Now, how can I proceed? Trying a BC550C cascode to even lower the input cap or...
This is addictive and fun.![]()
Nice that you have found the FFT on view menu after clicking on analysis window and the proper input settings. Can't tell you how to better patch a Mad's Pacific, I never made any Maison's or Mad's Pacific. Your LT sim will let you try things I suppose. Its FFT looks optimistic at 0.25% 2ndH and suspiciously enough 3rdH VS 0.6% almost only 2ndH you measure non the less. Or you measure more 3rdH now too? As I wrote, the Simplistic gives repeatable results at 0.02% ballpark, or even 0.01% in my LowMC attached. You can look at my CCTs link I gave you before, for ideas.
Attachments
1. Local feedback, gain control. 2. SE stands for single ended i.e. non symmetrical, not source to earth. 3. Weird circuit better don't make hit or miss changes. 4. Reduces bandwidth for some reason. Prone to oscillation with its BJT buffer drive could be.
Hi Salas,
I am hoping you can answer a couple of further questions I have.
There have been several interpretations of the original "Pacific" RIAA circuit shown in this thread but they all seem to have the following:
* Rdrain of 2-3K ohms (with a 24v DC supply).
* the first (series) resistor of the RIAA network (R1) is 22K-33K ohms.
Q1: I understand the Drain resistor sets the output impedance of the input JFET?
Q2: What would the effect be of having R1 at 220K?
I know this will cause the RIAA component values to change - what I am interested in knowing is ... is it a better/worse load for that input JFET than 22K/33K? (And why.)
Thanks,
Andy
Last edited:
A1: Yes it sets it.
A2: It would be an easier but noisier load for the first stage and a high source impedance for driving the second stage. Overall worse.
Now if we take the example of the circuit you made, its driving 240K from 51K and 240K then driving 510K. Noisy Riaa & bad drive of EQ circuit, and 2nd stage.
A2: It would be an easier but noisier load for the first stage and a high source impedance for driving the second stage. Overall worse.
Now if we take the example of the circuit you made, its driving 240K from 51K and 240K then driving 510K. Noisy Riaa & bad drive of EQ circuit, and 2nd stage.
A1: Yes it sets it.
A2: It would be an easier but noisier load for the first stage and a high source impedance for driving the second stage. Overall worse.
Now if we take the example of the circuit you made, its driving 240K from 51K and 240K then driving 510K. Noisy Riaa & bad drive of EQ circuit, and 2nd stage.
Thank you very much, Salas (you're up early?).
Re. "my" circuit - as I understand it, because they used an Rdrain of 51K they needed to put 240K at R1.
But If I have, say, somewhere between 2K & 3K as the Drain resistor, I could use 33K for R1, right? As you say, this will drive 510K much better.
(I have to play with this value next weekend; I have temporarily mounted a Bourne PCB pot as the Drain resistor and will work out the value needed to give me a voltage difference of 8v across the JFET (Drain to Source). Rsource will be 33 ohms (not 51K!!).)
Regards,
Andy
Don't know why it was done like that in your phono, its odd. Could be having high Vgs(off). Those Rs are bypassed in yours and you must consider the new values that will bypass 33R to infrasonic, also a new Riaa filter calculation based on Rd1stStage+R1. If Rs will be left bypassed you lose 6dB. You got to know your Fet's bias curves to know what you are changing. I would make a different phono from scratch if I was in your shoes. Why you wanna change it, does not satisfy? Make the Mad's or the Mad's based one Radioman is going to end up with, which he is going to document, for instance.
Don't know why it was done like that in your phono, its odd. Could be having high Vgs(off). Those Rs are bypassed in yours and you must consider the new values that will bypass 33R to infrasonic, also a new Riaa filter calculation based on Rd1stStage+R1. If Rs will be left bypassed you lose 6dB. You got to know your Fet's bias curves to know what you are changing. I would make a different phono from scratch if I was in your shoes. Why you wanna change it, does not satisfy? Make the Mad's or the Mad's based one Radioman is going to end up with, which he is going to document, for instance.
Thanks, Salas,
All good points.
But are you saying that because I will now have 33 ohms for Rsource, I need to change the 100uF caps?
Yes, I know I will need to calculate new RIAA values - however, my understanding is that these are based on Zout of the JFET stage + R1 - which is Rdrain + R1 (not Rs)? I have the filter calculations from this very valuable paper by Simon Shilton:
Phono Stage Design
Why am I changing the circuit? Because after building it and listening to it (it was recommended to me by someone who said "This is the best phono stage I have ever heard!
So I want to finish it - although from reading this thread, I understand that these are pretty deep waters I'm swimming in!
Thanks for your help, anyway.
Andy
The 100uF ARE the main Rs bypass caps. Not the 0.1uF. Those are ''quality'' enhancement to the main body of uF needed. You remove those, or bigger for lower Rs, you go 750mV/2 (-6dB).
Not good bass and highs is confirming the bad impedance relationships between stages and EQ network we discussed.
Finally, good luck, you gonna learn something in practice, and if it does not go, make something of related variety you can find around, like Mad's.
Not good bass and highs is confirming the bad impedance relationships between stages and EQ network we discussed.
Finally, good luck, you gonna learn something in practice, and if it does not go, make something of related variety you can find around, like Mad's.
The 100uF ARE the main Rs bypass caps. Not the 0.1uF. Those are ''quality'' enhancement to the main body of uF needed. You remove those, or bigger for lower Rs, you go 750mV/2 (-6dB).
Not good bass and highs is confirming the bad impedance relationships between stages and EQ network we discussed.
Finally, good luck, you gonna learn something in practice, and if it does not go, make something of related variety you can find around, like Mad's.
Efharisto, Salas.
Andy
Some of my amatere work
I have somewhat updated may homepages and particular the RIAA pages on my hobbysite DIY Audio.
There is an amateur test that doesn't say that much but is anyway comparable and was fun for myself to do. Please have a read on my DIY pages.
Maybe I strive some more with the pacific to find out, or let be..... The Q is to cascode or to not cascode.
![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
I have somewhat updated may homepages and particular the RIAA pages on my hobbysite DIY Audio.
There is an amateur test that doesn't say that much but is anyway comparable and was fun for myself to do. Please have a read on my DIY pages.
Maybe I strive some more with the pacific to find out, or let be..... The Q is to cascode or to not cascode.
Hello Salas,
I have done a bit more experimentation with my JFET-based RIAA circuit and have decided on RSOURCE and RDRAIN values for the JFETs, and also the RIAA network component values. So I have ordered some new parts and, hopefully, will be able to finish it off and measure it next weekend.![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
I would be very grateful if you could answer a question I have. The original circuit had RSOURCE of 51K with a bypass cap of 100uF. However, my new circuit will have a RSOURCE of 1K.
I understand this is an RC network that potentially rolls off the bass? Is the equation which gives the low-end -3dB frequency, the following?
Thanks,
Andy
I have done a bit more experimentation with my JFET-based RIAA circuit and have decided on RSOURCE and RDRAIN values for the JFETs, and also the RIAA network component values. So I have ordered some new parts and, hopefully, will be able to finish it off and measure it next weekend.
I would be very grateful if you could answer a question I have. The original circuit had RSOURCE of 51K with a bypass cap of 100uF. However, my new circuit will have a RSOURCE of 1K.
I understand this is an RC network that potentially rolls off the bass? Is the equation which gives the low-end -3dB frequency, the following?
F = 1 / (2*pi*R*C)
Thanks,
Andy
The bypass capacitor will roll off for a time constant indeed. Calculated like a 1st order high pass RC filter. 1000 Ohm and 100uF do fortunately give deep infrasonic turnover still.
Thank you for the info, Salas.
- Home
- Source & Line
- Analogue Source
- Ultrasimple MM/MC RIAA preamp 2