RIAA/Inverse RIAA Circuit - Jumper Selectable

[Wayne Kirkwood]

Hi Trolls. I'm new here to The Secret Society. Having stumbled into the forum numerous times and finding it a treasure-trove of information I figured it was time to give back. I make a few widgets useful to MEs and while working on a new project I stumbled (as I always seem to do) into a circuit that may be of use to this community.

The project is a "Swiss Army Knife" phono preamp/transfer rig. Its main purpose in life is to provide RAW, flat, un-EQ'd output to a DAW for capture, de-clicking/cleanup, and post clean-up, DSP-based, RIAA EQ. The flat preamp will have a fully-balanced instrumentation amp for MM carts. It's designed to be remotely-powered and located at the turntable. It's "line level" output will then feed a dual path record and monitor switcher located at the DAW about 30 wire feet away from the turntable. With un-EQ'd playback being painful to listen to, an analog RIAA playback circuit is available for monitoring or EQ'd (not RAW) transfers. When completed I will be able to simultaneously record RAW files as I edit and declick previously-recorded files monitoring through analog RIAA decode. Monitor switching, between the record path and playback path, is non-destructive. I can do two things at once. Having used the proto to rip and edit at the same time I know I can fly through an LP.

One suggestion for the "Swiss Army Knife Phono Preamp" came from this forum's Paul Gold. Paul suggested insert points to apply inverse RIAA EQ for cutting. Most of the inverse RIAA circuits are passive EQ/attenuators used for phono preamp testing. Few modern line-level inverse RIAA circuits seem to be published. I decided to see if it was possible to make the existing RIAA playback EQ into an Inverse RIAA.

The RIAA playback EQ I used is a unique topology that to the best of my knowledge was first described by John Roberts in his March 1981 Popular Electronics P10 phono preamp article. Solid State Music, "SSM," re-published John's design in the SSM2015 datasheet some years later.

What's unique about the P10 RIAA EQ is where feedback is taken. In the P10 EQ topology the 3180 and 318 us networks are split: The 3180 us time constant is in the feedback loop; the 318 us time constant is applied as pre-emphasis that works against the 3180 us time constant to form the 3180 to 318 us "shelf." The -6dB/octave falling response of the 3180 us pole working against the rising +6dB/octave 318 us pole produces a net shelving response from 318 us to 75 us. The 75 us pole, located at the input, is passive. The resulting transfer function is identical to the "conventional" RIAA EQ network without the 75 us/3180 us interaction. The result is an accurate RIAA network with minimal component interaction.

What I found with the P10 RIAA playback EQ is that the circuit elements could be reversed (with jumpers) to easily provide inverse RIAA EQ.


A switchable RIAA/Inverse RIAA EQ Circuit for Playback or Cutterhead Modulation.

The jumpers are shown in the "A" RIAA playback position.

The 3180 us network is formed by the 10 nF (1%) and 316K+2K. At their ideal values the tc is exactly 3180 us.
The 318 us network is formed by the 10 nF (1%) and 31K6 + 200R. The tc is exactly 318 us.
The 75 us pole, located at the input, is not obvious and for all practical purposes is passive.
The 75 us pole is formed by [(9K76+53R6)||(31K6+200R) * 10 nF. With ideal values it calculates out to be 74.99 us.
It should be noted that the 75 us pole is passive by virtue of the fact that the right-hand side of the 10 nF is held at virtual ground by the op amp.

When the jumpers are placed in the "B" position, which reverses the position of the 75/318 and the 3180 feedback networks, an inverse RIAA response is obtained.

One problem with any inverse RIAA network is that the response rises into daylight or until the op amp (a 709 in the SE-66) runs out of steam.
Neumann used second-order filters to limit ultrasonic response.
If I'm not mistaken both the SE-66 and SAB-74B used a Sallen-Key topology with a Butterworth, or slightly peaked Butterworth response.
The SE-66 used a 33 kHz post pre-emphasis filter, the SAB-74B a 50 kHz pre pre-emphasis filter.

In the inverse RIAA circuit shown above a 3.3 us time constant has been added, by the 332 Ohm resistor, to reduce the ultrasonic response.
In addition, without the 332R, few op amps would be able to drive the 3180 network due to its low AC impedance at HF.
An additional ultrasonic pole is added with an optional 100 pF cap.
(This cap should not be used in the RIAA playback position because it creates a small error.)
With only the 332R the 20 kHz response is down by about 0.5 dB from the ideal.
With both the 332R and 100 pF the response is attenuated at 20K by about 0.75 dB.
For extremely critical applications one could make the input resistor 100R and pre-filter using a Sallen-Key 12 dB/octave 50 kHz fc filter.

Does it work?
Yes it does.
I verified the RIAA playback response using the Jung/Lipshitz network as my inverse RIAA.
I then cascaded the "P10" Inverse RIAA into the P10 playback RIAA and drove the chain with both sine and square wave inputs.
The sine response was ruler flat until the minor 20 kHz rolloff - a rollof that both the SE-66 and SAB-74 exhibit.
The squarewave response was excellent and verified the amplitude response measurements.
As expected, the last octave showed the presence of bandwidth limiting from the ultrasonic poles.

Those wishing to follow the preamp thread can find it here. It's a long and winding thread and still a work in progress. I recommend anyone following it jump to the later schematics using the link in the original post.

http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=423

I hope all of you find the circuit useful.

Wayne

[markrob]

Hi,

That's a really nice clean implementation. Thanks for posting it.

Mark

[tubeactive]

Truly Fantastic Efforts Here ! Literally, Very, Very Applaudable ! Well Done ! Thank You and those involved in the linked work !

[Wayne Kirkwood]

Thanks everyone for the views and comments.

I wanted to provide some follow-up information.
As a reality check I decided to check the Inverse RIAA feeding the RIAA playback EQ end-to-end.

I first checked the playback RIAA EQ network accuracy using the Jung/Lipshitz network through the preamp low-level input.
The playback RIAA network on the protoboard used the Roberts P10 "8300 pF" (2.5%) caps. (1500||6800 pF, 383K, 38K3, 11K8).
The overall accuracy was ruler flat using the Jung/Lipshitz Inverse RIAA with the RIAA playback network.

The line level Inverse RIAA network was an identical topology to the P10 playback network but used 10 nF 1% caps and series resistors to get the R exact. (316K +2K etc.)
These are the schematic values in my OP and those used by SSM.

Once the playback network response was verified I used a sine input feeding the line level Inverse RIAA output into the playback RIAA network.
The cascaded networks were "ruler flat" with the exception of the minor 20 kHz droop due to ultrasonic roll off in the Inverse network. (About 0.75 dB at 20 kHz.)

A picture is often worth a thousand words and time domain accuracy - square wave response - can speak volumes about frequency response.
Though we don't record square waves they can tell us a lot about the frequency domain - over a lot of data points - in one quick glance.
If the Inverse RIAA and RIAA curves are precisely matched, then a square wave in should produce a square wave out.

As all of you know RIAA pre-emphasis is a brutal, transient, thing.
The top trace is the input, the bottom trace the output from the Inverse RIAA filter.
To give you an idea of the dynamic range a pre-emphasized square wave generates, the input trace is 200 mV/div, the output trace is 5V/div.
Total square wave discombobulation.
Headroom is precious and to say that Inverse RIAA is a differentiator is an understatement.


Output of an Inverse RIAA network driven by a 1 kHz square wave.

So what happens to waveform fidelity after RIAA de-emphasis?
Can this be put back together (re-combobulated) accurately?
Yes it can.


Encode-Decode Response of Inverse/Playback RIAA networks passing a 1 kHz square wave.
Top trace is input, bottom trace is after encode-decode.

Though there is some slight RC rounding due to the ultrasonic roll off, the square wave response illustrates the complimentary, inverse, responses of the networks.

[gold]

A really cool feature of this project that isn't apparent from Wayne's post is that the front end of the preamp can be remote located. In other words you can locate the input PCB of the preamp directly next to the tonearm and have power and audio on an umbilical back to a base station with all the rest of the circuitry. It is also uses a balanced topology. To get the most out of the preamp use twisted pair from the tone arm to the preamp input.

Often it's not practical to locate a phono preamp next to a tone arm. The most common source of noise in a playback system is from the cable run. This lets you have cables that are very short from the tone arm to the phono pre input. Couple that with a balanced topology and it should be easy to have a whisper quiet playback system.

[EpicenterBryan]

Hi Trolls. I'm new here to The Secret Society. Having stumbled into the forum numerous times and finding it a treasure-trove of information I figured it was time to give back. I make a few widgets useful to MEs and while working on a new project I stumbled (as I always seem to do) into a circuit that may be of use to this community.

You are THE Wayne, the guy who did the Precision Mid Side MS Matrix Board over at proaudiodesignforum.com? Which is also available for purchase on the site and also on E-bay?
I'm going to put a link in this post for people who have not seen:
http://www.ka-electronics.com/kaelectronics/MS_Matrix/MS_Matrix.htm

Oh man, when I saw your schematic on this post I recognized your handwriting, and style. How cool, or weird is that?
Wayne, I love this idea! I want one. Are you planning to offer a board or kit?

At some point I planned to buy a pair of your M/S Encode/Decode boards and build a specialized box to route signals to and from a pair of multi-band compressors... Kind of like a souped up Poor man's Fairchild 670 dual compressor with vertical and lateral mode but with multi band. Again, at some point I planned to share the idea here at the lathetrolls....

If you have time to start a new thread about your Mid Side MS Matrix Boards perhaps with a link to the extensive info over at proaudiodesignforum.com on the design and theory I think many people will be interested. This signal path is kind of magical and mysterious, and dives deep into the non disclosed tricks that a disc mastering engineer does, so it's kind of like exposing how a magic trick is done. I for one would do it, just because it's all so new to so many people, and even though I'm an old dude, I'm just now discovering how cool all this is...

Bryan

[tubeactive]

Back to the RIAA thread, Wayne is the man ! Thanks again, Wayne for lighting this fire and putting the info on this forum ! Playback seems to be a slow responding forum. Threads either get unnoticed or the good folks might have their own "special recipes" which some may want to keep private; or so it seems. Phono playback is of utmost importance to me. Sharing our knowledge will only share the wealth (of enjoyment) possible.

While I am thoroughly delighted to see near perfect square waves, I can only wonder about actual sound quality. Specs and waveforms can look great on paper but descriptions of the emanating sound, the sonic flavors or neutrality, the speed and dynamics do not make a visible appearance. Wayne describes some listening results on the proaudiodesign forum linked thread using some of his reference discs; which is imperative to believe any commentary. There will always be lurkers or responders who doubt our opinions, but when a sound engineer builder/reviewer actually believes in his work and then proves aloud to himself in writing, his "opinions" are more easily trusted, IMHO...

So, Wayne, once again THANK YOU ! How would you describe the "speed" or "action of the music" with your preamp's phono playback ? While pretty square wave waveforms indicate wide, flat bandwidth with little phase error, the sounds of music like the plucks of the strings, fingering of the keys or wind instruments, strikes of the drum skins et al, hovering in space within a holographic imaging soundstage are not evident on our scopes. You used Santana Abraxas as one of your ref discs, which is clearly one of mine. Another Columbia LP I adore is the second (self-titled) Blood, Sweat & Tears LP; the one with Spinning Wheel and God Bless The Child. With my favored preamps, usually tubed, along with my very "fast" sounding amps(little or no feedback loops in my ref amps), I can easily hear the "action" of all instruments with a serious, defined 3D image. I am not trying to boast at all. I simply wish I could use a solid state preamp (or power amp) that could match or even do better in the areas of speed and believable imaging. I use classic speakers and do not rely on types that force a certain image or soundstage. My speakers of choice are Altec Lansing VOT with old 16 Ohm drivers for large room listening, while my nearfield duties are carried out very capably with full range 8s in transmission line towers.

In the early stereo era, circa 1958-1960 there were some very cool preamps and amps available; still worthy of resto and upgrading today. One of the earliest stereo amp/preamp (integrated amp) offerings was the Madison- Fielding model 320; the one with the mahogany on the faceplate. Along with the early sixties Ampex 008 stereo preamp, these appear to be some of the only holdouts from the mono era that utilized passive EQ at the input for RIAA Treble Deemphasis along with active feedback EQ for the RIAA Bass Boost EQ. Playing my stereo ref discs through these, I hear a "different time and space" playback than when I use my favored all active feedback EQ preamps. I have a super-rare Fairchild 602 preamp module which uses passive EQ RIAA networks that I also wish would sound right; but it does not. None of the passive EQ preamps I have tried seem to get the timing right; and I have owned and tried some very expensive ones. The breeds of preamps with split EQ networks among different gain stages also seem to be at fault with respect to the speed or timing of the instruments.

My ultimate ref disc is a live recording, done by the famed Wally Heider studios in 1969, the first (self-titled) Hot Tuna LP on RCA; a very vintage 2S stamper. Even later stamper discs can easily elicit that "there they are...hovering between and behind the speakers" feeling. Every pluck and fingering by the bassist and acoustic guitarist is clearly discernible with the "right" preamps. I believe that most of the LPs of that sixties through the seventies and even into the eighties era were cut with feedback equipped cutters. My playback philosophy is to play the discs back retaining that timing, the action of the instruments, the "time and space" perspective. I seem to always prefer the active feedback EQ; even with my late seventies and eighties fave discs. Dire Straits "Brothers In Arms" and the Cars LPs sound more "right" with feedback EQ, IMHO and IME.

So, Wayne, have you tried any version of your cool preamp with all feedback EQ ? How about dual mono power supplies or even separate DC Voltage rails for each channel ? Even with low current devices, separate power supplies for each channel will always be sonically rewarding. In fact, even if you use a stereo amp with a single power supply, a twin power supply preamp should still yield better separation, height, width and depth, along with superior dynamics...

[Wayne Kirkwood]

Thanks everyone for your replies. I expect to post the balanced input flat preamp schematic here soon and then follow up with the EQ/Monitor switcher.

It's interesting that the same albums tubeactive mention are ones that I use for listening tests. Santana Abraxis, Blood Sweat and Tears, and Hot Tuna's Watersong from the Burgers LP. Another big favorite is Steely Dan and Donald Feagan. I may need to buy a new copy of Nightfly.

I don't normally use the gizmos I design and view my role as one of toolmaker. I depend greatly on the users of those tools to tell me if the sound is right. Those folks have the ears, source material and monitors. The goal is usually no added coloration. I try to make things the best I can at reasonable cost with available parts and have the people who build and use them tell me what they think.

This project is different in that I'm designing it for me - a guy with a bunch of 70s and 80s LPs - who needs to easily transfer flat RAW files and de-click them. Since editing and confidence monitoring RAW files is painful and fatiguing I incorporated RIAA playback EQ primarily for use in the monitor path. The final files will have DSP-based EQ. My objective was to make a transfer rig and not a preamp for general listening. That space is very crowded.

My thought was is that if I make it flexible enough other people might find more uses than archive and transfer. One of the potential applications is test and measurement of cutting chains. For potential test and measurement customers I needed very accurate RIAA EQ. (An insert for a cartridge correction filter is also included.)

The choice of active or passive RIAA EQ seemed simple enough and based on intuition and the complexity of passive EQ I chose active EQ. Because the gain of the flat front-end is high - it drives the A/D or sound card directly - an RIAA topology capable of attenuation is required. A conventional active, non-inverting, all-in-one EQ stage cannot produce gain below unity. The mid-band gain has to be more than 20 dB or it won't accurately follow the 75 us response of the RIAA curve. With unity gain at mid-band being a requirement, the active non-inverting stage was ruled out. Active inverting stages (as well as passive ones) produce attenuation and can precisely follow the RIAA curve into the ultrasonic. (The 1 kHz gain is actually around -2.5dB.)

Is the 75 us pole in the P10 schematic passive or active?

Good question. It depends on how you look at it. When I originally used the word passive it should have been in quotes. Analysis of the 75 us pole is a lot easier if you look at the right hand side of the 10 nF capacitor as being grounded. The 75 us time constant is then the product of the parallel combination of the resistors in the input leg - their Thevenin equivalent is 7.49929K - and 10 nF. But is the 10 nF really grounded? No. It's held at virtual ground by feedback through the 3180 us network. The servo action of the op amp holding the inputs in null makes it ground. Because the 75 us pole requires feedback, my view is that it's really active by definition. As long as the op amp keeps itself in null the effect is the same as if it were passive but its requirement for feedback makes me want to call it active. One can have an active EQ with elements both in the input and feedback arms. What makes the P10 different is where feedback is taken from: The 3180 us and 318 us time constants become separate networks rather than the 318 defined by the interaction of the 3180 and 75. Calling the 75 us passive makes it easier to visualize.

Another requirement for my projects are readily-available components out of the Mouser catalog. In this area the P10 RIAA EQ shines with standard E12 cap values and E96 resistors hitting the time constants almost exactly. The modified P10 values (SSM version) use E3 cap values and series-connected E96 resistor values. With the cap values being equal, the matching of units within the same lot should be excellent. Mouser also stock the 10 nF Wima PP in 1%. (505-FKP20.01/63/1). I looked at some of the earlier work of Jung in TAA and I stock the polystyrenes for his JW-33 EQ. I tried that EQ back when it was published and always liked the fact it rolled off indefinitely. In the end I decided on the P10 from John Roberts due to its simplicity. Having the 75 us pole at the input is also a big plus.

I do not doubt that all-in-one RIAA FB networks sound differently. The majority of all-in-one networks are non-inverting and do not roll off at -6dB octave in the ultrasonic. It may be that the combined response of the cutting chain - which can't boost at 6 dB/octave forever - combined with the reproduce chain - which can't fall off at 6 dB/octave forever combine in such a way that they compliment nicely. When time permits I'd be curious to test that theory.

[Wayne Kirkwood]

I have a small run of pre-production RIAA/Inverse RIAA EQ Monitor Switcher boards now and wanted to update the Inverse RIAA thread.

The jumpers look like this:


RIAA EQ/Inverse RIAA EQ jumpers

The EQ circuit on the left has J5A and J5B jumpered for normal RIAA EQ.
The right channel has jumpers for Inverse RIAA.
(It's recommended for inverse RIAA that C27 and C28, 100 pF also be fitted to provide additional ultrasonic rolloff.)

These are the frequency response plots of the RIAA and Inverse RIAA outputs.


RIAA EQ/Inverse RIAA EQ Response Flatness Plot

This plot has the Inverse RIAA Output feeding the RIAA playback EQ.


Inverse RIAA EQ Response with RIAA playback EQ applied.

The playback RIAA EQ has a subsonic rolloff of about 2 dB after the EQ stage which can be seen in the decode response.

Its easy to switch from record to playback RIAA EQ by simply moving four jumpers.

[EpicenterBryan]

Hey Wayne,
Are you at a point where you can share prices?
Totally interested.

Bryan

[Wayne Kirkwood]

Thanks Bryan.

I did a small pre-production run of four and have two sets of preamp and EQ boards available.
Will be sending you a PM.