Hello,
I'd like to discuss with you some details in the phono section of the venerable Audio Research SP-10 preamplifier. It's phono stage comprises four 6DT8 double triode tubes per channel, which heavily resemble the 12AT7. Both triodes of each tube are paralleled, most probably to decrease noise by 3 dB. There are SS arrangements in each stage's supply which I believe are inductance simulators/gyrators. RIAA equalisation is done per NFB, which not everyone would follow. So far so good.
What I really don't get is the stage arrangement. I would have swapped stages 1 and 2, as I think, the cascode works best as the 1st stage. Additionally, Miller capacitance of the 1st stage is dramatically increased by 100 pF capacitors between the grids and the plates and adds to another 100 pF between grids and cathodes. A very heavy capacitive load for the cartridge, I think.
Does anyone have a clue what might have Audio Research led to design the RIAA phono preamplifier this way? Thank you!
Best regards!
I'd like to discuss with you some details in the phono section of the venerable Audio Research SP-10 preamplifier. It's phono stage comprises four 6DT8 double triode tubes per channel, which heavily resemble the 12AT7. Both triodes of each tube are paralleled, most probably to decrease noise by 3 dB. There are SS arrangements in each stage's supply which I believe are inductance simulators/gyrators. RIAA equalisation is done per NFB, which not everyone would follow. So far so good.
What I really don't get is the stage arrangement. I would have swapped stages 1 and 2, as I think, the cascode works best as the 1st stage. Additionally, Miller capacitance of the 1st stage is dramatically increased by 100 pF capacitors between the grids and the plates and adds to another 100 pF between grids and cathodes. A very heavy capacitive load for the cartridge, I think.
Does anyone have a clue what might have Audio Research led to design the RIAA phono preamplifier this way? Thank you!
Best regards!
The semi-con stuff is used for B+ voltage regulation, and resistive loads are used for the signal valves. Miller effect capacitance at the input is reduced by the same proportion as gain by RIAA loop feedback, and somewhat by the 20pF caps between grid and cathode ("bootstrap").
One could argue that the node between first and second stages is also sensitive to Miller C, so is a good place for a cascode, but probably few of us would make that choice today - or maybe we would.
All good fortune,
Chris
ps: I notice lots of changes to later schematics:
https://audioresearch.com/new_website/wp-content/uploads/2024/11/SP10_Manual_Schem.pdf
One could argue that the node between first and second stages is also sensitive to Miller C, so is a good place for a cascode, but probably few of us would make that choice today - or maybe we would.
All good fortune,
Chris
ps: I notice lots of changes to later schematics:
https://audioresearch.com/new_website/wp-content/uploads/2024/11/SP10_Manual_Schem.pdf
Last edited:
I'd be inclined to build according to my idea, i. e. a RIAA amplifier with a cascode at the input, followed by a gain stage and a CF at the output, using four 6922's with paralleled triodes, RIAA network in the NFB, and see what happens. I'd omit these overly complicated supply gimmicks in favour of a Maida regulator for the complete amp.
Best regards!
Best regards!
It looks like you already want to change the whole thing...I definitely would NOT (be inclined to build). This is way too complicated and I am highly dubious of the sonic benefits. It is from the 80s, a time when Minnesota Lutherans were convinced that only by over designing and including excess parts they could ask for high end sticker prices. 
😛
😛