Thank you Hans.Here's a copy of the AD797 documentation showing 1.5nV/rtHz@20Hz and a copy of a measurement done with the ZTX851 showing less then 0.3nV/rtHz@20Hz.
Two ZTX trannies in parallel per input amp. giving a noise improvement of 3dB per amp Two amps in a balanced config causing 3dB additional noise.
So, the remaining noise just that of one single ZTX 851.
I will reproduce Dick's measuring circuit diagram in LTSpice and see what it brings.
One thing is already obvious, measuring transistor noise with a Cap in series with the base will result in wrong results at LF.
Hans
You further the argument for direct-coupling of phono inputs.
A well-known author here at DIY scoffed at the idea a capacitor could introduce noise.
The reactance in series with the source certainly does.
FWIW the PNP ZTX951 is even quieter, by a small margin, than the ZTX851.
The ZTX951 is considerably quieter than the 2SB737.
Pardon the hum: https://proaudiodesignforum.com/forum/php/viewtopic.php?p=14790#p14790
As to the consistency of the ZTX851 I've used hundreds and never found a noisy one.
I've only thrown away one device: It was born with a base emitter short.
Jan, you have to diffenciate Current and Voltage noise. For an MM cartridge, BC550 and 560 work well. You can make a halve decent MC Stage too by paralleling say 8. In the old Time we called that Faulkner circuit. I think He came up with that idea in the 60th of 70th.
Faulkner : "Optimum Design of Low-Noise Amplifiers"
https://www.keith-snook.info/Stuff-you-should-read/Faulkner - The design of Low-noise audio frequency amplifiers.pdf
https://www.keith-snook.info/Stuff-you-should-read/Faulkner - The design of Low-noise audio frequency amplifiers.pdf
I remember in the old company we used 4 parallel BC550C and 4 560C in the complementary
input of the MM stage (with emitter resistors 47R). Other models only had a single pair.
input of the MM stage (with emitter resistors 47R). Other models only had a single pair.
.. but I take my hat off to the fact that he received funds from a "National Sustainability Programme project" for this.
.. with emitter input.
You should always show your picture source, one is
https://www.bonavolta.ch/hobby/en/audio/ats.htm
You should always show your picture source, one is
https://www.bonavolta.ch/hobby/en/audio/ats.htm
I have drawn the circuit a little bit different to understand it better.
The datasheet of the 2N5515 shows that the parameters of that Fet are not very well defined.
For example Gm can vary with a factor of 4. Interpolalating I get the approximate idle current.
After modern standards this is not a particular low voltage noise circuit . On the other hand my friends in Paris in the 80th usually used the DENON DL103 that has a rather high impedance of 40 Ohm and only 0,3mV output.
In comparison my Lyra Atlas Lambda has 0.56mV and 4,2 Ohm, pushing out approx 20 times the current.
I have an idea how this circuit can be modified for cartridges like that.
Also I am not sure if the 120 Ohm resistor is the noise budget or if it appears parallel to the rest of the circuit.
The datasheet of the 2N5515 shows that the parameters of that Fet are not very well defined.
For example Gm can vary with a factor of 4. Interpolalating I get the approximate idle current.
After modern standards this is not a particular low voltage noise circuit . On the other hand my friends in Paris in the 80th usually used the DENON DL103 that has a rather high impedance of 40 Ohm and only 0,3mV output.
In comparison my Lyra Atlas Lambda has 0.56mV and 4,2 Ohm, pushing out approx 20 times the current.
I have an idea how this circuit can be modified for cartridges like that.
Also I am not sure if the 120 Ohm resistor is the noise budget or if it appears parallel to the rest of the circuit.
Attachments
Interesting circuit. But I note that while they say it is to solve a large output swing with low distortion, they only quote distortion figures with 1V output.Here comes the paper.
Suspicious.
Jan
The study on the special Vas/Cascode you posted above: Cascode-based voltage-amplifier stageWhat circuit ?
Silentswitcher is passé. Manufacturing cost have almost doubled, as have the chips, if you can get them. It´s a mess.
Jan