John,
Sure, or there was parallel development (you did not put the second stage cascodes in). Looking at the schematic and with my 1980's EE Knowledge I'd call the design "blatantly obvious".
Yup, the LH0032 is basically Class B, nothing that judicious application of a resistor to the negative rail (then we end up with a Kaneda) or an external Follower cannot fix.
Sure. Now what was the noise of the TL082 so beloved by many HiFi Designers? Was it 18nV|/Hz at 1KHz? That is about even stevens. A little on the high side for MM maybe, I get maybe 75dBA post RIAA, guesstimate...
Might have, 1mA output stage current is a bit measily. But if we wanted to use this suckah in audio it may be up to us as the circuit designers to do something about it...
Precisely, I have yet to find any "Audio" Op-Amp that I actually like the sound when used for Audio. Now Video Op-Amp's can be hairy to work with but get it right, the results can be spectacular.
If we are really botherd by 18nV|/Hz there are also ways to get a quieter OPA (external LSK389 in Mil Spec case anyone?), but then a discrete solution (or a different Op-Amp) may be better.
I always wanted to try an AD811 fronted by 2SK146 as follower (with the neccesary resistor on the - input) for MC Phono, never found the time.
Ciao T
Sure, or there was parallel development (you did not put the second stage cascodes in). Looking at the schematic and with my 1980's EE Knowledge I'd call the design "blatantly obvious".
Yup, the LH0032 is basically Class B, nothing that judicious application of a resistor to the negative rail (then we end up with a Kaneda) or an external Follower cannot fix.
Sure. Now what was the noise of the TL082 so beloved by many HiFi Designers? Was it 18nV|/Hz at 1KHz? That is about even stevens. A little on the high side for MM maybe, I get maybe 75dBA post RIAA, guesstimate...
Might have, 1mA output stage current is a bit measily. But if we wanted to use this suckah in audio it may be up to us as the circuit designers to do something about it...
Precisely, I have yet to find any "Audio" Op-Amp that I actually like the sound when used for Audio. Now Video Op-Amp's can be hairy to work with but get it right, the results can be spectacular.
If we are really botherd by 18nV|/Hz there are also ways to get a quieter OPA (external LSK389 in Mil Spec case anyone?), but then a discrete solution (or a different Op-Amp) may be better.
I always wanted to try an AD811 fronted by 2SK146 as follower (with the neccesary resistor on the - input) for MC Phono, never found the time.
Ciao T
Hi,
Hawkesford Cascodes return (as Joachim already commented) to the emitters of the cascoded transistors.
It may also worthwhile to look at different transistors for the Hawekesford Cascode (low Capacitance ones) and the folded cascode (very linear, high beta, high gm ones).
Should be possible to tweak the model, not sure with LT-Spice, but easy with P-Spice...
Good MM and high output MC Level.
I expect more distortion with real devices, but still, looking good.
Yes, another 20dB Gain should not be hard. Even Open loop. Probably can use a much simplified design.
Alternatively, tripple up the 2SK170's, this will give around 10dB more gain and lower noise, the rest of the gain you can get by upping the RIAA Impedance around threefold.
Ciao T
Hawkesford Cascodes return (as Joachim already commented) to the emitters of the cascoded transistors.
It may also worthwhile to look at different transistors for the Hawekesford Cascode (low Capacitance ones) and the folded cascode (very linear, high beta, high gm ones).
Should be possible to tweak the model, not sure with LT-Spice, but easy with P-Spice...
Good MM and high output MC Level.
I expect more distortion with real devices, but still, looking good.
Yes, another 20dB Gain should not be hard. Even Open loop. Probably can use a much simplified design.
Alternatively, tripple up the 2SK170's, this will give around 10dB more gain and lower noise, the rest of the gain you can get by upping the RIAA Impedance around threefold.
Ciao T
Now gain is at 56 db...@ 1KHz
Have a little problems arround the CCS for the jfets...thinking about cascoding a high Idss jeft...would like to have more current trough the Jfets...than here running at 2.6 mA..would like app 5...
Had to change the cascode an current mirror bjt's as SOA came into play..
Have a little problems arround the CCS for the jfets...thinking about cascoding a high Idss jeft...would like to have more current trough the Jfets...than here running at 2.6 mA..would like app 5...
Had to change the cascode an current mirror bjt's as SOA came into play..
Have you tried?
I'm sure you can find a way to tune the offset, but in reality, this circuit's offset will wander all over the place with temperature, time of day and phases of the moon.
So, unless you use something like a servo, IN PRACTISE this can not be realised.
jan didden
Have a similar input stage in my curren riaa and it stays relatively stable even without a servo. I will offset it to deliver around -2 volts and see how much it varies. Report will follow
//A
Been playing a little with the circuit..
One thing that springs to mind with this type of transconductance Riaa is how the distortion changes with frequency...or with the riaa slope.. as frequency increases amplitude decreases.. and the same does distortion...So while in the deep base at 40 Hz .. the distortion is app -60dB.. then at 1 Kz the distortion is way lower at -80 db..and at 10 Khz...the distortion is as low as...-90dB..!!
In IRL distortion figures are surely higher...the main the point is the decreasing distortion with the rise in frequency...
This is quite opposite of normal voltage types where distortion normally gets worse as frequency rises..but at the same time the natural way of distortion with various frequencies...
For sure a very strong concept...some polishing left...and a quick build and listen...
One thing that springs to mind with this type of transconductance Riaa is how the distortion changes with frequency...or with the riaa slope.. as frequency increases amplitude decreases.. and the same does distortion...So while in the deep base at 40 Hz .. the distortion is app -60dB.. then at 1 Kz the distortion is way lower at -80 db..and at 10 Khz...the distortion is as low as...-90dB..!!
In IRL distortion figures are surely higher...the main the point is the decreasing distortion with the rise in frequency...
This is quite opposite of normal voltage types where distortion normally gets worse as frequency rises..but at the same time the natural way of distortion with various frequencies...
For sure a very strong concept...some polishing left...and a quick build and listen...
Hi,
Probably can up another 4dB safely.
How about BJT CCS (you can use the base voltage for Q1/7 I should think), using again the Hawkesford cascode? And I would put the Hawkesford Cascodes back on the folded Cascode transistors...
You can probably drop the Id on the 2SK170 to 3mA each...
Ciao T
Probably can up another 4dB safely.
How about BJT CCS (you can use the base voltage for Q1/7 I should think), using again the Hawkesford cascode? And I would put the Hawkesford Cascodes back on the folded Cascode transistors...
You can probably drop the Id on the 2SK170 to 3mA each...
Ciao T
John,
Works for me, I have a reel. But potentially harder to source for others...
Ciao T
Works for me, I have a reel. But potentially harder to source for others...
Ciao T
Hi,
Yes, really pretty good.
I would probably make the CCS a BJT one, basically repeat Q1/R4 instead of the two J-Fets, also maybe tie the two more LED's for the cascode of the frontend CCS into the the same Bias chain with R6...
For R3 maybe use a 51 Ohm resistor and a 100R trimmer in series, to reduce the range of adjustment.
For the RIAA, maybe settle on capacitors with standard E6 Values, for ease of availability and use parallel combinations of resistors to tune the RIAA precisely, E24 or even E96 resistors are easy to find, but even E12 high quality capacitors tend to be custom/special order items with high MOQ's...
There is a different structure of RIAA EQ that can use 4pcs identical value capacitors, may be worth looking at...
(Image Courtesy of Paul Kembles RIAA Page)
I am (obviously) talking about the middle one. Suitably scaled (4.7nF Cap's maybe) it should do the trick.
Ciao T
Yes, really pretty good.
I would probably make the CCS a BJT one, basically repeat Q1/R4 instead of the two J-Fets, also maybe tie the two more LED's for the cascode of the frontend CCS into the the same Bias chain with R6...
For R3 maybe use a 51 Ohm resistor and a 100R trimmer in series, to reduce the range of adjustment.
For the RIAA, maybe settle on capacitors with standard E6 Values, for ease of availability and use parallel combinations of resistors to tune the RIAA precisely, E24 or even E96 resistors are easy to find, but even E12 high quality capacitors tend to be custom/special order items with high MOQ's...
There is a different structure of RIAA EQ that can use 4pcs identical value capacitors, may be worth looking at...
(Image Courtesy of Paul Kembles RIAA Page)
I am (obviously) talking about the middle one. Suitably scaled (4.7nF Cap's maybe) it should do the trick.
Ciao T
Folks,
Yup, zere iz natrually a problem wiz Anglefire. Zo here ze schematic zcree zcraped...
Ciao T
There is a different structure of RIAA EQ that can use 4pcs identical value capacitors, may be worth looking at...
(Image Courtesy of Paul Kembles RIAA Page)
I am (obviously) talking about the middle one. Suitably scaled (4.7nF Cap's maybe) it should do the trick.
Yup, zere iz natrually a problem wiz Anglefire. Zo here ze schematic zcree zcraped...
Ciao T
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.