Phono stage hiss - Measurements

[thumb5]

[I wrote a long, rambling post and somehow managed to delete it!  Argh, will try to put it together again.]

Musing on the shape of the noise spectra both @Confused and I posted, it reminded me of the RIAA playback equalisation curve:



That being the case, I wonder if what we're hearing is just white noise that's been RIAA equalised?

Doing a bit of Googling, I found an interesting article that talks about theoretical noise limits for phono stages: https://www.renardson-audio.com/phono-1.html.  This quotes an earlier article published in Wireless World (1972) from which:

...ignoring amplifier noise the signal to noise ratio for a 50Hz to 20kHz bandwidth relative to 2mV at 1kHz was 58.5dB for circuit A and 72dB for circuit B.

The results are for a phono stage with a 47 kΩ resistive loading, and take into account the RIAA equalisation.

The "circuit B" mentioned has the loading connected across the input terminals, which is said to be the most usual design.  Based on something Mathieu Pernot mentioned recently at OAC, namely that the input loading can be easily measured with a digital multi-meter, I guess the Expert phono stage has this configuration.

Those theoretical limits are largely due to thermal (Johnson-Nyquist) noise from the cartridge itself and the loading resistor.  This thermal noise has a spectrum that's similar to white noise.

Working out how much noise a resistor generates just by sitting there at room temperature is quite surprising.  For example, the 47 kΩ loading in my phono stage generates about 4 µV RMS across the audio bandwidth.  In the context of a 1600 µV input level that's quite significant!  In the measurements @Confused and I posted earlier, where the cartridge is disconnected, this thermal noise is going to be present at the input to the phono stage and will be amplified by whatever gain it's configured to use (and then RIAA equalised).

For lower loading resistors, the RMS noise will be reduced (proportional to the square root of the resistance).  So for example @Confused's 30 Ω loading resistor would produce about 32 dB less noise than my 47 kΩ.  On the other hand, his phono stage is set up to be about 4 times more sensitive, which is about 12 dB, so if we look just at thermal noise from the loading resistor, without the cartridge present, @Confused should have about 18 dB better signal-to-noise ratio.

With the cartridge in place the effect is likely to be less marked, but it seems that using a high load resistance could be a contributing factor to phono stage noise.  Next time I get a chance I'll try dropping it down to one of the lower values and measuring again.