Ultra high spec opamp MC/MM phono, warp "elliptic" filter, line, headphone amps

No. The connecting runs to the supply/ground are a bit longer due to those components being in a bit of an island. I just wanted to reflect that change.
All of the 0.1uF cap ESRs are larger than the manufacturer specifies in an attempt to capture board issues.
The "cascading" filter sections proceed from the output amps- after the first 10 ohm Rs in series with the inductors- to the 2nd set of 10 ohm Rs and the highly sensitive input amp with the single 75us TC. The level of filtering increases as the circuit gets more sensitive, but the supplies and grounds are physically close to the output amps in order to ensure an optimal grounding scheme. As a result there is a minor island as far as the ground/supply connections are concerned in the first stage.
It's more important to keep the ground connection to the input pins as tight as possible than to maintain an ultra low impedance path to the output amps.
 
Last edited:
The final filter section provides about an additional 20dB of rejection from c. 10kHz up when compared to the rejection after the first set of Rs.
The gain of the input block is c. 20dB at 20kHz, and the PSRR of the AD797 is about 76dB- so the actual PSRR at this frequency is about 56dB. The filter adds about 80dB of PSR so the overall rejection is about 136dB (theoretically)!
The gain continues to decline monotonically after that at 6dB/octave until the amp hits unity closed loop gain at c. 200 kHz. Over this frequency band the opamp PSRR is therefore constant as we have not reached the intrinsic PSRR region. Beyond this the PSRR due to the opamp declines, but the increasing PSR due to the filter more than compensates.
 
Does anyone have any questions on the actual phono stage? The circuit has exceptional measured performance and there are subtleties that may not be apparent.
Its subjective performance also appears to be excellent- despite being opamp based and using a SMPS- both of which are generally perceived as anathema to audiophiles.
Some of the PSRR rejection discussion that just occurred touched on these matters.
 
Lacking the necessary technical expertise, I can only add my own observations that the phono stage is very detailed, has great dynamics seems very neutral, and is exceptionally quiet.

I have many phono preamps with which to compare; Threshold FET Ten/PC, Spectral DMC12, Soundsmith MCP2MkII, Hagerman Bugle 3 and Cornet 3, Vincent PHO 701, Accuphase C200, iFi Phono, Apt Holman with custom designed MC phono card, Clearaudio Smartphono, Forte Model 2, and Audible Illusions, and various SUTs, such as MA Cotter Mk 2L, Parks Budgies (both types of Cinemags), and a homebuilt SUT with HK XT-3 transformers.

Every one of these far exceeds the cost of Wyn's preamp, which sounds better than several in the list above.
 
Several? Damned by faint praise...
In fact, within the parameters of this design as I understand it, no preamp will be objectively better, but may be subjectively better due to deviations in frequency response and added acoustically pleasant harmonics - essentially their defects make them possibly subjectively "better".
 
Last edited:
Member
Joined 2011
Paid Member
I've observed over the years, that significantly more than (3/7ths) of the really great posts on diyAudio, are written on Friday, Saturday, and Sunday. I suppose there are huge numbers of members who don't read diyAudio except on the weekends, and that's when they type in their (excellent) comments.

I'm not saying this to belittle anything posted during the early part of the week; rather to encourage Wyn Palmer to wait till Monday before deciding whether he is disappointed with the responses he has received.

Does anyone have any questions on the actual phono stage?
 
Several? Damned by faint praise...

Not intended. My observations of the sound quality are of course entirely subjective, but actually, I'd say more than several, but I was trying not to appear hyperbolic, given the technical nature of the previous posts.

More accurately, none of them sound better to me than your preamp, but a few have sound qualities I like as much; the Threshold, my homemade SUT through the Cornet, and the Spectral. Of the ones I have, only these 3 and yours are in regular use or rotation. Least favorite? The Audible Illusions, Vincent, Forte.
 
Sorry Greg, I wasn't really giving you a hard time :)
As you know the goal of this exercise was to produce an objectively top flight performing design at an extremely low price- about what- $125 fully loaded-that would compete against much more expensive commercial designs, with objective calls being left up to the builder. Specifically, the design would employ commercial opamps and SMPS and extremely high levels of negative feedback- the anti Pass if you will.
The preamps I have compared it to include the Hovland HP100, two Classe designs- the DR5 and DR6, a couple of different Herrons, a VTL, and the Halcro DM10. It was indistinguishable from the Halcro to my ears, although the DM10 owner claimed that he could hear a difference- surprisingly in favor of the DIY one even though the DM10 even now goes for c. $7000 as a used unit- and distinguishable from the others, so the objective was met. The DM10 has almost identical specs, but a less objectively compliant RIAA characteristic, so it could have come down to a couple of tenths of a dB difference in overall frequency response that somehow seemed better to him.
So, a very subjective call indeed.
 
Member
Joined 2019
Paid Member
It does look good, the build notes are clear too. Only things I'm unable to work out is where to buy boards, I'm sure I read somewhere they're available. Or is it gerbers --> fab.

Also, with a -+15 supply, it looks to be an ideal candidate for the Jan Didden silent switcher that's been lurking in my spares box for the last few weeks.
 
I have a single board left of the v1.3 that does not have the dual opamp + elliptic filter section.
I sell the boards essentially at cost plus shipping and packaging.
The v.2 board that will have the latest functions is finishing layout and I will order a batch of 10 shortly. I will prototype one and offer the remainder for sale once I'm convinced that all is well. The dual opamp input/elliptic filter have been prototyped and the elliptic filter has been sent out as an add on board to prior users of the main board.
 
Wow, Wyn and Sean!

It really does sound fantastic! Extremely detailed, in fact - so much so I'd like the ability to roll off some highs for some recordinghs, but it is really dynamic, powerful, clean and quiet.

My original layout induced some PS noise, so I rearranged the board, made a cover for the PS enclosure I made, and re-routed all signal cables and the power wires. It is now so quiet it has to be turned up full blast to catch even a hint of electrons buzzing around. If I had played music at that level I'd be deaf.
In fact, it is so powerful I can't safely use the line jack on my Fiio hi-resolution player - have to use the volume-controlled headphone output jack.

Waiting for the glue to dry on the LEDs in the front panel, but after, will button it up and post photos.

I went retro-look; it is rather anachronistic to listen to 2-channel stereo in today's world, even more so to listen at home through headphones while, heaven forbid, reading!

Well done on the design and performance, Wyn. I am very glad you took the bait and designed this.

And Sean - you are getting really skilled at laying out a bunch of really tiny components for great performance.

Cheers,

Greg
 
The BOM will be posted shortly. We're trying to organize things to cover all of the optional features, and that is proving to be complex.
In essence the change consists of :
1. Phono input/75us TC opamp- from a bipolar to a FET input- for example one or two AD797s to one or two OPA1641/1642/1656.
2. Input opamp gain setting inverting input R to ground- from 21.5 ohms tp 105 ohms.
3. The interstage filter RC time constant.
4. The gain setting Rs in the output RIAA stage.
5. The DC offset control loop output R.
6. Phono input shunt R to ground- 100 ohms (typically) to 47k.