Hi,
If you want to be extreme, try socketed naked Vishay Bulk foil resistors, change to short for MC...
Well, if you do this for DIY I do not see any of these being a problem.
Ciao T
If you want to be extreme, try socketed naked Vishay Bulk foil resistors, change to short for MC...
Well, if you do this for DIY I do not see any of these being a problem.
Ciao T
Hi,
That was not the point here (obviously). It was not about non-linear capacitance per se, but input capacitance as absolute value. At the signal levels we are dealing with for MC/MM Phono the non-linear capacitance impact is still quite small, compared, for example, to a low gain Linestage application.
The concern was the 75pF Ciss per 2SK369, which would make a serious (but in actuallity not that serious) input capacitance, which may cause trouble with MM Pickups.
And then there is the 15pF Crrs, but that is largely neutralised in this circuit (it is also actually the more non-linear of the two), however that too is cut by local degeneration.
If we add local degeneration that will drop the Gain by 20dB we need to drop Gm by 20 dB and as a result we now "bootstrap" Ciss and reduce the Miller amplification of Crss cutting their effect to around 1/9 of the original undegenerated value.
Hmmm, did you notice that this circuit is cascoded? Okay, one halve of the input differential operate into a current mirror, this is something I personally would probably change, but the current mirror can be made quite low impedance, so it's performance will not be far from a cascode.
Still, I can see no reason to not make the Current Mirror into a folded cascode and shift the current mirror to the output stage, but honestly, the current mirror where it is not so bad either.
Ciao T
That was not the point here (obviously). It was not about non-linear capacitance per se, but input capacitance as absolute value. At the signal levels we are dealing with for MC/MM Phono the non-linear capacitance impact is still quite small, compared, for example, to a low gain Linestage application.
The concern was the 75pF Ciss per 2SK369, which would make a serious (but in actuallity not that serious) input capacitance, which may cause trouble with MM Pickups.
And then there is the 15pF Crrs, but that is largely neutralised in this circuit (it is also actually the more non-linear of the two), however that too is cut by local degeneration.
If we add local degeneration that will drop the Gain by 20dB we need to drop Gm by 20 dB and as a result we now "bootstrap" Ciss and reduce the Miller amplification of Crss cutting their effect to around 1/9 of the original undegenerated value.
Hmmm, did you notice that this circuit is cascoded? Okay, one halve of the input differential operate into a current mirror, this is something I personally would probably change, but the current mirror can be made quite low impedance, so it's performance will not be far from a cascode.
Still, I can see no reason to not make the Current Mirror into a folded cascode and shift the current mirror to the output stage, but honestly, the current mirror where it is not so bad either.
Ciao T
Joachim,
Quite an interesting graph. You have quite an undamped arm resonance there at 300Hz and of course the main one at around 7-9Hz (do I see a double hump from slightly sticky bearings?), which will contribute massively to the LF noise rise.
The Noise at 20Hz shows around 12dB excess noise over the midband noise plus RIAA influence.
Midband noise is around 20dB below the LP's background noise, so if this was 4pcs 2Sk170 & 4pcs 2SJ74 it should be quite similar to the Vendetta (which is after the Yardstick in this thread) and shows that as much as 20dB more noise than that may be endured with 0.3mV Cartridges...
And of course, there is around 1% 2nd HD already, at a signal level that is basically -14dB. So our 2nd HD at full signal will be around 5% (more at lower frequencies). I presume (well, I actually I know) that this HD is not your Phonostage, but the LP/Cartridge Combo itself...
Some lessons here for anyone making a phono stage, don't loose sleep over a bit of noise or THD, unless you intend to have John Atkinson or the German Magazines measure it and want them to say "excellent" while Mikey or the other Reviewers gripe about "lack of microdynamic resolution", "solid state colourations", "insufficient macrodynamic expression" and the like...
Ciao T
Quite an interesting graph. You have quite an undamped arm resonance there at 300Hz and of course the main one at around 7-9Hz (do I see a double hump from slightly sticky bearings?), which will contribute massively to the LF noise rise.
The Noise at 20Hz shows around 12dB excess noise over the midband noise plus RIAA influence.
Midband noise is around 20dB below the LP's background noise, so if this was 4pcs 2Sk170 & 4pcs 2SJ74 it should be quite similar to the Vendetta (which is after the Yardstick in this thread) and shows that as much as 20dB more noise than that may be endured with 0.3mV Cartridges...
And of course, there is around 1% 2nd HD already, at a signal level that is basically -14dB. So our 2nd HD at full signal will be around 5% (more at lower frequencies). I presume (well, I actually I know) that this HD is not your Phonostage, but the LP/Cartridge Combo itself...
Some lessons here for anyone making a phono stage, don't loose sleep over a bit of noise or THD, unless you intend to have John Atkinson or the German Magazines measure it and want them to say "excellent" while Mikey or the other Reviewers gripe about "lack of microdynamic resolution", "solid state colourations", "insufficient macrodynamic expression" and the like...
Ciao T
That must be from Putin's breakfast meeting where he leans to a Naim CDX2's tray. For Medvedev the word is he also (later?, elsewhere?) has the $60K Vitus phono and Tidal speakers.
Thorsten, the measurements are from an Olympos in a Triplanar arm. I have now changed to a Titan i and will repeat the measurements with a bipolar phonostage i made. Sooner or later i will change the arm too. I have a Schröder Artemiz and wait for a Spiral Groove Centroid. To tame the main arm resonance at ca. 7 to 9Hz i will try a subsonic filter or i could fill more oil into the damper of the Triplanar or both.
QSerra, the measurement with the lowest noise is the residual of the measurement aparatus. Then from bottom to top comes the 2SK372V, BF862 and the 2SK170.
QSerra, the measurement with the lowest noise is the residual of the measurement aparatus. Then from bottom to top comes the 2SK372V, BF862 and the 2SK170.
Hi,
The testing of the K372 and BF862 at 1.2mA Id is impressively low noise.
Attempting to get low noise from a 2SK170 at 1,2mA Id is a fools errand though, just look at the Noise vs. Id curves in the datasheet.
Ciao T
The testing of the K372 and BF862 at 1.2mA Id is impressively low noise.
Attempting to get low noise from a 2SK170 at 1,2mA Id is a fools errand though, just look at the Noise vs. Id curves in the datasheet.
Ciao T
Joachim,
They are very interesting. If I ever make a 1980's Japanese Style compound RIAA circuit with Op-Amp and two J-Fets for RIAA clearly the BF862 is ideal, low current for low noise, very nice possibilities...
I agree at running at least > 50% Idss with the whole K170/J74/K147/J72 range, noise is lower based on my tests, though the rise at LF remains.
Ciao T
They are very interesting. If I ever make a 1980's Japanese Style compound RIAA circuit with Op-Amp and two J-Fets for RIAA clearly the BF862 is ideal, low current for low noise, very nice possibilities...
I agree at running at least > 50% Idss with the whole K170/J74/K147/J72 range, noise is lower based on my tests, though the rise at LF remains.
Ciao T
The 1kHz tone turned out to be the carrier(?) for some kind of HVAC control system that talks through the power lines. Nice of them, it now contaminates the entire building.
These were for microphone circuits, but in general I don't go by noise vs. Id. For FET's that generally just tells you where the parasitic resistances kick in. Many of these FET's maintain <4nV @ 25uA of Id, opens up a bunch of neat battery powered circuits.
The massive paralleling of FET's at high Id to meet some stake in the ground spec on noise seems brutally obvious to me by now, I think it's done, yes?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.