Laboratory Page Memphis MC 1
Active or passive Equalization Passing a sine wave through a passive RIAA filter will reduce its amplitude but preserves its exact form. A feed- back loop within an active RIAA filtering takes the sine wave from the output back to the input. Thus it is more prone to distortion due to setting times, slew rates, input / output impedances and other variables. Of course there are top notch amplifiers with active RIAA filtering but they require pretty sophisticated engineering. For designing DIY projects passive RIAA filtering is substantially easier to handle IMHO. RIAA Equalization and Simulation Several web sites with active or passive RIAA network calculators can be found on the internet. Mh-audio, KAB, hagtech and others provide calculators for passive RIAA networks. Obviously they all are based on the equations published by Stanley Lipshitz in the 6/1979 Journal of the Audio Engineering Society and at least these calculators are a good point to begin with. A convenient way to evaluate the RIAA equalization in computer based simulations is to feed the virtual phono signal through an inverse RIAA filter into the circuit. The resulting frequency response should be flat within the tolerances allowed by the designer. Searching the internet I found an interesting inverse RIAA filter (fig. 1) on the hifisonix web site. I simulated the inverse RIAA equalization using LT spice and evaluated its accuracy by comparing the results from the simulation (numbers in red) with the ideal RIAA response (tab. 1). Differences are between 0.01db to 0.03db between 10Hz and 20kHz which might be considered sufficient. Thus I used this circuit as the anti-RIAA filter to analyze and adjust the RIAA response of the Memphis MC1 phono stage (fig. 2). Fig. 2: Phono stage with inverse RIAA filter Setting the values for the RIAA filter The Memphis MC1 RIAA filter is fed from a current source (transistor Q7) with L6 and R14 in parallel being the load. A basic problem of this configuration is the limited low frequency response due to the choke. The inductance of the LL1667 is 810H (copper resistance 2.4kOhm). The figures 3 to 7 show a simplified simulation of the low frequency roll-off (-1db) with different standard resistor values of 100k, 47k, 22k and 10k in parallel to the choke. As expected, the lower the resistor value, the lower the frequency at the -1db roll-off (39Hz @ 100k, 4.7Hz @ 10k). On the other hand a lower resistance also decreases the gain (0db @ 100k, -20db @10k). As a compromise I accepted the -1db roll off @ 9Hz and chose a resistor around 22k as the basis for the further RIAA filter calculations. Figure 8 shows the simulated and the measured RIAA responses.     Fig. 3 to 7: Simulation of the low frequency roll-off with R1 = 100k (4), R1 = 47k (5), R1 = 22k (6) and R1 = 10k (7) Fig. 8: Simulated (red) and measured (blue) deviations from the ideal RIAA response (db) between 10Hz and 40kHz with MC input transformers and output capacitor included.
-0.004      -2.60     -8.18      -15.25 -17.14   -19.59   -25.53 19.71 19.24    16.92     13.06      8.19    2.63 Fig. 1: Precision inverse RIAA filter to evaluate, adjust or design RIAA equalization networks in phono stages.  Tab. 1: Ideal RIAA response (black numbers) and simulated inverse RIAA response (red numbers) from LT spice simulation (fig. 1).
R = 100k V = 0db @ 1kHz V - 1db @ 39 Hz
R = 47k V = -6.6db @ 1Khz V - 1db @ 19 Hz
R = 22k V = -13db @ 1kHz V - 1db @ 9 Hz
R = 10k V =  -20db @ 1kHz V - 1db @ 4.7Hz  Hz
4
7
5
6