No it's not. It's his own polemic that is based on personal hypotheses. There is nothing wrong with a good single-ended amplifier and his remarks about about the high end are simplistic generalizations that are way off base.
John
I always keep a tube chain at hand, first because tubes can sound very good and second as a reality check how far my silicon designs are from ancient but proven technology.
The amps i use are LA Audio 6550 monoblocks in push-pull triode mode and my preamp is a rebuild Velleman 8020. Originally i had rebuild it to Sander Sassens specs. The topology is a simple anode follower DC coupled to a conventional kathode follower. The result was not bad at all but two thinks bothered me : first, the sound is a bit smoothed over, quite "tubey" if you like and second i could not get rid of a small hum that was very system dependent. That combination of that flaws kept me from listening to the tube preamp much. I do not need much gain in my system so i decided to rebuild the stage again to a buffer. After studying various White and Taylor followers and buffers with constant current sources i eventually decided on an Aikido buffer. inductors where out of the question because of space constrains. i designed an Aikido buffer around the ECC82 tube in use. I like the Aikido buffer much. It has good PSU rejection and a measure of distortion cancelation. I set the tubes at a bit under 7mA and got the following measurements: -3dB @ 500kHz, unclipped output voltage over 40V, my measurement limit. Hum is gone now and the sound is much more open and dynamic. Why do i tell you all this: well the reason is that the ECC82 can be an excellent line tube if used right. See the simplified schematic and the unfortunately not very tidy look of the modified PCB board. P.S. my circuit editor does not allow to draw tubes so i use a FET symbol.
The amps i use are LA Audio 6550 monoblocks in push-pull triode mode and my preamp is a rebuild Velleman 8020. Originally i had rebuild it to Sander Sassens specs. The topology is a simple anode follower DC coupled to a conventional kathode follower. The result was not bad at all but two thinks bothered me : first, the sound is a bit smoothed over, quite "tubey" if you like and second i could not get rid of a small hum that was very system dependent. That combination of that flaws kept me from listening to the tube preamp much. I do not need much gain in my system so i decided to rebuild the stage again to a buffer. After studying various White and Taylor followers and buffers with constant current sources i eventually decided on an Aikido buffer. inductors where out of the question because of space constrains. i designed an Aikido buffer around the ECC82 tube in use. I like the Aikido buffer much. It has good PSU rejection and a measure of distortion cancelation. I set the tubes at a bit under 7mA and got the following measurements: -3dB @ 500kHz, unclipped output voltage over 40V, my measurement limit. Hum is gone now and the sound is much more open and dynamic. Why do i tell you all this: well the reason is that the ECC82 can be an excellent line tube if used right. See the simplified schematic and the unfortunately not very tidy look of the modified PCB board. P.S. my circuit editor does not allow to draw tubes so i use a FET symbol.
Attachments
The fact that you use tube amps as a reference for sound quality is most significant to me.
What is drawn I would not call Aikido. It is a cathode follower with triode load with injected PS ripples. Try the same tubes in an augmented cathode follower. The upper triode would improve PSRR as well, decreasing hum from your ripply PS.
Wavebourn, my unknown friend, the inventor of the aikido circuit calls it that way :
Aikido Cathode Follower & More Triadtron
Aikido Cathode Follower & More Triadtron
P.S. hum is non existing audibly in my implementation of the circuit athough one channel is dead quiet and the other makes an extremely low wideband noise. A potmeter instead of the 3.9kOhm resistor could make that trimmable. We found that one of the tubes i am using ( using a tube tester) measures a little "fat" so adjusting that resistor we could find the optimum null. The triode load is unlinear so is the kathode follower in the oposite direction. The same happens with noise from the PSU. It is in antiphase so cancels to a certain degree. I did the calculation for a mu of 17. That can vary in reality.
Hello Joachim and All,
The circuit you posted is exactly an Aikido; it is an Aikido Cathode Follower. Wavebourn see this page Aikido Cathode Follower & More Triadtron
Solid is the one word description that comes to mind when speaking of the ACF circuit. I am using the ACF to drive SET output tubes. Lots of full scale dynamics here.
A question for you Joachim. You have built SS amplifiers. Have you built any class A single end amplifiers? If you have how would compare the sonic signatures and dynamics to single end triodes?
DT
All just for fun!
The circuit you posted is exactly an Aikido; it is an Aikido Cathode Follower. Wavebourn see this page Aikido Cathode Follower & More Triadtron
Solid is the one word description that comes to mind when speaking of the ACF circuit. I am using the ACF to drive SET output tubes. Lots of full scale dynamics here.
A question for you Joachim. You have built SS amplifiers. Have you built any class A single end amplifiers? If you have how would compare the sonic signatures and dynamics to single end triodes?
DT
All just for fun!
It is hard to make a solid state amp that has the midrange liqiudity of a single ended tube. Class A design helps of cause but you loose the efficiency advantage. The problems in tube design have a lot to do with the output transformer. A friend of mine is able to do a transformer with airgap for a 300B that has the -3dB point at 450kHz. Then we are really cooking. He talked a lot to old engineers at Siemens, Telefunken, Saba etc. and did experiment over a period of more the 20 years. On the other hand i am working with a team of High Tech designers to decode the silicon transistor and it is theoretically posible to design a circuit that takes care of the flaws in semiconductors.
In the end it boils down if you want to create a "pretty sound" or a "transparent" sound.
Considering the flaws in recordings and playback, a little prettyness can not hurt. I think the human race is not prepared for truth yet.
In the end it boils down if you want to create a "pretty sound" or a "transparent" sound.
Considering the flaws in recordings and playback, a little prettyness can not hurt. I think the human race is not prepared for truth yet.
It is practically possible, as well.
There is how I am doing so:
http://www.diyaudio.com/forums/tube...er-iii-high-end-hybrid-amp-4.html#post1889739
A semiconductor device works between voltage follower and counter-modulated current source. As the result, variations of voltage and current through it are kept minimal. I used such output in hybrid power amps, in line output buffers, in headphone drivers.
Similarly, you can linearize your CF's transfer function.
I do not know. Probably not. Maybe he just invented the name. The only thing i can say is that it does not disappoint sonically.
SRPP is not for me. Here in Europe are at least 5 books about SRPP on the market that all have different interpretations of the circuit. As far as i understand it, it has low output impedance until you load it. When i see your circuit elaborations i shut up. I am an ok tube designer but not world class. It is just a hobby for me. Keep on trucking and enlighten us. I keep on fighting with my solid state.
SRPP is not for me. Here in Europe are at least 5 books about SRPP on the market that all have different interpretations of the circuit. As far as i understand it, it has low output impedance until you load it. When i see your circuit elaborations i shut up. I am an ok tube designer but not world class. It is just a hobby for me. Keep on trucking and enlighten us. I keep on fighting with my solid state.
I am more a specialist for low level, low noise design. Our latest current slope retracking circuit has a Gain-Bandwidth product of 50Ghz in one stage. Theoretically it is just a chain of diodes.
- Status
- Not open for further replies.