A MM Phono Amplifier
It is a lot of variants of the design Le Pacific (by Walter), one reprint is shown at the end of this description. The following design is my variant of his simple design for the amplification (and equalization) of signals from moving magnet (MM) cartridges. In the figure below is shown a simplified circuit diagram of my version.
It is a two stage amplifier, the first stage built around the JFET Q104 and the second stage built around Q116. The resistors R103 and R115 are giving a reduction in the drain current and linearising the amplifier stages. They are adding some noise, but since their value is low, it has no impact on the final signal to noise ratio.
In this kind of amplifier the choice of JFETs are important. I have a bunch of 2SK170V transistors, so these have been used here. The saturation current (IDSS) of these transistors was selected to be about 12 mA. This is reduced to some 5,5 mA by R103/R115. This gives a transconductance (gm) of about 33 mA/V. The gain is then about:
Essentially the components R108-C112 implements the inverse RIAA characteristics. However, R107 acts as a source resistor for the network and R114 as a load. This makes the calculations somewhat less straightforward, but measurements confirm that the accuracy is good with the values shown.
The resistor R102 makes up the resistive loading for the cartridge. The capacitor C101 is shown to be 100 pF in the schematic. Together with the equivalent capacitance from the gain stage around Q104 this makes up the capacitive loading for the cartridge. However, this is quite high, about 200 pF, so even without C101, this capacitive loading is too much for most cartridges. The phono cable capacitance will also add to the above.
To reduce the input capacitance, a cascode comes to help. This is shown in the final schematic below, where the transistor Q106 is the common base transistor. The base voltage is 6.8 V, thus reducing the drain-source voltage for the JFET to about 6 V. The cascode makes the input capacitance exclusive C101 about 25 pF. The necessary cartridge loading can then be chosen by simply selecting a proper value for this capacitor.
At the output it is added a buffer around Q118 and Q119. It has no voltage gain, but reduces the output impedance from 2.4 kohm to about 35 ohm. It is strictly not necessary, but can be used for long cable connection and/or low input impedance in the following stage.
Further-on it is shown an RC-filtering of the 24 V power supply. It ensures that the amplifier has a very quiet power supply. The power supply section itself is shown in the next figure.
It delivers 24 V to the two stages
and the additional 6.8 V to the cascode transistor Q 106 . An
unregulated DC voltage of about 32 V (V+ and V−) is fed to the power
supply with a 10 ohm resistor to ground. This resistor applies an
additional RC filtering. The network around Q134 is a capacitance
multiplier and removes most of the ripple and noise from the applied DC
voltage. Two zener diodes D136 and D137 make a total voltage of 24.8 V
on the base of the transistor Q140. The zener diode D137 also supply
6.8 V to the cascode. It is used several 2200 μF capacitors for the
filtering. Thus, both the 6.8 V voltage and the 24 V voltage is very
clean. The only drawback with all this filtering is a slow start-up for
power-on. A two layer PCB has been made for the amplifier with the
power supply. It is laid out for two channels, please look at the
picture on the next page.
The two channels are identical, the
component numbering is 100-etc for one channel and 200-etc for the
other channel. The power is fed to J 146 /J 246 for V+ and J 147 /J 247
for V-. The signal input is J 142 /J 242 and the input signal ground
(SG) is J 143 /J 243 . The signal output is J 144 /J 244 and the output
ground is J 145 /J 245 . J 141 /J 241 is an extra ground connection.
The bottom side, component side and the component
placement is shown on the next pages. If you are interested, the PCB and/or the Gerber files are available. As mentioned, the JFETs is 2SK170 from Toshiba, which is not in production any more. Therefore an alternative is LSK170C, produced by the US company Linear Systems. Another alternative is BF862 from NXP, but this is a surface mount component. The cascode transistor was 2SC1815, which was at hand, but can be replaced by any low noise NPN transistor. But if my layout is used, be aware of the pinning. The two BD139 in the power supply may also be replaced by other types, and these transistors are not critical. The two zeners are ordinary 0.5 W types.
The gain at 1 kHz is 100x (40 dB), so a 5 mV cartridge will give an output of 500 mV. There is an endless list of tweaking possibilities for this amplifier, so it is very difficult to give some guidelines. I have limited my measurements to only THD and RIAA accuracy. There is low distortion for an input signal up to 50 mV. The RIAA accuracy was very good since I have used only components with 1 % tolerance for the resistors and plastic capacitors. The unregulated DC power supply is kept at a good distance from the amplifier. This ensures that this amplifier is very silent. This is also because the PCB is built into an Al enclosure. Finally, I must say, the reputation of this amplifier is understandable. It sounds remarkably neutral and good sounding despite its simplicity. Yes, it sounds better than amplifiers with many more transistors. Or, maybe it is the low number of transistors thatmake this amplifier sound great?
Knut Harald Nygaard