Hifisonix X-Altra RIAA Phono EQ Preamp

Important note September 2024: The BOM dated prior to September 17th 2024 was severely corrupted (my fault entirely by the way – I had multiple copies open and they were cross-syncing into the cloud).  Here is the link to the audioXpress supplementary data that now has the completely corrected BOM.  

Here is the link to the correct BOM Supplementary Material | audioXpress

Note, the top level download file is named ‘Russell-XAltra-AX-2021’ but the contents have been updated – the September 17th 2024 BOM is listed in there.

Do NOT USE any BOM that is dated prior to September 17th 2024!

You can buy a set of X-Altra MC/MM preamplifier boards for $49.50 + $15 for shipping over in the Shop. 

Here is a video about the preamp that discusses the key design aspects  https://youtu.be/L1CA2_t4fu8https://youtu.be/L1CA2_t4fu8


The X-Altra MC/MM RIAA preamp is a state-of-the-art design that brings new levels of performance to vinyl EQ featuring best-in-class measured thermal noise on both MM and MC inputs and tight RIAA EQ conformance. On moving coil inputs, the RTI noise contribution from the amplifier stage was measured at just 232pV/rt Hz (pico Volts per root Hertz), or -192.6 dBV, placing it amongst the very quietest MC phono preamplifiers available at any price, be they commercial or DIY offerings. Tight RIAA conformance extends from below 20 Hz to >50 kHz and over 20 Hz to 20 kHz remains within +-0.15 dB, while distortion remains below 50ppm up to 8V RMS output from the preamplifier (MM input) and better than 0.01% at 8V RMS output on MC inputs (50 ppm at 1V RMS output).  These performance figures come with class-leading input overload of >30 dB on both moving coil and moving magnet inputs. A novel input offset servo keeps the MC input at 0V +-25uV ensuring unwanted MC coil DC bias currents remain negligible and allowing direct coupling of the transducer to the amplifier, further contributing to the outstanding noise performance.

MC and MM transducers require very different amplifier characteristics if the designer is looking for ultimate performance.  For this reason, the X-Altra MC/MM separates the two input stages and uses amplifier topologies that offer the best performance in each case. The MC amplifier is configured as a current injection stage around a complementary pair of ultra-low noise bipolar transistors (ZTX851 and ZTX951).  This treats the source as a current transducer, rather than the more conventional approach where it is treated as a voltage source driving a 100-200 Ohm resistor load and is a major contributor to this preamplifier’s outstanding noise performance. In practical terms, compared to a very low noise JFET (en typically ~ 0.9nV/rt Hz) or high-performance op-amp-based MC head amp (for example an AD797), the X-Altra MC/MM is 3.8 x quieter (>11 dB), and 6 dB quieter than Douglas Self’s MC preamplifier (Small Signal Audio Design (2010), pages 213-218 with performance quoted as -141 dBu = 479pV/rt Hz).  Note that the figures quoted for the X-Altra MC/MM are all based on measurements of the full signal chain from the input right through to the output of the preamplifier. If the X-Altra MC/MM is quoted similarly to the Self-design, the spot noise performance is -147 dBu. Compared to other commercial products using discrete input stages, the X-Altra MC/MM is typically 6 dB quieter.

On the MM input, a dual ultra-low noise JFET (Linear Integrated Systems LSK389 or TI JFE2140), operating in parallel single-ended mode, is paired with a high-performance opamp to realise a single all-active equalisation gain stage with noise levels to within 0.3 dB of the theoretical limit, along with >30 dB overload performance.  The X-Altra MC/MM achieves better than -76 dB SNR (unweighted ref 5mV 1.3k + 500mH cartridge source). With the input shorted and ‘A’ weighted, the input-referred noise is -89 dB, ref 5mV. 

The design makes extensive use of SMD components, so the result is a very compact unit with small loop areas which in turn allows the unit’s power supply to be co-located on the amplifier board.  The mains peak noise floor of the finished preamplifier is 20 dB below the thermal noise floor of the amplifier with a standard MM cartridge (1.3k + 500mH) connected i.e. inaudible.

 

You can download a copy of both Part 1 and Part 2 from the audioXpress website here :-

Part 1    https://audioxpress.com/files/attachment/2735

Part 2   https://audioxpress.com/files/attachment/2736

 

Here is a link to the February issue overview and towards the bottom, a quick intro to the X-Altra MC/MM Phono Preamplifier :-

https://audioxpress.com/news/the-audio-technologies-that-matter-more-than-ever-in-audioxpress-february-2021

Here is a picture of the March issue of audioXPress with second part of the X-Altra article

 

If you have any questions, or require technical support with your build, contact me via the comments section below or email me at bonsai(at)hifisonix.com (replace the (at) with @)

There is also a diyAudio thread about this preamplifier here :-

Bonsai’s X-Altra MC/MM Phono Preamp | diyAudio

Errata: Kindly note that in the circuit diagram in the published article C9 on the left channel (and C25 on the right channel which is not shown in the circuit) is the wrong way around. However, the PCB silkscreen is correct, so please follow the silk screen orientation for these two capacitors when assembling the board.

This document shows you how to wire the Filter T/Over switch, located on the rear panel, to the PCB :-

Here is a table that allows you to get an idea of how noisy a MC preamp is given the cartridge coil resistance and the amplifier input referred noise. Once you have gotten the number off the table, use the following formula to calculate the S/N Ratio: 20 log [(Rdg*144)/Cartoutput] where Rdg is the reading from the table and Cartoutput is the MC cartridge output in Volts.  It is clear from this table that the X-Altra MC/MM is in a league of its own in terms of thermal noise (and overload capability).

For the X-Altra, the measured equivalent input thermal noise resistance is 1.4 Ohms, so its ABOVE the 2 Ohm row, but we will use the 2 Ohm row since it is close enough. The figures in RED show the expected X-Altra preamp S/N ratio for a 200uV output cartridge. The ‘A’ weighted figures will be about 4 dB better than those shown.



Some photos of the completed preamplifier, along with my Michel Gyrodec are shown below.

 

 

 

Comments

11 responses to “Hifisonix X-Altra RIAA Phono EQ Preamp”

  1. Raph (blacknoz) says:

    Hi Andrew,

    Hope you are doing well.

    I love your nx-amp and would definitely like to build this RIAA.
    If you go for a group buy, just ping me, I’m in 🙂
    I’ll mix it with the Doug Self pre from Linear audio #5 I just finished last month and which sounds great with the nx-amp (at least to my ears).

    Best,
    Raph
    P.S. if I may, would you share / recommend a PCB supplier for small audio projects as I am trying to design some at the moment?

    • Bonsai says:

      Hello Ralph,

      I’m still doing some work on the preamp. I have been approached to potentially have it printed in a magazine, so once development is completed, I will submit it to see if they want it. If not, it will go up here on the Hifisonix website.

      I’ll keep everyone updated on the X-Altra page

      Rgds

      Andrew

    • Bonsai says:

      For PCB manufacture, I use ‘PCB Cart’ who are based in China. I found the quality very good.

      Rgds

      Andrew

  2. Raph says:

    I was coming back randomly to your site and found your answer. I’m glad that a magazine decided to publish your work, although I guess that it will be hard to find here in France 😉
    Your photos look great, I’ll come back here to read your updates.
    Thanks for sharing your PCB manufacturer and one more time thank you for the nx-amp which I enjoy nearly every day.

    Take care,
    Raph

  3. Matej says:

    Hi,

    congratulations on great design and making it to real product!
    I like low noise designs, but I like simple solutions too.

    In a real world of vinyl records I would say we can get S/N about 60dB(A).
    If this is true, why just not use NE5534 instead of (unobtainable) FETs at front-end?
    I am asking from practical point of view (I completely understand engineering challenge of building the best phono pre-amp). Do we need discrete solutions (for MM)? I mean good PSU and well designed RIAA should do, right?

    Best,
    Matej

    • Bonsai says:

      Hello Matej.

      The S/N improvement over an NE5534A with a MM cartridge connected approaches 5 dB which is significant.

      Certainly vinyl surface noise is much higher than the preamp stage.

      The X-Altra MC/MM provides SOTA performance levels for very reasonable cost and provided a great design challenge. Similarly performing solid state vinyl preamps sell for more than $2000.

  4. Hello Andrew,

    I read the articles in Audio Xpress, very good! You mentioned the headroom of 31.5dB across the audio band, although not clear in the text, not to me anyway, I think you mean for signals that are RIAA coded, not for siganls that appear after the cutting process like warped and eccentric records, cracks and plops and so on. This is relevant because most of the discussion about headroom concerns this type of signals.

    • Bonsai says:

      Thank you Joost.

      Yes, the headroom is for RIAA encoded signals. However, good overload recovery with RIAA signals also helps with non encoded signals like clicks (lot of HF content) and warp and arm resonance (LF). Stereophile has reviewed commercial phono preamps with only 6-9 dB O/load capability at HF, and one recent product was less than 3 dB. This is not enough in my view and most sensible practitioners would agree.

  5. Thierry Vitale says:

    Hello

    I’m about to order the PCB kit and had a look at the Bom. I see that the 4 x 12 nF caps (GRM2195C1H123FA01D #12) are absolutly unavailable anywhere and the #8 are supposed to arrive in mid June.
    Is there a way to find them or is there an equivalent ?

    Thank you
    Best

Leave a Reply

Your email address will not be published. Required fields are marked *