Simplistic NJFET RIAA

diyAudio Chief Moderator
Joined 2002
Paid Member
The first stage on my schematic has 11k Zout (Ro=Rriaa(V1GOhmload-VoRIAAload)/Vo. If I go 50xtimes that for series RIAA resistor, then the RIAA capacitors will be fictionally small and bad for tolerances or quality dielectric availability. If the Zout is embedded in the RIAA, where is the difficulty? Q2 with 2.2Meg input is 50x the 47k combined driving impedance. Will it get seriously loaded?

P.S. Which schematic is your tube RIAA? Can you post it? So I can see why you had to do those mods. If its not proprietary of course.
 
Administrator
Joined 2004
Paid Member
jackinnj said:
Kevin posted triodeguy's (John Sears) RIAA -- I tweaked it a bit with my "LaPlace-i-nator". The 2N3819 is a fraction of the cost of a 2SK170, I haven't measured the noise, however.

An externally hosted image should be here but it was not working when we last tested it.


Richard Sears.... Otherwise cool work..

IMO The 2N3819 might prove to be just a bit noisy in this application.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Zoran said:
This is good concept...
..
Use passive inverse riaa network for simulation
not only straght signal...
try haggerman inverse riaa
cheers
You will be amazed

Hi Zoran. Did the Bode plot driven from Hagerman inverse. I added the ''lost constant'' too.:D
Well it looks doable, wasn't off. I also sized the source resistors bypass capacitors more generously for better low bass phase shift. I will use 500R trimmers there. That and matched 2SK170BLs should do the channel matching trick.
Hope that it will remotely resemble the gain and time constants in real life test!:smash:
 

Attachments

  • frriaa.gif
    frriaa.gif
    6.1 KB · Views: 3,998
diyAudio Chief Moderator
Joined 2002
Paid Member
P.S. The plot above appeared as a joke when uploaded. :xeye:
Info on its scaling:
It has 10Hz to 100kHz limits, its 20Hz and 20khz are -0.4dB. Its octave spaced per main lines horizontally and 0.1dB per vertical dent.

And the ''promissing'' circuit. Shown driven from a Hagerman inverse RIAA with 60dB attenuation of input generator.
 

Attachments

  • inversedriven.gif
    inversedriven.gif
    9.7 KB · Views: 4,230
diyAudio Chief Moderator
Joined 2002
Paid Member
I would terminate with 560R to 1K2 if I had a 60R Zo cartridge. The Hagerman assumes 60R Zo. That is arbitrary. We are free to choose how to terminate any given cartridge. I can load it down with 100R and sim for less effective input voltage (the 0.25mV will halve with 60R Zo on 100R load) and 150pF input cable parasitic. The output I have tested low before, I have a buffer for that, no worries. The 1 Meg is there to get shunted by loads, but to keep the output cap static happy when not.
Anyway, here it is as you wanted to see it. Still solid. :) ...In theory.:cool:

*For anybody who might be attempting this before I get a chance, be sure to remember that the gain is based on 10-12mA Idss 2SK170BL @ 9V
 

Attachments

  • screenshot1.gif
    screenshot1.gif
    17.4 KB · Views: 4,724
salas said:
Different software, and models probably. Show your best Bode and RIAA values. In case it deviates in practice we should be having a plan B.;)

I use LaPlace transform with values of 3180, 318 and 75uS -- Walt Jung shows how to do this in his text "Op Amp Handbook". What I saw with R17 was a pretty serious peaking in the ultrasonic region. If you hunt around on the ADI website you can find that chapters of the book are archived.

Dennis Colin in his article "Tweaking the Passive Inverse RIAA Network" assumes the 3.18 uS compensation (Audio Xpress August 2007, a PDF can be found on the AX website http://www.audioxpress.com/magsdirx/ax/addenda/media/colin2808.pdf --

You might want to adjust the values of the Inverse RIAA net you are using, Denis has them acurate for a micro-deci-Bel ! But regardless, the values are going to be different than those derived using the LaPlace transform.
 
LaPlace --

For the enhanced RIAA include in the numerator 3.18 uS

.PARAM T1 = {3180}
.PARAM T2 = {318}
.PARAM T3 = {75}
.PARAM T4 = {3.18} ;Reference Enhanced RIAA Constant
ERIAA 5 0 LAPLACE {ENORM*V(1)}={((1+(T2*1E-6)*S)*(1+T4*1E-6)*S)))/((1+(T1*1E-6)*S)*(1+(T3*1E-6)*S))}

Edit -- there appears to be one wrong number in Table 1 of the article.