No, because we are trying to change the voltage at TP4 relative to the emitter voltage of Q27. Lowering the rail will decrease both voltages by approximately the same amount.
Thank you for your help ungie, I believe we are very near to sorting this out.
Unfortunately I can not keep working tonight so I will measure the points tomorrow.
Now I see the objective of raising R34... it will be very easy
You are very welcome Ricardo! I'm pretty sure that the offset you are seeing at RIAA is because Q27 is not conducting and hence we are only seeing the positive voltage due to Ic of Q28. Have a good night.
Andrew
My apologies for being the newbie in this thread, but what is Diamond Dove?
Andrew
That is my solution to solve the noise issue in MC Phono Stages. The best stages have around 0.3nV Hz. Like the JC Orion, the big Pass and the Accuphase. With bipolars or Fets,
no matter what, you reach a kind of thermal limit at ca. 5 Ohm. Also you have a rise at lower frequencies, that is 1/F noise, worse in Fets then in bipolars. Bipolars have shot and flicker noise. The BC327-337 are actually very good in that regard.
Usually the trick to lower noise in active stages is to use many elements in parallel ( Faulkner circuit). At around 8 in parallel the mentioned effects comes into place and paralleling more is not productive any more. One reason is that you can not have infinite gain in the input stage.
The detailed physical reasons go beyond posting on the Net. The best source of information is Burkhard Vogels book "The Sound of Silence". I recommend to buy it and read it. It is hard bread though.
My solution is a passive I/U converter that is part of a special feedback structure that i call forward coupling. It is not feedforward and it is not error correction. It is not NFB, neither passive or transimpedance like we do in the Paradise. In my system the micro information is directly coupled to the output and nowhere shunted to ground or send back to the input for feedback. It is balanced-floating-differential. The signal from the cartridge is not connected to ground and also not referenced to ground like in a conventional balanced stage. This is not for DIY, mind you, but i think it is interesting to publish the measurements. It has no 1/F so the S/N is maintained in the bass. It has 0,1nV/qHz and a noise impedance of slightly under 0,9 Ohm. The S/N referenced to 1mV is -94dB, 10dB better then any active solution i know. So there is room fore progress.
See the measurements. The black trace is the Lyra Titan producing 0,8mV at 1kHz.
Because it is a relative measurement you should deduct ca. 10dB. The signal peak is actually at ca. -60dB/V.
Red is the Diamond Dove shorted with a special shorting plus. Dark blue is the noise with Titan connected. You can see some hum pickup from the turntable motor and the rise in noise simple because the Titan has an impedance of 6 Ohm. In this case the cartridge produces more thermal noise then my stage. The light blue is 20cm of balanced-shielded cable connected to the stage. that cable is also shorted at the end but still it picks up the motor. When we go that low in noise, cabling and connectors show really bad effects.
no matter what, you reach a kind of thermal limit at ca. 5 Ohm. Also you have a rise at lower frequencies, that is 1/F noise, worse in Fets then in bipolars. Bipolars have shot and flicker noise. The BC327-337 are actually very good in that regard.
Usually the trick to lower noise in active stages is to use many elements in parallel ( Faulkner circuit). At around 8 in parallel the mentioned effects comes into place and paralleling more is not productive any more. One reason is that you can not have infinite gain in the input stage.
The detailed physical reasons go beyond posting on the Net. The best source of information is Burkhard Vogels book "The Sound of Silence". I recommend to buy it and read it. It is hard bread though.
My solution is a passive I/U converter that is part of a special feedback structure that i call forward coupling. It is not feedforward and it is not error correction. It is not NFB, neither passive or transimpedance like we do in the Paradise. In my system the micro information is directly coupled to the output and nowhere shunted to ground or send back to the input for feedback. It is balanced-floating-differential. The signal from the cartridge is not connected to ground and also not referenced to ground like in a conventional balanced stage. This is not for DIY, mind you, but i think it is interesting to publish the measurements. It has no 1/F so the S/N is maintained in the bass. It has 0,1nV/qHz and a noise impedance of slightly under 0,9 Ohm. The S/N referenced to 1mV is -94dB, 10dB better then any active solution i know. So there is room fore progress.
See the measurements. The black trace is the Lyra Titan producing 0,8mV at 1kHz.
Because it is a relative measurement you should deduct ca. 10dB. The signal peak is actually at ca. -60dB/V.
Red is the Diamond Dove shorted with a special shorting plus. Dark blue is the noise with Titan connected. You can see some hum pickup from the turntable motor and the rise in noise simple because the Titan has an impedance of 6 Ohm. In this case the cartridge produces more thermal noise then my stage. The light blue is 20cm of balanced-shielded cable connected to the stage. that cable is also shorted at the end but still it picks up the motor. When we go that low in noise, cabling and connectors show really bad effects.
Well, actually, there is a rise in noise in the bass even in my solution that amounts to ca. 10dB, That comes from the active second stage that is already at 1,3nV and has 20dB of gain before it from the passive I/U converter. I COULD parallel active stages too, say 4 for a gain of 6dB, but so what. I do not work for the military and it is not a life critical mission.
It is just for fun and enjoying the music, and music making it does.
It is just for fun and enjoying the music, and music making it does.
I build it all myself, by hand. So in a certain way it is not commercial. It is more a customized work of art. I also compensate the cartridge-turntable combination, so that it does fit ideally in a certain system.
I started to think about ultra low noise since 1980. That was the time gallium arsenide transistors where introduced. With liquid cooling i thought this could be much better then Fet`s or Bipolar`s. Unfortunately this kind of transistor has hefty 1/F, so it is very good at RF but not so good at audio frequencies.
The noise bug had bit me. It is like audiophilia. It is chronic disease that gets worse with age.
Anyway, yes you can buy it. If you sign a subscription and wait for 3 month.
I usually give away all that i know and when you have followed this thread and understood 10% of it you are a good phono stage designer and can find your own holy grail. This time i keep my mouth shut, sorry. I found a phenomenon that i have to study in detail.
If somebody wants to listen to it, visit me any time.
I started to think about ultra low noise since 1980. That was the time gallium arsenide transistors where introduced. With liquid cooling i thought this could be much better then Fet`s or Bipolar`s. Unfortunately this kind of transistor has hefty 1/F, so it is very good at RF but not so good at audio frequencies.
The noise bug had bit me. It is like audiophilia. It is chronic disease that gets worse with age.
Anyway, yes you can buy it. If you sign a subscription and wait for 3 month.
I usually give away all that i know and when you have followed this thread and understood 10% of it you are a good phono stage designer and can find your own holy grail. This time i keep my mouth shut, sorry. I found a phenomenon that i have to study in detail.
If somebody wants to listen to it, visit me any time.
Hi Ricardo,
could you put a 10K trimmer in parallel to R8 and see if reducing its resistance (start with 10k and go down slowly) impacts the offset? And, from your schematic it seems you have 10k in parallel to R4, can you remove that one please?
Could you also please check Q7, Q6 and Q24?
Also, it might be interesting to see the voltages without the servo OP being inserted (hope you have used a socket ;-)
good luck
as, that is the Diamond Dove that i find revolutionary.
For me it solves the noise problem.
It is more an intellectual satisfaction.
A stage that has 1nV/qHz and no flicker and popcorn noise plus low 1/F must not sound worse then a stage that has 0,1nV/qHz. Because of the self noise of the vinyl there is a lot of masking going on.
The Diamond Dove does sound exceptional not only because it has very low noise but also because it is balanced, has very low distortion and it is fast. RIIA is extremely accurate because of the way I implement it. Common mode suppression is very high because i trim both sides of the circuit with air caps. It also has a 6dB better overload margin then conventional stages run on the same voltage.
For me it solves the noise problem.
It is more an intellectual satisfaction.
A stage that has 1nV/qHz and no flicker and popcorn noise plus low 1/F must not sound worse then a stage that has 0,1nV/qHz. Because of the self noise of the vinyl there is a lot of masking going on.
The Diamond Dove does sound exceptional not only because it has very low noise but also because it is balanced, has very low distortion and it is fast. RIIA is extremely accurate because of the way I implement it. Common mode suppression is very high because i trim both sides of the circuit with air caps. It also has a 6dB better overload margin then conventional stages run on the same voltage.
Or better (I think [I do not like sockets
]) a third bridge point on the output of the servo OpAmp.
well, it did happen.... so much for late night building.... this is a beast of 56 bipolars, after all
I will check if the footprints of the bipolars can be changed to add a P or N, to help with this issue.
Frans, what do you think about having bridges everywhere so we can trouble shoot the complete circuit.
I did check this, but there is no 'everywhere' component in the schema