[gurevise]

Hello Jim,
What is the approximate accuracy of RIAA EQ with specified types of resistors and capacitors? I have used them without matching values.
Some people go to great length trying to match all values exactly.
I'd like to know the worst case scenario of deviation from RIAA EQ curves without component matching.
May be some suggestions would be welcome on what specific components in EQ circuits have to be more accurate.

Thanks
Sergey
Dayton, OH

[eico1]

I can recommend to start by ordering this paper from aes.org:

New Factors in Phonograph Preamplifier Design
Vol. 24, Number 4 pp. 263 (1976)
Author:   Tomlinson Holman
Abstract:  An investigation into why various phonograph preamplifiers sound different has produced some new measurement techniques which correlate well with subjectively observed differences. A preamplifier design which performs well on these new tests as well as

As I recall it has equations for the typical eq topologies. At the time I wrote a basic program to pick resistors and caps based on standard values for minimum error.

steve

[hagtech]

I did a study on RIAA accuracy using SPICE several years ago.  The following plot shows what happens when you modify capacitor values by +/-5%, with resistor values at nominal.



It's the middle two time constants that are more important, the 318us and 75us.  These will affect your transient and phase response more than the others.  Placing the 3180us pole can really be sloppy.  As you can see, a 5% tolerance leads to a +/-0.4dB error.  Not exactly tragic.  Especially when compared to the response of your loudpeakers or room effects.

I believe it is more important to match channels than it is to match to an absolute.  It does help a little to make sure components in both channels are reasonably close in value.  But it's easy for some people to go overboard here.  Your time and money are better spent on tubes.

jh