground and a psu/Zobel earthing point. I wonder ? Should I add the natural filtering of an input tube. Just seen your latest posts here Carlos, will read later, wondered why I hadn't heard from you ! Cheers .......... Graham. destroyer X You understand very well my meaning. When i made P3A i had already constructed the Leach Amplifier... the Super Leach amplifier, the one more complicated...not satisfied with the bass...it was there...but my cassetes that was very near the woffer did not shake (1 meter, not erased guys!) I could make them shake with a 35 watts amplifier,a Sony model that have a circuit very alike AKSA 55. P3A was something histerical.... very harshing in my idea..but do not remember very well...was dismouted alike thousands...i cannot keep them all here...no space enought...and too much money to keep them all here...so, parts are travelling from amplifier to amplifier..... with same parts and different sounds always. A friend was called to say something about Leach...i was not happy with bass, told him nothing....he listened and told me not a good bass, and that a brazilian copy of Leach, the Cygnus are better, and that he had one in his home...... so, we put them together in AB comparison..... the Cygnus home made and the Leach also home made....really..... Cygnus, the copy, referenced in the leach amplifier sounded better. I have it dismounted, when Rasmussen told that the guilty was the feedback line electrolic condenser, that i had to increase it...i was thinking to re-construct and do not did it yet....but maybe, the electrolitic condenser modifications will fix the whole thing...and a lot of good guys apreciate the Leach amplifier, so, it may be good..... do it and tell me what you think, please. I will give to myself a brake...to listen the Diamond...the first published, by me, of the Graham schematics, and the last in his productive mind (more dozen may be already created since that one)...i am just listening and having pleasure with it....will discover defects too...... have to hear and hear..... stop and continue....when you continue you realise new things...as long hearing, the brain make the needed adjustments...if not, how can a human beeing listen to a small portable radio with a 1 inch speaker without be boered.....brain adjustment i think. regards, Carlos destroyer X A simplified schematic...one model use TIP41 and TIP42 as drivers... a very low frequency unit.... i think 3 db roll of may be around 3 Megahertz of something around this limit. It is beeing used to Public Adress now a days...and also may units are inside enormous trucks, the Electrical trios, with people dancing over the truck's ceiling and all instruments there....a refrigerated inside,with double power generators and many PA units, surrounded with hundred of enormous speakers, exponential baffles, use to make our carnival parties, and some out of date carnival parties too...travelling whole country, during the year. And there is Cygnus...pumping air like a hell...producing pressure in my abdomen... and hitting "down countries too".... your arm hairs start to move...and you cannot avoid to protect years...as 20 or 30 kilowatts are not uncommon. here is the Cygnus schematic...the smaller one of the series...i think this one is 440 watts over 4 ohms...if i can remember...the model 400 was 200, this one 440 and have another near to ! kilowatt each channel...only more transistors and bigger supply. regards, Carlos destroyer X They were made based on Leach. Many model where made, with fans helping the heatsinks. PA400 PA-800 PA-1800 Those i can remember...the same circuit... some transistors changed because of bigger voltage, bigger swing, and supply was changed also, heatsink size of course increasing in the proportion the sound was increased. Also they used a TL071 or TL081 in the input, to increase the level to the power amplifier...but in general, all schematic the same. Good amplifier...very good amplifier. regards, Carlos destroyer X And man!...to put German tourist to dance...the sound must be very good. regards, Carlos destroyer X This link has a music playing. Click in "Galeria de Fotos" http://www.triopadilhao.com.br/ regards Carlos destroyer X Here is. Carlos destroyer X Here is Daniela Mercury and Caetano Veloso. Carlos destroyer X Regards, Carlos destroyer X The picture was made over the truck ceiling. regards, Carlos destroyer X He asked not to identify him...he is shy. Of course, his changes suggested in text, will make the amplifier alike Aksa..... of course, this guy is an Aksa fanatic..he use AKSA T Shirt and already travelled to Australia to have a picture together Hugh....poor Hugh, may felt embarrassed with that. Here is the schematic, sorry Greg, i use your schematic to put modifications over it, because your schematic is very pretty. regards, Carlos Graham Maynard Hi Carlos, Yes I was laughing exactly where you said I would be. If you have not readjusted your biasing yet it should run at 500mA per rail. With 36V per rail this should be a 36W total of quiescent heat. It has been warm here too recently but the circuit is nicely stable if the bias is re-trimmed after it has stabilised. This might take three adjustments. I was going to go for higher power rails but I don't want to blow my speakers. Soon maybe. Carlos this has all moved so fast. I had not named the design because I had not been able fully audition it. I am going to call this amplifier The 'GEM', in honour of my late father Gordon Ernest Maynard, who was one of three GEMs all named by my Grandmother. They were all in the RAF, my Dad a wireless/navigator in Mosquitos. It is he who encouraged my wireless/electronics interest from an early age, and he who made my investigative efforts possible. I had named an earlier JLH based class-AABB the GEM, but it too was shelved as a stepping stone towards this final classA//AB version. A diamond is a gem after all. No I do not like parallel capacitors along the audio path; when within the global NFB loop they turn the amplifier inductive. No I do not like series inducance either; the loudspeaker generated back-EMF works this component like an electric spring against a NFB loop amplifier's very low output impedance. Both of these inductivities affect the transient response and hf definition. I do not like Miller C at the VAS of a low bias class-AB amplifier because it slows down drive continuity when output stage conduction must be suddenly pulled through bias from one polarity to the other. As long as the value is small a Miller C will not have a similarly bad effect when the output stage operates in class-A. Your 'oscillation' solution is quite acceptable, but I had hoped it would never be necessary, and I know exactly what you mean about not wanting to change anything, as if there is too much to lose. Only 35W. The neighbours will never hear that ! Come on, add another speaker and increase the input. Hi Greg, From what I remember Carlos was so taken aback by his 'Blameless' construction it would not have lasted long enough to get its photograph taken. Cheers .......... Graham. Graham Maynard Hi Carlos, I need to get back here quickly. That circuit in Post#57 could blow someone's bass cone out or smash its voice coil. There is no start-up bias to Q6/Q8/Q10/Q11. Q7 and Q8 will generate a monstrous power-up thump. The 470 ohm Q6 base resistor will need 2.3mA, thus C3 will need to charge to 23V before bias stability can be established (too late speaker trash) The +ve rail diode will ruin class-A bias control; ie. 390 ohm for bootstrap pulldown, but 1390 ohm for release. The cross-coupled 47R/220n will also affect class-A biased control because Q7+Q9 work with Q11 in class-A, and Q7+Q9 with Q8+Q10 in AB. There is very also high gain with poor bass because R6 has not been reduced to 390. but hey, don't let me stop you having fun ! Cheers .......... Graham. destroyer X I never known that Aksa has a brand printed on T shirts...hehe, will ask one for me, i am a customer... my money gone to Australia, and has it value returned in sound pleasure, each dollar already return. Hello Hugh!...i think a T Shirt, Hiper, Extra, Super large, Heavy Duty may fit...remember...190 centimeters and 340 pounds heavy.... a musculous man!...ahahahaha! live is wonderfull, in special when people try to modify schematics. there's one component in 55 Aksa.... that one, if you remove....pluft!... half the quality goes! So, not a good idea change designs....i will fix my wiring Graham, but will make in the rigth channel amplifier, this one will stay the way it is... will close inside a wood cabinet, with holes to air flow, fan and cables and finished....because i will not hear it's beauty..will hear it's sound. Thank's all the informs, the US fanatic Connor guy will read that. regards, Carlos destroyer X I am also a Radio Amateur Graham...one of the last to make a CQ DX call (long distance radio calling) Second World war air figther was wonderfull Graham, you may be very proud, in your place i will feel pride because of that. Mosquito, if i am not making confusion, is the one with two motors and 2 vertical rear stabilizers... pilot go in the middle, between the motors... a beautifull airplane, could make economy of fuel flying high and with slower speed related others. I have not so pretty past.... a Religious man, a father of Christian religian arrived Brazil Northeast, around 1540, and could not resist to the Indians beauty...hehe, Here i am proving that religious people can be crazy for an indian girl (now Germans are turning crazy here, returning Germany with our girls), the father had a daugther, and some more generations passed , and one of the Portuguese Catholic Church father descendent married one Holland invader...and them one more marriage with an Italian woman and some more generations...here i am. My wife was worst, a pirate from Dennmark had one Enginner descendent that arrive Brazil to construct a railroad inside Amazon Jungle.....he saw an Indian very pretty, and runned to catch her with a rope.... they turn friends, first using energy, and time passing, they get marriage, and this produces my first wife....Red hair.... 180 centimeters Nordic woman. The second wife is result of an Spanish man, a revolutionary man that scaped from Spain. so, i have not so wonderfull past as you, ..... sorry, ...your Majesty, the king of UK. Very nice History, thank you to share that with us...in special with me. GEM is the name we call the stones used to jewellery here....so, we go calling GEM, not only because you want that, but also because is a name that is very adequated. Regards, Carlos jacco vermeulen quote: Originally posted by destroyer X could not resist to the Indians beauty I just knew you had a feather up your S, Carlos. :clown: nice story though, regards, as always. Stocker quote: Originally posted by Graham Maynard I am going to call this amplifier The 'GEM', in honour of my late father Gordon Ernest Maynard, Cheers .......... Graham. I thought you already had a "GEM" amplifier. In fact I am fairly certain, since the whole thread is printed and sitting on my shelf awaiting the time when my test equipment is built (by me). Should I build this one instead? Maybe Carlos is the one to answer that question! :) destroyer X Thank you Jacco, very deep message. Stocker: This amplifier can produce a lot of fun to you, but, observing your message; "Jesus loves you" I deduce you have something precious, already having something important to believe...those other things, are really a toy for big boys alike us. You see, even beeing a competent designer, even having fun with his constructions, famous to publish texts in important magazines,...you see the importance Graham give to his family, and how proud and sad he feel about his lost father. Graham is much more important to me since i discover the good heart he has inside the "crazy scientist envelope"....as important things that happens with him, fast he remember his father and dedicate all good amplifier to his father memories. I have to increase the size of my "feeling meter" related some persons...as Graham already top the up end limit. Jesus loving us, we can construct, or not, amplifiers.... Jacco, God is love too...so...let's have fun inside our shacking cars..ahahaha!...may be old car vibrations man? regards, Carlos Graham Maynard Hi Stocker, The GEM was a long time being developed. Th class-AABB did not sound right, this is the outcome of dozens of dead ends, and my finished design but for possible adjustable input filter, and temperature controlled line powered ultra low noise AC fan. I won't be building anything else. It is better than my transformer limited (even though expensive) KT88 chassis. Cheers ......... Graham. Stocker Wow. Are you going to start a new thread when it is "done"? This one sort of exploded out of the woodwork... I don't want to miss the final design! ShinOBIWAN Hi Carlos, Have you built the KSA 50 yet? Was wondering how it compared to Grahams 'Diamond' :) Graham Maynard Hi Carlos, The Mosquito - multipurpose - yes twin engines, Spitfire type, but the plane was faster than a Spitfire - my father said capable of about 450mph ( wooden airframe ) in a dive after munitions released, but not allowed unless for emergency escape purposes - pilot had to push through the throttle stop and then there would be a powerful kick of acceleration, but a report would have to be written giving details. Single tail only though. My Dad thought the war was awful; he did not like talking about it. His family home in London was not directly hit but suffered damage from a stick of bombs; they had to move. Of course it is not the amplifiers that are the real loves of our lives, merely aspirations ( our desires for better things ). Hi stocker, It already is the final circuit. Now named, as attached. I could do many things to this circuit, but every option has an associated disadvantage. At <0.01% this circuit does not have the super-low thd performance that some designers claim, but then nor does it have the NFB loop created output inductivity or poor phase coherence they don't even mention ! At present it has approximately 125W - 4 ohm maximum output, but I recon it could be taken to +/- 45 V rails for 100W - 8 ohm simply by changing the 390 ohm bootstrap feeder to 680 or 820 ohms, and I will try this, but not just at the moment. I want to think about a novel variable input filter where you can gradually pull down the hf response for less dynamic background/leisure listening, though this would not be just a simple tone control. I'd be happier with auto-speed silent fan cooling, though no problems so far; its just that an automatic fan would give total peace of mind. Why not try it for yourself ? Its not complicated ! Give it a new thread ? If it is any good it will make its own way out of here ! Though come to think of it, I wonder when Carlos ( or myself ) last put an amplifier in a case ? Cheers ....... Graham. Graham Maynard Couldn't get page to 'submit' destroyer X Do not know the mistake i have done.... i checked but could not find the mistake.... i have some time limit to try to fix my own errors of assembling...not more than 5 hours...passing that i accept that i fail and dismount the unit. If sound bad, with normal VBEs.... i check twice the circuit to see errors in passive... and if no error, i use to give a shot over the board.... my revenge...send it to hell.....BOOOOM! Will try again, someday, as i could see enormous thread related this circuit, and another very good Engineer, Mr. Jan Dupont appreciate that circuit, and this to me, is a strong sign that the circuit is very good. I also will produce another leach, because the first one was not good enougth..will try it once again. regards, Carlos destroyer X So, i will put more transistors in the output...and double extra NPN...will increase to 1K2 the resistor, will change electrolitics to higher voltage and will put power on it. hehe...wonderfull...and two fans pumping air to that cook machine. GEM is wonderfull...try it Stocker...easier to construct than you can imagine. Can call me Charlotte if not satisfied! regards, Carlos Workhorse Hi CARLOS, remember me..... You are one of the most amazing persons on the forum who fascinates me ...your style of writing is too wonderful......continue postings for ever..... When does the Brazilian Carnival festival starts...[on trucks] could you know the exact date..... Do you know the Selenium loudspeakers? At last sorry for little bit of off-topic. regards, K a n w a r;) X.G. hi,Graham I saw the power amp circuit whic similar to your A//AB on the Janpanese DIY magazine a few years ago.the circuit is AB class PP+CCS,and said that improving the sound. X.G. destroyer X You many times...you are very nice too. I thank you very much. Carnival is around february, the firsts days if i am not making mess...but it is in the same period in all Christian countries. We have a lot of "out of season" Carnival.... in special they use to be jumping from State to State, mainly in Brazilian Northeastern States, going to Bahia, that is the place were that out of season Carnival was invented....normally start at friday nigth and finished around monday 1 O clock morning. This is organized by State Governments...and there are the "Cold circuits too"...where those trucks visit the German towns we have here, normally in high and cold weather places.... "Winter festival" and things alike...my people is happy, irresponsable happy, having so much problems and smilling all the time...but this is better than find wars to figth and other crazy things.... we are not belicist...for us, brazilians, what matter is make love, eat good food, dance in good parties and never kill no one.... never! War is not our problem...we are always out of that!....i am afraid if someone invade if we will or not invite them to have lunch with us. Yes Selenium is already alive, but we have Bravox as the main company to produce speakers here...but very snobish with us, brazilians, you cannot find...all speaker are exported, mainly to United States...we have many other brands...to automobiles competition of sound loudness...but Selenium, Bravox, Arlen, Novik are the main brands here (i am not sure if Arlen and Novik are alive or not) Selenium produce an armoured tweeter, fluid cooled, silk dome that sounds great. Now the text i prepare of my last explosions here: Well boys, i used 54 Volts plus and 54 volts negative into GEM incredible amplifier. Also i replace the 390 volts resistor that is connected to positive line, the voltage divider to feed bias to all output transistor and trans resistor responsable for bias control, the Vbe multiplier. Also i had to increase the 1K resistor that is in the the feedback line, the one that control gain and is connected into feedback diferential transistor. I had to increase the electrolitic condensers value because of the supply increased value. Also i installed more transistor in the output, each one of them with 0.33 ohms resistor, the wattage capacity was increased to 420 watts each rail, two fan blowers where attached, one pumping air and the second one exausting.... heatsinks one facing the other, Insulated by wood and making a tunnel to air flow. Stand by current was increased to 1 ampere to each 54 volts rails that immediatelly reduced to 51.5 volts each rail. The extra transistor was adjusted increasing its Vbe till reach 620 milivolts...some nice current to control the amplifier, keeping one guaranteed class A current always flowing The other transistors are really AB with very bigh bias...the heat did not appear as my Heatsinks are enormous, fan blower pumping fresh air, natural 25 degrees wind crossing The room at 4 km per hour speed (around that, i have no anemometer...wind speed meter) First observation, the gain was reduced, even with the 1K feedback resistor replaced to 2K2....also the high treble turns less beautifull, and i think was the result of increased bias, and also because of transistor slew rate....having to achieve bigger voltage, delayed more time...so, speed to achieve maximum was reduced.... first time to observe the speaker hitting the end...hehe, very nice pop, this is real power...very high power, at least 80 Rms to make this speaker to touch end or compliance movement. Could not test further, because i touch the Miller Capacitor, very thin leads bent and a short was made between base to colector in my very good 2SC1819.... very hard to find Transistor this one...and burn a lot of transistors also...one fuse open, the other no..but both rails where damaged... beautifull Sankens gone.... Will prepare a cerimony and will be buried alike heroes in the Cemitery of D’Arlington, Having 21 shots of cannon 75 milimeters and Marines Military Band playing... a very Sad moment to me guys...thank you by the messages of support to that pain moment in My life....no way to have Sanken in this third world land...the best here is 15003 and 15004 or something alike...very slow and treble eater transistors. Now i am hearing the Symassym, another amplifier designed by Mr. Michael Bittner, a Class AB unit that sounds great and boomy...because of Tip 3055 in the place of the old And good Sankens that dead in battle to defend the high fidelity research in South America. At least, the one is very good even with that loss of brigth.... nice amplifier, but cannot, and could not, beat the GEM...the sound of GEM, was powefull and very clear...you Feel immediatelly that is in front of the Royalty. Regards, Carlos destroyer X But near future i will have big transistors here, and GEM will born again in 2 beautifull channels. Now the Symassym image. regards Carlos destroyer X This is the result of many designs and tests made by Michael Bittner. A very nice amplifier also....but cannot reach GEM. regards, Carlos destroyer X A good friend knowing that i lost my Sankens told me nothing. 10 minutes ago he called me by phone... and say, already gone. Then i asked him what is gone? He told me one pair of Sankens gone to my home adress by mail..hehe...in 24 hours more Sanken...and will hear GEM again! Here is the board dismounted, one of them, will make it again. This board was made for the second Graham circuit published in this thread, and missed one trimpot to adjust bias. This board is not for the last GEM version. Never give up! I will be back! regards, Carlos Greg Erskine Carlos, Thanks for the info on the P3A, Leach and Cygnus. I don't think I'll be starting a Leach amp soon as it takes me as long to do the schematic as it does you to build a complete amp, but the GEM looks very tempting. I think I got your recommendations for the GEM. Monitor the temperature of the zobel resistor for indication of oscillation. Add a Miller cap if necessary, but because of the class A component, it shouldn't be necessary. Be careful setting the bias pots because incorrect setting will possibly blow up the amp. I like the idea of putting in safety resistors. Any ideas on the appropriate values? How about a sketch? I'm not 100% clear on setting the two bias but I would worry about that yet. It will be months before I get to that stage. Graham, Thanks for the feedback on my schematic of your amp. I don't quite understand the implications of the low input impedance. Most amps range form 10k to 100k but the GEM is only about 600 ohms or so. Does this mean you must use a pre-amp? Also, the lack of an input cap means the device connected to the input must be free of DC. You mention bypass caps, which I assume are the 2 rail filter caps but should the feedback cap C6 be bypassed as well. Can you give a few details of the PSU? Carlos has sort of hinted at what you have but I'm not clear. With big caps at the amp you don't really too much on the PSU. What sort of DC offset can I expect? Thanks destroyer X I think explain now, how to adjust this bias, may be lost in your memory, as monthes will pass till you start them, but i will do it. Those things are good to make immediatelly after explanation, not only to avoid memory problems, but also to fix the information that will turn real practice.... this way will never forget...doing! I will construct other GEM, as you know i had my problems here...i have burned my circuit because an accident. (Stupid constructor) When re-constructed, i will put the safety resistor and will adjust it doing.... to limit current into a reasonable level...but this can be calculated too...because will be 2.4 volts flowing into 1K...we have to limit to 0.680 the voltage from base to emitter in the Vbe multiplier.... i will do it and will send you results. The first idea Graham had, was to adjust 400 miliamperes of current, passing inside the extra NPN transistor. The first step was reduce the normal transistor bias to zero... joining the Vbe multiplier base to emitter, turning the trimpot to that position, this way no current in the normal output.... and them adjust the other trimpot that increase extra NPN transistor VBE, till the current passing inside colector to emitter, reach 400 miliamperes..... and the last adjustment part was to increase the normal transistors bias to make them conduct something around 100 miliamperes...so, bigger current for the lonely one, the extra transistor, so, it will work in class A.... and low current to others, that will work class AB. But i made a mess and resulted very good... i made different... i adjusted the normal output transistors to 400 miliamperes, and them i turn the other adjustment (100 ohms trimpot) measuring the voltage from base to emitter in the lonely transistor...when voltage from base to emitter was around 630 to 640 milivolts, i stopped there,...this current will produce good current in Colector to emitter output circuit So, resulted in a class A amplifier with some class A aid...and even Graham class A has not with too big current..... we can call big biased class AB or underbiased class A...if you prefer bias between AB and A class amplifier. Graham decided to put 1 ampere flowing in the main transistors and more 100 miliamps to the extra one. (If i remember) So, things are not well defined in my mind, and Graham must pronunciate about that to define and clarify. To obtain the current that is circulating you can measure the voltage over emitter resistors.... will be 22 milivolts over 0.22 ohms, resulting in 1A...or ....100 milivolts over 0.1 ohms resistor, resulting also in one ampere. The extra transistor can be measure the same way....install a resistor in series with the colector and measure the voltage there...lets imagine 40 milivolts for instance,...using ohms law, you have to divide that voltage expressed in Volts (0.040V) divided by the resistance expressed in Ohms (0.1 ohm).. this will result in 400 miliamperes. To first power on, install 5 ohms resistor, 5 watts in series with positive supply and another one in series with negative supply, this is some protection, because 5 volts will be over that resistor when the current reach 1 ampere...check if both rails have the same voltage over those protective resistors. They can be used to obtain the current that is circulating, as you can divide the voltage you find over those resistors by its own value to obtain current...let's imagine you found 7 volts: 7 / 5 = 1.4A In this case, have to reduce current in the main bias trimpot, the one that have the auxiliary transistor. If something was wrong, and enormous current cross the circuit rails, the voltage over the protection resistor will be enormous..let's imagine that you measured 31 volts over the resistor... this have to be done fast, if not the resistor will melt or explode..too much power over it... the supply switch must be near your hand...or install fuses to burn and open the circuit if something turns bad to you. 31 volts measured over 5 ohms resistor will be 6.2 amperes. The interesting is that only some volts will be applied to the circuit, this will limit the power circulating... if 6.2 amperes are flowing, and 4 volts are measured (35 the supply voltage less the 31 volts that was developed over the 5 ohms resistor give the result of 4 volts....or the supply lost voltage and over the board only milivolts or you can reach 4 volts there. Power is voltage multiplied by current...so.... 4 Volts multiplied by 6.2 amperes result in 24.8 watts...no doubts that 100 watts transistor will be alive in that sittuation.... reason why people call that resistor protective resistor to power up. Well, save this to future references...needing more details i can do it and send direct to your mail adress...just tell me if you need something more, telling me your mail adress to. If i am explaining something that you already know..something you are tired to know.....very basic informs to you, excuse me, i am sorry, have no intention to disturb you...but will use that to inform some details to the very beginners that use to read our forum to learn some...as i use to do also. Greg, i am not receiving messages from the forum administration, so, do not know the reason...profile and all options are to send me messages...not receiving, so, needing some, use my direct mail adress: nanabrother@yahoo.com regards, Carlos destroyer X If some error, Graham will correct for us. It is missed in the sketch, but the Q5 lower resistor, the 470 ohms, is in the reality one trimpot. regards, Carlos Graham Maynard Hi X.G. Didn't get the chance to post this yesterday. Thanks for your feedback, I'd love to see the circuit - can you show a copy ? I have seen class-AB + CCS before, and I have seen claims that it maintains class-A operation. To me all it does is maintain class-A in one half and offset the zero current class-AB crossover for the other, though such an arrangement can work better when the load is a *resistor*. Here I have a separately active class-A device helping error correction at all times, and particularly during class-AB crossover at whatever voltage the loudspeaker might induce a class-AB current crossover. The lower class-A device is constantly working with the upper output half in the same manner that a JLH class-A amplifier operates. I also retain JLH class-A bandwidth and low output inductance. Cheers ....... Graham. EUVL Hi Graham, Congradulations on a very interesting design. Could you perhaps kindly highlight the advantages (or perhaps disadvantages) of this design compared to JLH Class A (Geoff's latest update), upon which it is apparently based ? One obvious aspect is power, but one could presumably build a high-power bridged version of the JLH using say a total of 8 output devices. This would also give close to a 100W ??. Thanks, Patrick Graham Maynard OH CARLOS !!! Straight to 54 volts is a big jump. My experience of bipolars has shown that every time you make a jump in rail voltage you *must* re-examine stability because the device characteristics change too. You know I always use 22 ohms per rail to check before applying full power, well even then I still use 0.22 ohm 2.5 to 5W wirewounds until I am completely sure. Sometimes these will smoke/smell to warn of a high current draw that might not be enough to blow a fuse. Also, I always keep an AM radio beside a new amplifier and listen for oscillation at start-up. All you should hear is mains bourne interference and the diodes rectifying. Very sorry to hear about this Carlos. Which position transistors blew, or did they all go ? Once oscillation sets in NFB controlled drivers can cross-conduct output devices to destruction. I also wonder why you had to alter the feedback ratio - the gain should not have altered, had it been oscillating from the start ? The quiescent currents do not need to be increased, 1A was rather a lot. I need a cuppa before I go through this biasing stuff. Hi Greg and Euvl, Will be back shortly. Cheers .......... Graham. destroyer X 37 volts, as things are stable there...guaranteed stable using this voltage. So, forget those dead Sankens...i am making efforce to fogive myself because of that.... stupid i was. But now, a friend re-arranged the things to me, and another GEM will born next week This weekend my birthday, so i will be around young girls in the beach and will have fun shopping...and eating as a thin dog! Next week will have GEM, this time a pretty assemble...decent wiring, as it will be done to stay...from now to ethernity..will go to my son and to my grandson....hehe, near future the amplifier will be inside some watch...and speaker will be electrodes touching skin and transfering direct to brain. I am afraid my grandson will use the aluminoum block to hold the door openned when we have strong wind days here...very common here strong winds. Destiny is very bad when we go turning old..hehe...really bad. Well, more two generations, if my grandson have one enemie... when young, can put the aluminium block inside one football ball...and ask the enemie to give a pretty good kick in that ball...for future they will find use for now a days amplifiers. I will post some informs, a simple sketch, not pretty, but really informative... big utility to the ones decide to produce this lovely amplifier. regards, Carlos destroyer X Parts are ready too, all them tested. So, will construct monday morning. Will be listening 9 AM, monday, next week Will have my Birthday watching "60 seconds" movie... "Eleonora" the car, the Mustang...the most wide world pretty automobile. 700 Horsepower and 300 watts of sound in my room...4 AKSAS and one strong Symassym. regards Carlos Graham Maynard Hi Carlos, _________________________________________ Bias adjustment. I initially said link out the class-AB bias pot and set up the 400mA class-A quiescent. Then remove link and add 100mA class-AB bias. Next I suggested it would be okay to set the AB bias preset at 50% before switch-on, then set 400mA class-AB bias, then the 100mA class-AB. No link would be needed using this method. Either method is suitable. __________________________________________ For the first power-up I start with 22 ohm seies power rail resistors and no load. Don't set the bias but check check for unusual heat and the output terminal being at zero, probably within a few millivolts. Next I remove 22 ohm and insert the 0.22 ohm rail resistors and set the bias with no load by measuring the voltage dropped by one of these resistors. 88mV for setting up the class-A, and then 110mV when the class-AB bias is added.. If everything is okay then test for a few hours and expect the current to increase. Then back off both biases, remove the 0.22s, and rebias without load whilst still at normal running temperature by using a meter in place of one of the fuses. For a stable zero at the output keep the input pair away from heatsink radiated heat. If one gets more radiation than the other, you could see drift up to about 25mV. ___________________________________________ Carlos. If the base-emitter junction of the bias control transistor is short circuited before switch-on, then it will be unable to conduct, maximum class-AB conduction will be set up and then the transformer will hum before the fuses flash. If the 0.22 ohm series psu resistors I mentioned are in place you should get away with it (that's my experience anyway) but if not, you would be much more likely to be unlucky. I once accidentally knocked resistor wires together; this shorted the upper driver base directly to the +ve rail before switch on. Those 0.22 ohm resistors saved the day because only the fuse blew, not even the 2SC3421 ! See, I am not immune to these kinds of mistakes either, but I have had no losses with this circuit, and with the 4 small transistors on a pcb and all others hard wired together on the heatsink with the pre-sets I cannot provoke oscillation by touching anything other than both the input and output connectors simultaneously. As with any amplifier this could cause damage, but it should never arise in real life. Hi Greg, As far as I am concerned every amplifier should be running at no more than the original 600 ohms studio standard. Higher audio frequencies are cleaner when pre-amps have a low impedance output, and 600 ohms amplifier input helps keeps the feedback limb impedance equally low. Lower feedback loop impedance is less influenced by circuit fields, but needs a high quality electrolytic and possibly smaller parallel C(s). (Exactly what you asked.) It depends on the component used and can easily be checked out by parallelling up during listening tests. It is easier to fit and be sure though. My PSU is no more than big transformer, big bridge and 2x 10mF per rail before each fuse. I have more 'C's in the parts box ready for final construction. Hi EUVL, I built a 100W JLH class-A 30 years ago, but it is not bridged. It is big, and oh-boy does it and the room get hot after a few hours, much hotter than a 100W tube amplifier ! It became a test bed for literally dozens of JLH derivatives. This is a recent photo, still with some test variations attached. It is 24 inches (60cms) long. Shown here on its side, with one side removed. The transformer is about 6x6x5 inches. There is a thermal overcurrent cut out on the rear panel; it would pop if you permanently shorted the output wires for more than a few seconds at full output. (Can't even find them these days.) You can spark its speaker wires noisily together at full input and nothing blows, not even the trip, but that is because it has pure class-A output. Also, put it beside a 100W class-AB amplifier and it is like having a choirboy beside a rock star; it does not have the dynamics !!! Pure class-A output current is naturally limited, and it cannot maintain drive through back-EMF induced loudspeaker impedance dips !!! I had the parts for two, but was so disappointed that the second was never built. I regard the GEM amplifier as retaining excellent JLH sound quality at normal listening levels, but it is also fully mature and ready to rock, as I am sure Carlos will confirm. I also expect this circuit to simulate reasonably well on computers other than mine. My finished monoblocks will use 2SC5200 where 2SC3281 are indicated. Cheers ........ Graham. destroyer X Of course he nows...he is the father of that children Related JLH comparison i agree with Graham, It seems to be two amplifiers with some logical switch on it...a genuine JLH beeing the first amplifier, and when power is increased, it turns a very good AB amplifier... will reproduce more bass and a very clear audio in a full range basis, because of beeing direct coupled reproduce 1 hertz without any problem, and bring with it the advantages of more dinamics and same clarity in trebles range than JLH basic first 1969 design...the one to substitute Williamson tube amplifier. The ones that apreciate JLH, without any doubt, will apreciate this one too. Observe that this amplifier is not for sale, so, this is not a Propaganda, this is just our fun to show you something good. No profit around this subject, in my case, nor an increase of pride, as i did not designed nothing, just cooperate with Graham, that feel always good when people construct his amplifiers and talk real things, making clear and true critics. My happyness is hear another good amplifier, and to share this with you all. Yes, maybe some payment, to see Graham happy, and related him also, the happyness is to share with you, i am sure about that. There's not GM or GEM factory or industry, and there's none intention to be created, maybe, plantpots of things alike, in the future, and as a result of family interests, but no electronic industry or design sales related Graham, the search is for quality and enjoyment. The ones that construct, please, tell him...he will be happy...me too regards, Carlos sam9 In a very broad sense it sounds like the same concept as Self's Trimodal. Two or three amps in one. However, the individual amp types are rather different. :devilr: :devilr: destroyer X I will start to hit my head into the wall. The one turned me really nervous...all my printer ink gone to save all those pages. I was expecting something very good. What a surprise i had... i am start shaking, and sudoring, sweating, transpiring...the skin is turning redish, arterial pressure is jumping tgo 25 to 21..... If i died, you are forgived..you do not know how i love to hear this name. regards, till the ethernity Carlos:headbash: :headbash: :headbash: :headbash: Graham Maynard Hi Sam, Trimodal ! ARGHH NO NO NOOOOOO. ( Carlos - a Trimodal is a Blameless with switchable biasing. ) I really don't want to say any more about it here. The fact that Douglas Self himself said you can hardly tell the difference in amplification between class-AB and A, says more than he realised. Cheers ....... Graham. destroyer X Had never tried blameless.....Aaaaagh! regards, Carlos mikeks quote: Originally posted by Graham Maynard