Wyn Palmer Low Output Moving Coil Phono Preamp
4-2022
This project started for me as a build of an AudioKarma project called the AD797 Phono Stage or alternately, the Hypnotized Low Output Moving Coil (hereafter LOMC) phono preamp. That design starting back in 2013, details of that project are here (note: its about 1,600 posts!):
https://audiokarma.org/forums/index.php?threads/ad797-phono-stage-build-and-help-desk-thread.501186/
I bought the PCB from AK member Sachin (Thanks!). As I continued researching the preamp (and plowing thru all those posts), I discovered that AudioKarma contributor Wyn Palmer suggested some modifications in the above thread to make the RIAA equalization more accurate, so I ordered those parts and planned to build that version of the preamp. But I kept reading...
In 2019, Wyn started a new thread at AudioKarma to design an improved LOMC phono preamp. The first version improved on the Hypnotized design by moving the 75us time constant to the first stage for improved overload characteristics and adding a DC servo to eliminate the output coupling caps. He soon added a novel warp filter (that rolls off the low frequency vertical signal). As the preamp evolution continued, he added a buffer/preamp circuit to use it as a full featured preamp, and lastly he added relay outputs to remove any thumps at power up/down. The first thread on the new phono preamp is here (it makes great background reading):
After over 2100 posts, a new thread was started to incorporate the latest changes to the design (a mere 1,500+ more posts to read!):
For those interested in the design, or building one of these preamps, I'd suggest starting with the first post in this last thread, as it includes a summary of the latest schematic, bill of materials, build notes, etc.
Wyn Palmer, the designer of the preamp discussed here, is a retired Senior Design Fellow at Analog Devices where he, among many other things, designed integrated circuits. So, a very well credentialed engineer! Many thanks to Wyn and all those involved in this project for doing the R&D work and sharing their results.
A summary of the design, copied from the above thread, by the designer, Wyn Palmer:
As the title suggests, it is opamp based, and is hardly unique, but it has some unusual features, has stellar specs, and is pretty easy to build.
Some features:
1. Optional parallel input amps- 2x LT1115/AD797/OPA1611 or OPA1612 bipolar opamps for MC- extremely high S/N ratio for low output moving coil cartridges. 2x OPA1641 or similar or OPA1656/OPA1642 FET input opamps for MM- also providing extremely high S/N with high cartridge resistance/inductance. Cartridge load components, both R and C, can be altered without soldering, allowing for loads from a few 10s of PF to uF.
2. 75us TC in input stage to optimize overload characteristic- approx. 28dB at rated -9dBv output at all frequencies up to greater than 50kHz with +/-15v supplies- an important feature matching high performance, commercial designs.
3. Additional standard RIAA TCs in output stage- again maintaining maximal overload margin.
4. Gain partitioning/opamp choices made to keep negative feedback factor high over the entire audio range to minimize high order harmonics and many tone IM distortion. No visible degradation in noise floor at levels much larger than rated output due to IM even with many tone inputs, high order harmonic distortion (e.g 9th) below noise floor minimizing degradation of sound quality.
5. Inter stage TC added to compensate for the transition to unity gain of the input gain stage to maintain RIAA characteristic compliance. No unity gain transition in output stage due to choice of architecture. This contributes to the compliance with the standard RIAA characteristic.
6. Phase adjusted DC offset correction in the phono amplifier - so no coupling caps in series with the signal path and no LF peaking due to the correction loop.
7. Ultra Low distortion, high stability, high accuracy Polypropylene caps used in all of the signal paths.
8. Off the shelf SMPS used with custom designed input LC filter. >100dB of supply noise rejection at SMPS switching frequencies.
7. Extensive on board power supply decoupling.
8. MC and MM versions can be built with some component changes, with essentially qualitatively identical performance.
9. 62dB gain for standard MC stage, 42dB for MM. Nominal sensitivity- 250uv @1kHz, 5cm/sec for the MC, 2.5mv for the MM.
10. RIAA simulated compliance c. 20mdB p-p, 20Hz-20kHz. Measured with non selected off the shelf components to be <50mdB p-p. Simulated worst case <100mdB p-p.
11. Unmeasurably low (limited by measurement system) distortion at rated output, including harmonic and many tone. Also applies to line and warp filter sections.
12. Phono Warp first order "elliptic" filter with vertical (L-R) rejection of c. 3dB at 140Hz and c. 36dB at 2Hz. Nominally 0dB rejection of horizontal (L+R) rejection. Great for rejecting noise from warped records or for some rumble components without disturbing the actual music frequency response.
12. Additional 10dB gain can be selected in the warp filter section to increase overall gain to 72dB/52dB.
13. Ultra low distortion, low noise high dynamic range Buffers/amplifiers for volume control and line driving added, 47 ohm output impedance.
14. Mono switch and balance control functions available.
15. Optional balanced output drive, 47 ohms output impedance/side (excludes "14")
This latest version of the phono preamp described above is the one I built and is detailed below. I wanted a phono preamp only, so I bypassed the buffer/preamp section and fed the output of the warp filter directly to the output relays.
But first, since I already had the Hypnotised PCB...
- My first Build Using the Hypnotised PCB
Since some of the parts, including a custom case and faceplate, had long lead times, I decided I could use the Hypnotize PCB to basically built Wyn's first version (1.3, no warp filter, etc), with the exception of the DC servo (leaving the coupling caps in place).
My original annotated schematic, showing the original Hypnotised design values and the Wyn Palmer revisions to correct the RIAA EQ.
Further revisions to my schematic drawing to use the Wyn Palmer version 1.3 design, which moves the 75us time constant to the first gain stage. There was no easy way to incorporate the DC servo of the WP design, so I kept the coupling caps at the output.
The Hypnotised PCB, unpopulated. Input at the bottom in this photo, output coupling caps and output pads at the top. The RIAA EQ components are primarily in the horizontal row in the center.
To allow plug-in resistors and caps for cartridge loading, I cannibalized the connectors from an IC socket strip.
This shows the cartridge loading cap (red Wima 220pf) in place in the sockets and the as-yet unpopulated sockets for the loading resistor.
This shows the PCB populated with the Wyn palmer V1.3 circuit, less the DC servo. As originally built, the input IC was an AD797, and the second stage was as OPA1611.
I later modified the circuit for lower gain. The circuit has 63dB gain by default, but the design documentation helpfully includes information on changing the overall gain while keeping the RIAA accurate. I've used an Audio Research PH3SE for many years, which has 54dB gain, and I've found it works with both high and low output moving coil cartridges. So I replaced a few resistors in the first stage to get an overall gain of 55dB. An associated change replaced the first gain stage IC , which was an AD797, with an OPA1641.
Just one note: the RIAA EQ resistors are all 0.1% tolerance, and the caps are 1%, which I found fairly readily available.
In order to build the Wyn Palmer 1.3 design on the Hypnotised PCB, some parts had to be mounted beneath the PCB, shown here. In general, however, the circuit fit fairly easily on the PCB - no cut traces, etc.
The PCB mounted in its chassis. Input connectors are at the rear bottom, to keep the low level signal path as short as possible to the PCB. The outputs are near the top, farther apart.
On the left side is the Jan Didden designed Silent Switcher power supply I used, behind an aluminum divider. This supply is fed 5VDC from a phone power bank battery or a phone charger wall wart, so no AC is in the chassis except the audio signal.
Rear view with input and output RCA's on the left and the USB-B connector for the power supply on the right. The RCA jacks are Vampire PCB2F/S gold/teflon connectors.
Some testing of my prototype indicated the RIAA EQ of this preamp was well within the +/-0.25dB tolerance of my Inverse RIAA. Channel balance was within 0.02dB. The power supply draws less than 1 watt, and the audio circuit draws 22.5mA from the +/-15VDC Silent Switcher power supply. Listening with a handful of friends indicated it sounded pretty darn good. This preamp now has a permanent home in by downstairs setup.
But Wyn had made a number of improvements to his preamp since his first version....
2. The real deal: version 3.2
The version 3 PCB (actually version 3.2). Obviously, this is considerably more complex than the Hypnotised PCB.
The input section at the left includes provisions for paralleled input ICs for lower noise, the middle section is that novel warp filter, then further right is a preamp/buffer stage, and finally on the far right is a set of output relays to ensure no turn-on/off thumps.
Thanks to AK member AlwinvdP for making these boards available!
Most of the parts to build the preamp. I had originally intended to use the Meanwell PD-25 switching power supply specified by Wyn (on the left here), but later decided to substitute the Silent Switcher, not shown here. I'll likely use the Meanwell power supply for the Wyn Palmer headphone amp, discussed here:
Most parts were bought from Mouser, some from Digikey (supply chain issues) and a few from my stock.
The aluminum chassis was custom made to my specs by Par-Metal.com, highly recommended.
The board partially populated, including the RIAA stage, DC Servo, and warp filter sections. No IC's in yet, only the sockets.
Again, critical resistors are 0.1% tolerance, and caps are 1%.
This preamp was built using the default gain setup for 63dB gain.
This is the completed board, ready for installation in the chassis. Inputs are at the left, outputs at the right, near the red relays. You may note that some of the ICs are standard 8 pin DIP (the LT1115s at the left and the TL052 for the servo in the middle), but the others, the OPA1611 and OPA1656, are SOIC form factor and had to be mounted on headers to fit on a standard 8 pin DIP socket.
That unpopulated section of the PCB is for the buffer/preamp that I omitted, as I plan to use this with a separate preamp.
Closeup of one of the OPA1611 SOICs on the DIP adapter.
This is the component side of the Silent Switcher power supply with the USB-B input. This side faces the outside of the chassis, and the other (inward facing) side of the Silent Switcher is a shielding ground plane. In a belt and suspenders approach, I also added an aluminum wall behind the power supply, shown here and in the photo below, grounded to the chassis thru the metal standoffs.
Preamp PCB installed for testing with power supply connected. No front panel connections at this point. Later, I'll wire up an on/off switch and LED from the power supply to the front panel. I also decided to wire a DPDT switch on the front panel for the warp circuit.
Rear panel showing the input and output RCAs (again, Vampire PCB2F/S gold/teflon connectors), and the USB-B power supply input on the right.
I designed a faceplate with FrontPanelExpress.com, shown at the bottom here. I had the holes for the two toggle switches threaded to match my switches, so the rear locknut and Loctite hold them in place, for a cleaner look.
At the top is the inner front panel (part of the Par-Metal chassis), after I punched holes for the switches and LED and tapped #8 holes for the bolts holding the panel in place.
Front panel with the switches and LED in place. This photo was taken as I was experimenting with dropping resistors for the LED to get the appropriate brightness.
Note that the Silent Switcher has a secondary supply at 6.3VDC for applications just like this, so the LED is not on the same +/-15V supply circuit as the preamp.
Some of my other gear is Benchmark, so I tried to mimic a bit of the same look with the black allen bolts attaching the front panel.
Inside view of the front panel: power switch on the right, LED at the center, both going back to the Silent Switcher. Warp filter switch on the left with wiring going back to the left and right channels of the preamp board.
I bought an inexpensive USB tester to measure the voltage from my battery and the current drawn by the preamp. This shows the steady state current of 0.64 amps after the relays turn on, at 5.1V, which is about 3.3watts.
The phono preamp in the system. The rechargeable power bank battery that supplies power sits atop the unit.