The FPS is basically a Tietze-Schenk low noise input stage, interpreted as a Fet input balanced INA. You can see it in T/S, Chap. 4 or in Burkhard Vogel´s "Sound of Silence", first Edition, page 48. John Curl send me a Fax with the same circuit from the 60th, also single ended, so this idea is ages old. A guy from Canada who´s name i have forgot made an ultra low noise single ended version although he is not a fan of the constant current sources. In praxis i found that the noise disadvantage is compensated by the gain advantage plus better PSSR. Two circuit i have in my lab prove that it is ultra low noise.
One is a single ended version with 4 x 2SK369. It is so quiet that Sam found it is more quiet then his line stage. The other is the AKSA Paris that has only one 2SK369 without source resistor at Idss. I have published measurements of both circuits before here.
I am absolutely sure that the constant current sources do no harm in noise IN THE CONFIGURATION I USE.
One is a single ended version with 4 x 2SK369. It is so quiet that Sam found it is more quiet then his line stage. The other is the AKSA Paris that has only one 2SK369 without source resistor at Idss. I have published measurements of both circuits before here.
I am absolutely sure that the constant current sources do no harm in noise IN THE CONFIGURATION I USE.
I dug out a measurement of the Paris, taken at -50dB/U.
That was at the time i posted measurements from my professional lab. The Paris is in green and the noise floor of my measurement system in blue.
You can see the single ended non NFB nature by looking at the second harmonic. That grows in proportion with the input signal and ultimately third is coming up too ( not shown ).
This is a single selected 2SK369 on Idss without source resistor. It is bipolar cascoded with a ground referenced cacode with a 470 Ohm, 470 Ohm base divider, signal going out of the base ! This is supplied by a bipolar constant current source with an RC filtered single green Led voltage reference. The SAM has 4 x 2SK369 on a 0,8 Ohm feedback resistor. The current source is referenced by two red Led´s with RC filter so the emitter resistor can be of somewhat higher value. Sigurd simulated that a 100 Ohm resistor adds around 1dB noise, so for lowest noise it should be somewhat higher, say 220 Ohm. this is all well researched and proven in praxis. It works.
That was at the time i posted measurements from my professional lab. The Paris is in green and the noise floor of my measurement system in blue.
You can see the single ended non NFB nature by looking at the second harmonic. That grows in proportion with the input signal and ultimately third is coming up too ( not shown ).
This is a single selected 2SK369 on Idss without source resistor. It is bipolar cascoded with a ground referenced cacode with a 470 Ohm, 470 Ohm base divider, signal going out of the base ! This is supplied by a bipolar constant current source with an RC filtered single green Led voltage reference. The SAM has 4 x 2SK369 on a 0,8 Ohm feedback resistor. The current source is referenced by two red Led´s with RC filter so the emitter resistor can be of somewhat higher value. Sigurd simulated that a 100 Ohm resistor adds around 1dB noise, so for lowest noise it should be somewhat higher, say 220 Ohm. this is all well researched and proven in praxis. It works.
Attachments
Here is the constant current source debate in detail. As far as i can tell we are taking about a maximum error of 2dB.http://www.linearaudio.net/userfiles/BV_to_OP_v1_pub(3).pdf
That noise discussion really bugged me. Phono Stages come and Phono Stages go at Maison Gerhard. Often i have not my best stages at home because they are on a loan. That situation is unbearable so i started to think about a REAL MPP that does not leave the house any more. To solve the noise issue ones and for all i designed a stage that consists of a pre-pre and a separate MM stage. That way i can test new input stages without that they need a RIAA. That pre-pre should be the lowest noise possible. So far the record was a single ended parallel symmetric transimpedance stage with 4 ultra low noise BJTs run quite hot, the same that i would like to use in the Masterpiece. I brought the noise resistance down to ca. 3 Ohm, most of the noise comes from resistors i need to set the gain and keep distortion under control. There are no emitter resistors. The signal goes straight into the emitters. That is a S/N of -82dB to -84dB, depending how it is measured or calculated. To improve on that i wanted a balanced input without the 3dB penalty so i designed a passive device, call it a "special" transformer. After consulting with several audio transformer makers and trying many prototypes i found a very satisfying solution. The noise impedance with the input shorted is only 0,75 Ohm. So without the cartridge the SN is even better by a factor of 4 giving over -90dB signal noise. This is held down to under 200Hz, after that frequency it is unavoidable that some hum is picked up. I even designed a "special" shorting plug because i found that even 20cm shielded-balanced cable shorted at the input picks up more noise then the pre.pre itself is capable of. To measure the noise more faithfully i use a relative method that does not show absolute readings but compares the noise of the phonostage to the cartridge output at 8cm-sec P-P. My Titan puts out a little bit less then 1mV under this conditions so the peak that you see on the black trace at 1kHz is actually at -60dB and not at -50dB. That makes the plots i will show you even more miraculous. Shown are the output of the Titan, the noise of the phonostage with the Titan connected, then with the cable shorted and then with the shorting plug. I will also make a measurement later that compared the "Diamond Dove" how i call it to the SAM, that is my lowest noise Fet design.
thats (unfortunately) the case.... working on my measurement hardware....
Often the case with designers, has been my experience.
The designer of my Classé Omega phono stage plus line amp mk3 has a surround setup with 4 of your Virgo (*) model plus subwoofer (good deal on them through friend Jochen Räke), last time he drove them with an Audionet ensemble.
Similar to cooks, star grade chefs are quite often fastfood junkies.
(* time when you used an Adyton pre/power set)