Hi,
Do anybody have suggestion for alternative filter component values for the Pearl? I guess every component needs to be changed in order to keep the same alignment. I am willing to play with the source resistance of the input stage as well if that helps, but would prefer standard values there as well (since I am using Holco's).
Current values are not standard (caps). Unfortunately I am inexperienced in this type of filter design and my current references out of Audio Amateur and elsewhere all use much smaller caps and larger resistors (which could be scaled I guess but I still need to match up to preceding and following stage).
I have solved this "problem" by making a PCB with room for parallell caps (mounted on both sides to conserve space as high-end caps are large and come in limited value families). However, I am posting this as I would feel more comfortable using single high-end caps as a first cut.
((((Current design uses 8 input JFETs and 4 output JFETS, all the same (input) type as well as same value drain/gain resistors all the way through for higher gain))))
Petter
Do anybody have suggestion for alternative filter component values for the Pearl? I guess every component needs to be changed in order to keep the same alignment. I am willing to play with the source resistance of the input stage as well if that helps, but would prefer standard values there as well (since I am using Holco's).
Current values are not standard (caps). Unfortunately I am inexperienced in this type of filter design and my current references out of Audio Amateur and elsewhere all use much smaller caps and larger resistors (which could be scaled I guess but I still need to match up to preceding and following stage).
I have solved this "problem" by making a PCB with room for parallell caps (mounted on both sides to conserve space as high-end caps are large and come in limited value families). However, I am posting this as I would feel more comfortable using single high-end caps as a first cut.
((((Current design uses 8 input JFETs and 4 output JFETS, all the same (input) type as well as same value drain/gain resistors all the way through for higher gain))))
Petter
Hello Petter,
This may or not be usefull to your query,
Douglas Self, with his pre-amp desgn in "Electronics World" magazine (UK) in the Mid 90's pointed out an "interesting statistical trick". I can't remember, or have access to the original article, but the argument goea a little like this...
If you parallel two capacitors (or whatever components you like) which have a 1% tolerance, then you end up with an actual capacitance whose tolerance is better than 1%. This is due to the normal dristributian of the tolerances.
In his design he paralleled around 5 or 6 standard values to get the value he wanted, but at much higher precision as the individual capacitors.
Hope this helps
Adrian
P.S. Are you the same Petter who lives in Stavangaar and who sent me and another New Zealander a number of copies of Audio Amatuer in 1998?
This may or not be usefull to your query,
Douglas Self, with his pre-amp desgn in "Electronics World" magazine (UK) in the Mid 90's pointed out an "interesting statistical trick". I can't remember, or have access to the original article, but the argument goea a little like this...
If you parallel two capacitors (or whatever components you like) which have a 1% tolerance, then you end up with an actual capacitance whose tolerance is better than 1%. This is due to the normal dristributian of the tolerances.
In his design he paralleled around 5 or 6 standard values to get the value he wanted, but at much higher precision as the individual capacitors.
Hope this helps
Adrian
P.S. Are you the same Petter who lives in Stavangaar and who sent me and another New Zealander a number of copies of Audio Amatuer in 1998?
passive RIAA-Network formulas
Hi Petter,
here are some equations taken from: "S.P. Lipshitz: On RIAA Equalization Networks" in Journal Of The Audio Eng. Society, Vol27, No.6, 1979. Refering to the schematics in Fig1 in the Pearl article:
R23*C10=750µs
R22*C13=318µs
R23/R22=6.877
C13/C10=2.916
Note that the value for R23 should contain the source resistance (499Ohm?) for the calculation. One of the Components can be choosen independently.
C11, C12 and R24 add a further rolloff at 8Hz. I dont know why they paralled two Caps here for this unimportant pole.
The values given in the Pearl-article are not 100% consistent with the formulas shown above. You will get the exact values for C10 if you use R22=989 and R23=6250Ohm. In this case C13 should be 350nF. Maybe here is sonmething messed or i dont understand the circuit right. The latter is more supposable. Hope that hepls.
joe
Hi Petter,
here are some equations taken from: "S.P. Lipshitz: On RIAA Equalization Networks" in Journal Of The Audio Eng. Society, Vol27, No.6, 1979. Refering to the schematics in Fig1 in the Pearl article:
R23*C10=750µs
R22*C13=318µs
R23/R22=6.877
C13/C10=2.916
Note that the value for R23 should contain the source resistance (499Ohm?) for the calculation. One of the Components can be choosen independently.
C11, C12 and R24 add a further rolloff at 8Hz. I dont know why they paralled two Caps here for this unimportant pole.
The values given in the Pearl-article are not 100% consistent with the formulas shown above. You will get the exact values for C10 if you use R22=989 and R23=6250Ohm. In this case C13 should be 350nF. Maybe here is sonmething messed or i dont understand the circuit right. The latter is more supposable. Hope that hepls.
joe
Pearl
The cap and resistor values are all interelated. The resistors are all standard 1% values. These values don't exactly match the Riaa curve but get very close. Compared to rooms and speakers .1 dB is not much. The 350 nF stated by Petter is indeed closer and the very first one used this value because it was a 330 nF that was high in value. The current caps are 2% Panasonics and easy to get.
The Values for resistors and caps were used because it keeps the impedance and noise down.
Wayne@passlabs.com
The cap and resistor values are all interelated. The resistors are all standard 1% values. These values don't exactly match the Riaa curve but get very close. Compared to rooms and speakers .1 dB is not much. The 350 nF stated by Petter is indeed closer and the very first one used this value because it was a 330 nF that was high in value. The current caps are 2% Panasonics and easy to get.
The Values for resistors and caps were used because it keeps the impedance and noise down.
Wayne@passlabs.com
HI
What are the non interrelated values for C13, C10 and C11 + C12 ?
I am asking because I dont get those Panasonics easylie and so I want to select and match the caps anyway. If i know the values, I can chose the caps which are closest to them.
Bernhard
What are the non interrelated values for C13, C10 and C11 + C12 ?
I am asking because I dont get those Panasonics easylie and so I want to select and match the caps anyway. If i know the values, I can chose the caps which are closest to them.
Bernhard
Re: Pearl
Wayne,
Maybe you can help me with a small issue. The very bottom end response is rolling off early. The cap value in the RIAA were matched to 0.1% along with the resistors.
The output coupling cap is 22 ufd. But the low end starts to droop below 80 Hz. The load is 23K.
Can I diddle with the subsonic filter to low it to 4Hz instead of 8Hz? The front in feeding the Pearl is not world class, but is more than good enough.
Thanks,
George
wayne said:
Wayne,
Maybe you can help me with a small issue. The very bottom end response is rolling off early. The cap value in the RIAA were matched to 0.1% along with the resistors.
The output coupling cap is 22 ufd. But the low end starts to droop below 80 Hz. The load is 23K.
Can I diddle with the subsonic filter to low it to 4Hz instead of 8Hz? The front in feeding the Pearl is not world class, but is more than good enough.
Thanks,
George
- Status
- Not open for further replies.