phono stage building blocks...

Hi all,

So although no electronics idiot (okay, I probably am), I've never had anything to do with phono stages, other than owning them. I would like to understand fully what actually goes on in an RIAA phono stage (other than the RIAA equalizer). I understand the concepts of time constants, etc, and understand that in various regions of the RIAA curve different time constants are required. There are many excellent kits and DIY phono stages out there, but I just need some help coming up with a block diagram of what is needed (I do have an intended use for the outcome of this thought experiment).

So as I understand it, an RIAA can be completely passive (no active components, except an output stage), Transformers can be used on the inputs to step up voltages (or a small amplifier to provided 20dB-40dB input gain prior to the actual phono equalizer if the cartridge is a LOMC.) and another small amplifier can be used as a gain stage on the output of the phono equalizer to allow lower impedances to be driven (so essentially a buffer stage). Is this correct?
 

Attachments

  • phono stage scheme.jpg
    phono stage scheme.jpg
    48.5 KB · Views: 470
Nanook said:
So as I understand it, an RIAA can be completely passive (no active components, except an output stage), Transformers can be used on the inputs to step up voltages (or a small amplifier to provided 20dB-40dB input gain prior to the actual phono equalizer if the cartridge is a LOMC.)
No. The output from MM is far too small for just a transformer. You need some power gain; all a transformer does is change impedance. MC may use a transformer plus a gain stage.
 
Even for a MM cartridge, the active devices need to provide, say, 78 dB of gain which the RIAA network will reduce by 20 dB at mid-frequencies relative to the lowest.

Occasionally, those involved in transferring old recordings (78's) use passive equalisers or even do the equalisation in the digital domain but this is after a flat preamplifier with considerable power gain and commensurately low output impedance.
 
So as I understand it, an RIAA can be completely passive (no active components, except an output stage), Transformers can be used on the inputs to step up voltages (or a small amplifier to provided 20dB-40dB input gain prior to the actual phono equalizer if the cartridge is a LOMC.) and another small amplifier can be used as a gain stage on the output of the phono equalizer to allow lower impedances to be driven (so essentially a buffer stage). Is this correct?


Hi,

Whilst fully passive RIAA is allegedly the only way to do it
according to hifi snobs, the technical reality is very different.

RIAA stages slug the response with a baseball bat 1st stage and then
repeat the same treatment 2nd stage. Much nonsense is written about
the subtleties of what is effectively pretty much a glorified integrator.

Passive is a bad way to do it, as any technical analysis will indicate.
Active is far better in all technical respects. Nobody builds passive
integrators because the active version is just so simply superior.

However if you want to go pointlessly complicated, go passive.

rgds, sreten.

FWIW you can make a killer specification MM RIAA
stage with a 5532 dual op-amp wired in parallel
and all the RIAA EQ built into its feedback loop.
 
Last edited:
There is a very interesting chapter on equalisation in "Understanding hi-fi circuits" by Crowhurst
(1957). His conclusion is that RIAA is not most accurately achieved in a feedback loop.

Hi,

The calculations to do RIAA all at once in a feedback loop accurately are
horrendous (due to interaction) compared to a two stage implementation.
(Two stage can be passive or active, neither is complicated for accuracy.)
However they were done by Lipshitz in 1979, 22 years later.

rgds, sreten.

See D.Self's "Precision Preamplifier" for values.
 
Last edited:
sreten said:
Much nonsense is written about
the subtleties of what is effectively pretty much a glorified integrator.
It isn't an integrator, but a low pass plus a shelf. I'm sure you know that.

Passive is a bad way to do it, as any technical analysis will indicate.
Active is far better in all technical respects. Nobody builds passive
integrators because the active version is just so simply superior.
As it isn't an integrator, arguing from integrator implementations can lead you to the wrong conclusion.

There are pros and cons to both passive and active RIAA. Both can be done well or badly. Doing the chosen method right is more important than choosing the 'right' method.
 
It isn't an integrator, but a low pass plus a shelf. I'm sure you know that.

As it isn't an integrator, arguing from integrator implementations can lead you to the wrong conclusion.

There are pros and cons to both passive and active RIAA. Both can be done well or badly.
Doing the chosen method right is more important than choosing the 'right' method.


Hi,

An integrator is simply a 6dB low pass. Its not far off
and my comments are more valid than ignoring them.

rgds, sreten.

A little known fact is RIAA inputs were nominally
compatible with ceramic cartridges due to the
impedance mismatch causing a 6dB roll-off.

The last quality ceramic cartridges used RIAA. e.g. :
Connoisseur SCU 1 Manual - Stereo Ceramic Cartridge - Vinyl Engine
 
Hi all, I leave for a day...

Thanks to all for pointing out my misconceptions.

I was looking at a first stage to provide gain for a MC. But could it be set up to provide a unity gain stage with a high input impedance, and low output impedance. Then send it through the passive stage, then set the input impedance of the second stage to vey high, allowing the first amp stage to drive through the passive stage and into the second? Thus the end result would be a lower output impedance to drive through whatever switching, etc?

I do understand that the active chip phono stages could be much easier to build. Iam just trying to understand the fundamental building blocks.

Regarding calculations and simulations for RIAA equalization, could one not use a table of values, plotting them out, then use a regression of the curve to come up with a curve of best fit? Again, just a question.

If any have a good resource for understanding the mathematics of the RIAA curve, please post them here. I'll look into them further.
 
A typical MC cartridge produces about 200 microvolts; the first gain stage must be low noise above all else. If I were doing the equalisation all in one network, I would amplify to about 1 volt before doing so and then have a further gain stage of 26 dB to get the signal up to line level.

The RIAA equations (for a single network) cannot be solved analytically but only by an iterative process; there is an online RIAA calculator which is very helpful and not limited to RIAA time constants-you could calculate for NAB, AES, American Columbia etc.
 
sreten said:
An integrator is simply a 6dB low pass. Its not far off
and my comments are more valid than ignoring them.
An integrator has no corner frequency, and so has infinite gain at infinitesimal frequency. Unlike a low pass.

A little known fact is RIAA inputs were nominally
compatible with ceramic cartridges due to the
impedance mismatch causing a 6dB roll-off.
Not sure how relevant that is, although the fact is true. I'm not sure how true it is that it is "little known"; maybe little known nowadays, but well known 40 years ago when ceramic cartridges were more common?

Nanook said:
Iam just trying to understand the fundamental building blocks.
The fundamental needs are gain and RIAA equalisation, arranged in a way which does not introduce too much noise and does not lead to premature signal overload. There are several ways of achieving this, which is why people do it differently. I'm not sure I understand what you mean by "the fundamental building blocks".

Nanook said:
Regarding calculations and simulations for RIAA equalization, could one not use a table of values, plotting them out, then use a regression of the curve to come up with a curve of best fit? Again, just a question.
Why? Circuit theory is well known, so from a circuit the frequency response can be calculated as a rational function. The RIAA characteristic is defined as a rational function, based on time constants (or, equivalently, corner frequencies). Nothing to fit; just use algebra. I'm sure Google can find you lots of information, including the classic Lipschitz paper.

piano3 said:
The RIAA equations (for a single network) cannot be solved analytically but only by an iterative process;
?
 
An integrator has no corner frequency, and so has infinite
gain at infinitesimal frequency. Unlike a low pass.

Hi,

You know that is not true regarding gain in a practical analogue computer.
Perhaps I should have said RIAA is a glorified 1st order low pass to be
more accurate, which is an integrator above its low corner frequency.

rgds, sreten.
 
Hi,

If you want to take the building block passive approach its relatively simple.

FWIW power output of MC's exceeds that of MM's and transformers can be used.

1st stage is low noise gain, both voltage and current gain, it must be active.
You need enough current gain to drive the 1st passive RIAA stage, which
is normally the top end of the RIAA curve. That then drives a further
active voltage and current gain stage, which drives the 2nd passive
RIAA stage, the low end of the RIAA curve. Output impedance at this
point depends on the passive stage, it will fall to a low level by 1KHz.
Any further buffering /gain is usually not needed, but can be added.

As transformers have no power gain they are only relevant when
considering the types of active gain stages to be used properly.

rgds, sreten.

Note I'm not saying the above is the way to do it, its the block approach.
 
Last edited:
thanks to all...

sreten: thanks for making the example for me. as I can be a little thick some days, I want to look at this as a group of building blocks and your last post helps. As always quite succinct.

DF96: Regarding using the existing math, I've always been like Einstein in that I am always seeking that 1 equation...and a method for obtaining a result to fit the equation and reality (even though it took 50-ish years for Einstein to be proven correct).

piano3: I'll look for the online calculator. I'd like to see what they are using for assumptions.
 
Nanook said:
DF96: Regarding using the existing math, I've always been like Einstein in that I am always seeking that 1 equation...and a method for obtaining a result to fit the equation and reality (even though it took 50-ish years for Einstein to be proven correct).
The one formula for RIAA can be found, As I said, it is a rational function. I don't know what else you are seeking. Invoking Einstein will not help: general relativity is given to us by the universe but we had to discover it; RIAA equalisation is something we decided to do.