Characterizing phono preamps

[bcarso]

My apologies at the outset for asking questions that have doubtless been discussed somewhere already, in the vast archives of diyaudio. However, search terms of high specificity didn't yield much, so I'm launching this.

It's been long noted (see e.g. the 1976 National Semiconductor Audio Handbook and earlier cited references; see pp. 6-13 - 6-16) that merely measuring the input-shorted voltage noise of a phono preamp to characterize its noise performance is almost meaningless. Yet Stereophile et al. continues to do this. I suppose we should be happy that some measurements are made at all.

This shorted-input measurement misses all of the preamp's parallel or so-called "current" noise. Now, with very-low-impedance moving-coil cartridges this isn't usually that significant an error. With moving-iron it's quite different. With FET input stages and their audio-frequency low parallel noise (an induced-gate noise term does start to manifest above about 100kHz), there is little effect from the active device noise. With bipolars and their base current, the typical moving-coil part shorts out most of the parallel noise. But the inductance of typical moving-iron cartridges is in the range of half a Henry, and some high-output MC parts have high inductance as well, and there is a significant contribution at high frequencies from the typical 47k termination and any active device parallel noise.

In addition, for measurements of distortion, the preamp input frequently has a voltage-dependent impedance, usually mostly capacitative, and this makes the distortion measurements with a low-impedance resistive source additionally unreliable.

One could suggest using a phono cartridge and placing it in series with the test generator. But this will contribute noise from the resistive part of the cartridge impedance, and as well a thermomechanical noise from the suspension. If independently determined, this noise could be "backed out" of the net noise measured, making it one way to go.

Anyway --- I'm throwing this out there in hopes that someone has addressed it, or will give it some consideration. I have some ideas for a test fixture which might be worth pursuit, starting with a very high quality inductor for the brute force approach.

Brad Wood

[bcarso]

Well, pondering some more, I'm reminded that the simplest things are often the most difficult to perceive.

One straightforward way to determine parallel noise is simply to measure the noise with an open-circuited input, and r.m.s.-subtract the previously measured series (voltage) noise (measured with the input shorted). The two noise sources are usually only weakly correlated, at least in the audio band.

The parallel noise will have itself a non-flat spectrum, and of course the RIAA equalization will weight this profoundly, as it does the series noise. But those with spectral analysis capability can characterize it further.

Once the two noise spectra are so determined, absent any nonlinear effects of source component loading, it should be possible to determine, for a given cartridge load, what the system electrical noise performance will be.

And it should be added, the cable capacitance should also be included in the analysis.


Brad Wood

[grufti]

SY worked through some of the real life noise considerations in his article on His Masters Voice on this site. He also has a calculator for phono preamp noise via RIAA filtering on his own website. And yes, I realize, that he may not have addressed every possible angle.

[bcarso]

Thanks grufti. I will check out Stuart's piece.

Brad

[bcarso]

Grufti, I looked for "His Masters Voice" and as well "His Master's Voice" and it seems nothing came up of relevance (or at least the thread titles seemed only distantly related). Could you point me more specifically?

Thanks,

Brad

[kevinkr]

Brad,
Here is the article you are looking for:
His Master's Noise: A Thoroughly Modern Tube Phono Preamp

[bcarso]

Thanks kevinkr. I had just found it (the unintentional misdirection having been the transliteration of the SY pun on His Master's Voice) as His Master's Noise. I will report back after perusal.

Best,

Brad

[ashok]

Hi Brad,
This must be what you are looking for
SYclotron Audio

cheers.

Edit: Oh..Oh.. am I so late !

[bcarso]

Thanks again for the link kevinkr. A couple of comments:

At least one important omission: recognition of the large excess series noise in tubes at low frequencies, making the equivalent noise resistance number usually far from realistic, especially in light of the substantial equalization boost at low frequencies. The article is also, unfortunately, not very apropos of my topic, since as far as I could perceive there is no discussion of series versus parallel noise. This is not an indictment of SY's article or design, as tubes are among the preeminent devices for low parallel noise (tiny grid currents and associated shot noise).

There is also no discussion of the contribution of the termination resistor to the noise. But again, this is wholly appropriate, as a typical low-inductance and low-resistance MC cartridge strongly shunts the termination resistor's thermal noise. As mentioned above, the situation with high inductance cartridges is very different.

[bcarso]

Originally Posted by ashok

Hi Brad,
This must be what you are looking for
SYclotron Audio

cheers.

Edit: Oh..Oh.. am I so late !

Thanks anyway!