When I worked is Switzerland from 1974-1976, I found mechanical engineering superior in Europe, but audio electronics somewhat backward, compared to the USA. Today, there probably isn't any difference in electronics design, but I would still favor Swiss mechanical engineering, just like I prefer German autos. Americans still make pretty good IC's, but I prefer Japanese j-fets and bipolar transistors. That was a real problem, perhaps 20 years ago, because Japanese customers wanted American components inside the Vendetta Research phono stages, and I refused to comply, as there was nothing comparable in the USA to replace the Toshiba complementary j-fets, at the time.
It is interesting that even naming the equipment that you used at one time, Joachim, shows the interconnection of individuals. For example, I presume that you used the Levinson LNC-2 electronic crossover? If so, then Dick Burwen designed it for Mark, and I modified it with discrete modules. Dick Burwen was one of the original circuit designers at Scott Wurcer's company were he works: Analog Devices. This is how closely we are all related in audio design. That is why I know 'everybody's' business history.
Not to go into details but the term here is used rather loosely. Ray Stata and Matt Lorber were the founders, Dick was a very early consultant and hybrid designer. He did not "translate" well to IC's and went off to do his own thing. In an early "Analog Dialog" Ray credits him with being a founder, what can I say he is one of the most genuinely nice CEO's to come along, but this statement was not exactly correct (AFAIK).
We have one advantage. We have the Grimm farytales. The Grimm Brothers where born in Arnsberg (30km from my place) and worked in Kassel ( that is 90km away).
The farytales come from an aerea where my grandmon and grandpop where born, the Rothaargebirge. That is right where i live. Thats the reason i beleave in miracles.
Yes John, i owned the LNC-2 and it was more important then girls and drugs to me.
Equipment like that was simply not awaillable elsewhere on the planet at that time.
The farytales come from an aerea where my grandmon and grandpop where born, the Rothaargebirge. That is right where i live. Thats the reason i beleave in miracles.
Yes John, i owned the LNC-2 and it was more important then girls and drugs to me.
Equipment like that was simply not awaillable elsewhere on the planet at that time.
Today i build the Low Noise Bal - Unbal Buffer. Now using Low Noise Fet buffers and the AD797 i thought i could even simplify without much disadvantage. When i look at the Self solution more carefully he uses quite noisy buffers and high value resistors. Then he doubles or even quadruples. It whould be more profitable to use lower value resistors and get rid of the quadrupling. I am using 470 Ohm resistors now plus 100 gate stopper resistors and a 75 Ohm resistor. Noise is dominated still by the resistors but i was searching a compromise between noise and distortion. The LSK389C has an Idss of 10 - 20mA. If i have bad luck i get one with only 10mA. That could still swing 4.7V class A into the 470 Ohm resistors and so i build it. It came out very nice. Looks and performance are high class. This is certainly one of the nicer dead bug style groudplane designs i build. I have new hock up wire that is much easier to work with as the teflon coated variety i used before so it took me maybe 4 hours. I build it double mono and generously decoupled because this circuit is very fast. The Fets reach over 5MHz and the AD797 goes up to over 50Mhz in unity gain.
It worked right away without problem. I shunted the negative sie and listened unbalanced. Noise contribution is nill in my system. I use a 4.2Kohm series resistor in my preamp because my volume and balance control is shunt. Noise fully cranked is a very faint hiss, ear on tweeter. That is the same when i switch off the preamp so my poweramp is limiting dynamic range and my preamp is limiting the dynamic range of the buffer. The sound was not exactly the same with the buffer inserted but not worse eather. Before i got a sound that was ead accurate but also dry as a bone sometime. The buffer added some warmth and humanity without loss of detail. That can have a lot of reasons. One is that my preamp and my long 5m cable puts quite a heavy load ( around 5kOhm under unfortunate conditions ) on my phonostage and the buffer isolates that with it´s much higher 47kOhm load. The other reason is that the Fet´s could suffer a bit into the 470 Ohm resistors so i was curious and measured the input voltage with a fast sampling RMS meter. On the Classic Records reissue of Janet Jackson "Velvet Rope" ( that is a very loud cut) the maximum was 0.6V so there should be some safety margin. I could double or quadruple the Fet´s but i think in praxis that is pointless.
I am totally happy with what i got and many that to Scott with this invaluable input.
It is much apreciated. A very simple, elegant and effective solution.
By the way, that 0.6V transfered to 12V at my 4 Ohm 90dB efficient speakers. I changed from my 86dB 8 Ohm monitors to these recently. So i was listening with 2 x 36W. That should transfer to 106 dB at the listening seat. I sit close. Not much more the 2m´s. THAT was loud ! Still i could understand what she was singing very well and my ears did not heart. Well, that was extreme and i do that rarely, but it was fun.
Some words to record care: i hear all that horible stories of bend and noisy records.
1: give the record back to the seller when it is bend or warped or badly scratched
( my mothers recipy)
2: do yourselve a favour and buy a decent record cleaning mashine
3: clamp the record down. My Spiral Groove SG1 has a rubber ring that lifts the center and a clamp that presses the record down. Much less trouble with warps is the result.
4: Keep you stylus meticulously clean. I use Lyra SPT and some time the skin on my fingers but that is only for the non mechanically chalenged.
It worked right away without problem. I shunted the negative sie and listened unbalanced. Noise contribution is nill in my system. I use a 4.2Kohm series resistor in my preamp because my volume and balance control is shunt. Noise fully cranked is a very faint hiss, ear on tweeter. That is the same when i switch off the preamp so my poweramp is limiting dynamic range and my preamp is limiting the dynamic range of the buffer. The sound was not exactly the same with the buffer inserted but not worse eather. Before i got a sound that was ead accurate but also dry as a bone sometime. The buffer added some warmth and humanity without loss of detail. That can have a lot of reasons. One is that my preamp and my long 5m cable puts quite a heavy load ( around 5kOhm under unfortunate conditions ) on my phonostage and the buffer isolates that with it´s much higher 47kOhm load. The other reason is that the Fet´s could suffer a bit into the 470 Ohm resistors so i was curious and measured the input voltage with a fast sampling RMS meter. On the Classic Records reissue of Janet Jackson "Velvet Rope" ( that is a very loud cut) the maximum was 0.6V so there should be some safety margin. I could double or quadruple the Fet´s but i think in praxis that is pointless.
I am totally happy with what i got and many that to Scott with this invaluable input.
It is much apreciated. A very simple, elegant and effective solution.
By the way, that 0.6V transfered to 12V at my 4 Ohm 90dB efficient speakers. I changed from my 86dB 8 Ohm monitors to these recently. So i was listening with 2 x 36W. That should transfer to 106 dB at the listening seat. I sit close. Not much more the 2m´s. THAT was loud ! Still i could understand what she was singing very well and my ears did not heart. Well, that was extreme and i do that rarely, but it was fun.
Some words to record care: i hear all that horible stories of bend and noisy records.
1: give the record back to the seller when it is bend or warped or badly scratched
( my mothers recipy)
2: do yourselve a favour and buy a decent record cleaning mashine
3: clamp the record down. My Spiral Groove SG1 has a rubber ring that lifts the center and a clamp that presses the record down. Much less trouble with warps is the result.
4: Keep you stylus meticulously clean. I use Lyra SPT and some time the skin on my fingers but that is only for the non mechanically chalenged.
Call me crazy, but i am in the flow. I started to desing a BJT version of the Discreet INA.
Why ? Because i can, honest, i think BJT´s have lower 1/F breakpoint and higher transconductance and i just got a batch of SA1085E, SC2547E. In my humble opinion they are the finest Low Rbb BJT´s ever made. They have very high and linear Hfe ( E batch = 400 - 800). They have a high transit frequency at the optimum bias point of 10mA. The NPN has 300MHz there and by coincidence the PNP type has it´s maximum of 150MHz just at 10mA where the noise into low source impedances is also lowest ( 0.5nV/qHz). So two NPN and Two PNP can be brought down to 0.25nV/qHz, and 4 plus 4 to 0.1875nV/qHz. That will not happen in reality because room temperature set´s a boundary, but it comes in handy to reduce the noise in my balanced structure. I also experimented with augmenting the current drive into the feedback loop successfully without restoring to the Burr-Brown device that syn08 is using. Nothing against that OP, it´s more then decent but i like more variety. Lowering the gain setting resistor to 1 Ohm is no problem any more. In the structure that i use at home mometarily i have 2 x 200mA on tab and that with a huge variety of OP amps theoretically. Sofar i tested NE5534, LME48710, OP27, OPA134. No luck with LT1115, LT1028, LT1468 and AD797 so far. They have stability problems in that circuit. That can of cause be solved by other compensation but i am really running out of time now.
Why ? Because i can, honest, i think BJT´s have lower 1/F breakpoint and higher transconductance and i just got a batch of SA1085E, SC2547E. In my humble opinion they are the finest Low Rbb BJT´s ever made. They have very high and linear Hfe ( E batch = 400 - 800). They have a high transit frequency at the optimum bias point of 10mA. The NPN has 300MHz there and by coincidence the PNP type has it´s maximum of 150MHz just at 10mA where the noise into low source impedances is also lowest ( 0.5nV/qHz). So two NPN and Two PNP can be brought down to 0.25nV/qHz, and 4 plus 4 to 0.1875nV/qHz. That will not happen in reality because room temperature set´s a boundary, but it comes in handy to reduce the noise in my balanced structure. I also experimented with augmenting the current drive into the feedback loop successfully without restoring to the Burr-Brown device that syn08 is using. Nothing against that OP, it´s more then decent but i like more variety. Lowering the gain setting resistor to 1 Ohm is no problem any more. In the structure that i use at home mometarily i have 2 x 200mA on tab and that with a huge variety of OP amps theoretically. Sofar i tested NE5534, LME48710, OP27, OPA134. No luck with LT1115, LT1028, LT1468 and AD797 so far. They have stability problems in that circuit. That can of cause be solved by other compensation but i am really running out of time now.
Welcome to the world of the SA1085 and its complement. These devices were first announced, (at least to me) in a Japanese technical magazine that I found in Tokyo in 1978. I went all out to get them, including arranging an introduction to some engineers at Hitachi. Unfortunately, while these devices appeared perfect in every normal way, they generated extra distortion when put into my patented inverse comp circuit, because of the extremely low operating voltage between collector and emitter, in my circuit. I have never completely understood why these parts have this quirk. This is when I moved to more traditional low Rbb' complementary power devices, that worked better in my circuit, and finally to comp j-fets, when they became more available.
Have fun Joachim, at least you are trying new circuitry. That is more than I am doing, these days.
Have fun Joachim, at least you are trying new circuitry. That is more than I am doing, these days.
Voltage in my design between emitter and collector is 3.2V but i can arange for more if i run into problems. I did not know that these are so old. I am using the ROHM equivalents usually and i did not encounter any problem with the ROHMs so far. Malcolm prefers the BC550C/BC560C ( massive paralleld) even over the ROHMs. He told me why but this is 15years ago, so i can not remeber the reason. Have to call him anyway.
Here is the first idea. Maybe getting the bias-shunt feedback from T6, T8 is a better idea.
I am even thinkig about a Rush Cascode because it is self biasing but i have no experience with Rush Cascodes whatsoever.
Here is the first idea. Maybe getting the bias-shunt feedback from T6, T8 is a better idea.
I am even thinkig about a Rush Cascode because it is self biasing but i have no experience with Rush Cascodes whatsoever.
Joachim, when I found the Hitachi parts unsuitable for my patented design, I switched over to Fairchild PE8050 and PE8550 parts, that were rated to 1.5A or more. These parts allowed 0.4nV/rt Hz in 1979, without extra paralleling. The SOTA head amp (also the Michaelson and Austin headamp) was made with these devices. This is very much like Malcolm's use of the bc560 comp parts.
Last edited:
OFF TOPIC PLEASE?
JC. Did you design the M&A head amp? I had one, but it had maker's errors and due to the smothering of the components with a concealing goo no tech would go near it, and the then makers were unhelpful. But when it worked it was very good.