News / Articles

Long-Awaited Pearl3 DIY Phono Preamp is finally here!

Intro

Wayne Colburn of Pass Labs and PassDIY revealed to the DIY community his third iteration of the Pearl phono preamp at Burning Amp 2023. PCB boards and parts kit (above) are available for $150 at diyaudiostore.com. You'll find a comprehensive wealth of information about this product on the diyAudio forum.

Fig. 1

Circuit Review

Here are my initial thoughts of the circuit highlights (Fig. 1). The input is similar to the Pearl and Pearl2, utilizing four parallel low-noise JFETs with a cascode circuit and a voltage stabilizer circuit supplying clean power. This single-ended circuit is running off a 15V positive supply. Unlike previous generations, the amplified signal of the input stage is further amplified by an opamp U1A. A secondary function of this opamp is to lower the output impedance. The low impedance is required to drive the negative feedback from U1A to the input stage, which uses low resistor values to limit the voltage noise of the input stage. The selection of U1A is critical especially in the low-gain (MM) configuration, which loads U1A with 230 ohm in the feedback path in addition to the load of the next stage.

The coupling between the input stage and U1A requires a 1uF blocking capacitor that did not exist in the previous versions. The 1uF value assures a flat response down to 10Hz. It is possible to play with the values of resistors in the U1A circuit to cut down the size of the coupling capacitor, at the expense of marginally higher noise level. It's also possible to lower the value of C1 by sacrificing low-end extension. For example, a 0.33uF will give a flat response down to 30Hz and roll off below that.

The combined circuit gives a voltage gain of 27dB (MM) or 47dB (MC). The rest of the gain is implemented in the following output section of this phono preamp. I should note that the input circuit plus U1A have a theoretical open-loop gain of 83dB, which results in a hefty level of negative feedback in these stages, something that I consider less than optimal. Perhaps lowering R11 to cut down the gain of U1A might improve this circuit sonically, but in general, the new four-stage design has a lot of open-loop gain so it's inevitable that the level of negative feedback will be quite high. Whether it is better or worse compared to Pearl2 remains to be determined through extensive listening.

Opamp U1A drives the next stage opamp U1B with a passive partial EQ for the 2120Hz pole of the RIAA curve. Once again, the coupling between these stages requires another 1uF blocking capacitor. Opamp U1B implements the 50Hz pole, but this time it's in a feedback circuit. This is a smart distribution of the RIAA EQ, since low frequencies are sonically less sensitive to delays in the feedback loop. U1B drives a discrete push-pull output buffer Q6/Q7 that seems to be biased into class AB. A current source Q9 at the output of U1B biases the opamp into class A.

It is interesting that only two of the four stages in this design are biased into class A, with U1A opamp and Q6/Q7 buffer stage biased into class AB. One possible enhancement would be to add a small current source at the output of U1A to force it into class A operation and a minor modification for the Q6/Q7 circuit to do the same. The feedback in the output circuit is from the output of the buffer stage Q6/Q7 back to the input of U1B opamp. This feedback does not extend to DC, due to the addition of C8 and C25 in that path. This makes the circuit work also as a DC servo for the output stage which is also the output of the phono preamp. For a DC servo to work well, opamp U1B should be selected with the lowest DC offset.

In the actual PCB, U1A and U1B are two halves of a dual-opamp IC, putting a lot of demands on that IC. It should have relatively low-noise, high current drive and low input offset. It would be even better if it's biased internally into class A. I'm not sure that there is such an opamp IC in existence.

There are two notable features that were added in the Pearl3. The PCB includes an optional circuit that generates a balanced output by means of opamp U2. And secondly, the PCB includes provisions for cartridge loading, which were not available in previous versions.

Conclusions

Overall the Pearl3 circuit improves in some respects over previous revisions (cartridge loading, DC servo, balanced output option) while compromising other aspects (U1A not a class A circuit, excess of NFB, many more discrete stages in the signal path, DC blocking in the signal path and feedback). I personally prefer a low-noise shunt regulator for the input stage and a PCB utilizing single-opamp ICs so the user can optimize each circuit to the max. Nitpicking aside, the Pearl3 is a competent low-cost MC phono preamp that you can build around $400 using the available kit with some part substitutions, especially the coupling capacitors. You can even build it as a dual or triple chassis with a powerline filter in the power supply chassis, for best noise optimization that is only encountered in very expensive commercial units.

It would be interesting to listen to a maxed-out Pearl3 vs the "entry-level" (relatively speaking) Pass Labs phono preamps and against the low-price competition such as iFi Zen. or against my own mod to the Pearl2, which includes shunt regulators and a proper DC servo.

More to follow!