Passive Composite RIAA Equalisation Question

[andyr]

A typical passive RIAA network is shown on p3 of this interesting paper:
Phono Stage Design

I have modified a phono stage with the help of this paper but in measuring it, I find it has a bit too much roll-off from 2Khz upwards. (So, somehow I have not applied the equations correctly.)

If I wish to boost the HF from 2Khz upwards, slightly:
* which component value do I change - C1, C2 or R2?
* do I increase it or decrease it?

Thanks,

Andy

 

[DF96]

You can fiddle by ear (or trial and error) with an RIAA network but I would not recommend it. The figures quoted in the page you referenced seem correct according to the classic paper by Lipshitz, yet to me they look wrong. For example, I would expect R1'/R2=9 so you get a 20dB shelf cut in the LF. However, I realise that these networks can be counter-intuitive because of interactions so I may be wrong.

At this stage I would assume that Lipshitz is right, and I am wrong. So recalculate your values according to the formulas you have. Otherwise, to boost above 2kHz reduce the value of C2 but bear in mind this will have some effect at all frequencies.

 

[SY]

There's an ambiguity in the schematic. If you use 475k as the series resistor (the schematic indicates two different values), the RIAA is reasonably close.

 

[wrenchone]

What you also have to remember is that unless you are using a fairly robust follower to drive the passive RIAA network, you need to throw the impedance of the first stage into the pot when you calculate your resistance values. The impedance of the fist stage forms a part of the initial resistor feeding the network. It's the reason that for a lot of the jfet preamps I've built, I interpose a buffer stage between the first stage and the EQ network, even though it adds another stage/more parts - one is also always sure of the values. I've used the paper cited here as the basis for calculating the values for my EQ networks. Someone please chime in with a correction if this guy has got the equations wrong.

 

[andyr]

Thank you for your answers, guys.

DF96, reducing C2 to boost values above 2Khz was the answer I was hoping to get. And I suspected doing this would have an effect below 2K as well ... damn! So I will try it in a couple of weeks with several reduced values of C2 and measure the output at 2Khz, 10Kz, 15Khz & 20Khz ... and then at 1Khz, 500Hz & 100Hz. Hopefully, I can boost the HF without doing too much damage to the LF.

SY - yes, there is an ambiguity in the paper. Simon told me I was the first person to have pointed this out to him, when I emailed him about the paper, a couple of years ago. His response was:

"NOTE: R1 and R0 in the schematic have 2 values each.....use EITHER 475K with 1M0 OR 359K with 592K."

should instead read:

"NOTE: R1 and R0 in the schematic have 2 values each.....use EITHER 475K with 592K, or 359K with 1M0."

475K in parallel with 592K = 263K543
359K in parallel with 1M0 = 264K165."

Wrenchone, yes, Simon also confirmed you have to allow for the Zout of the first stage (it is very low in his tube-based phono stage, so can be ignored). So in the JFET circuit I am using, 51K has to be added to R1 (the Zout of the input pair of JFETs is set by the value of the Drain resistor).

What also has to be done is take account of the Zin of the second JFET stage as a parallel resistor with R0. I assumed Zin was 15Meg. So we have to calculate R0' before using this value to calculate R1'.

The situation I now find myself is that, having used Simon's equations to calculate the RIAA values, I have a close approximation to the RIAA curve below 2Khz (mostly within +0.1dB) but above 2Khz, I get a gradual dip from -0.2dB to -1.2dB @ 20Khz. So I'm hoping I can fix this!

Regards,

Andy

 

[SY]

Andy, this is controlled by the 2n7 cap paralleled with 150pF. It needs to be trimmed to be a bit smaller (that's not easy! It will take a couple of new caps). I suspect that the Miller capacitance in the next stage is the culprit.

 

[andyr]

Andy, this is controlled by the 2n7 cap paralleled with 150pF. It needs to be trimmed to be a bit smaller (that's not easy! It will take a couple of new caps). I suspect that the Miller capacitance in the next stage is the culprit.

Could be, SY ... except that someone (an amp designer mate) told me Miller capacitance was not an issue with this JFET circuit?? I will see if I can post the schematic - I'd be very interested in your input.

Regards,

Andy

 

[DF96]

Unless the next stage is a follower (unlikely?), it will have some Miller effect. Even a cascode has a tiny Miller effect (it doubles the stray capacitance).

 

[wrenchone]

The impedance values seem rather high for a fet-based preamp circuit. I'm generally using 1/10 the impedance, with 100nF as my first capacitor and 34nF for the second. With the higher impedance, strays are going to affect the second break point, especially when Mr. Miller gets involved. This is especially true for fets like the 2SK170 and similar devices that have highish capacitance to begin with. It's one of the sneaky reasons why I use fets like the PN4393 and PN4303 - they both have much lower gate and feedback capacitance.

 

[andyr]

Andy, this is controlled by the 2n7 cap paralleled with 150pF. It needs to be trimmed to be a bit smaller (that's not easy! It will take a couple of new caps). I suspect that the Miller capacitance in the next stage is the culprit.

SY, I have drawn the schematic (in Powerpoint, as I don't have an electronics workbench).

Hopefully it will come out in the post.

Interested in your comments. As I said before, my next task is to snip a capacitor or two off the C2 bundle and re-measure the RIAA, to see if I've been able to boost the HF.

(Currently, there is a gradual droop from -0.2dB @ 2Khz to -1.2dB @ 20Khz ... which is what I want to try and correct by reducing C2.)

Regards,

Andy