Because of the inherent Johnson noise of any resistance, phono cartridges have a noise floor. It’s useful to know what that is so that design targets for phono preamps can be set. The cartridge resistance is surrounded by several reactances (e.g., the cartridge series inductance, the preamp input shunt capacitance…), so the effective noise resistance varies with frequency- you can’t just plug the DCR into the Johnson equation and get the right answer. Also, the noise from the cartridge is modified by the RIAA network in the preamp. So if a cartridge has an inherent noise of (say) 2uV, it really doesn’t matter if the preamp has 0.5uV or 0.005uV of noise, it’s buried behind the cartridge noise. In order to calculate this accurately, the audio spectrum needs to be broken into slices, calculations done for each slice, then the noise voltages RMS-summed. This is a major pain to do by hand, but trivial for a spreadsheet.
Here’s a handy-dandy little spreadsheet riaa calculator which allows you to calculate the maximum signal to noise attainable for a phono cartridge. Input the DC resistance, the inductance, the load resistance and capacitance at the preamp end, and the nominal output voltage, and voila, you know what you’d get with a perfect, noiseless preamp. This calculator is based on the equations given in the 1980 National Semiconductor Audio/Radio Handbook, using Q equations to transform the various series impedances (e.g., the cartridge L and R) into parallel impedances.
User inputs are shaded blue. I have used the spectrum slicing suggested by National, but you can vary that. The nice thing is that with the noise broken out into discrete bands, you can see the relative contributions from each frequency segment.
This was written in Office 2010 (xslx format), but if you have an earlier version of Excel, you can get a free converter from Microsoft.
Update: at the suggestion of Hans P, I have added an RIAA-weighted S/N as well as an unweighted S/N.
The calculator does not seem to be squaring the nanoVolts correctly on the 4th line. I tried some different values for DCR, such as 400 (V15V cartridge), and the squared answers were incorrect.
Ray, looks OK to me- it may be that because the cell is set to display zero decimal points, you see a slightly higher result for the square than you expected, but the calculation is correct. If it bothers you, you can right click on the cells, then use the “format cells” command to give you all the decimal places you’d prefer.
Hello Sy, trust you are keeping well.
I’ve taken the liberty of downloading your RIAA noise spread sheet and adding a facility for including the amplifying device noise current and noise voltage.
In your spread sheet, am I correct in assuming that the RIAA equalized resistance that should be used to calculate the total noise per band is the line marked ¦Z¦Ohm (and the line directly above it is the ‘unequalized’ source resistance)?
Great site by the way -I’ve also taken a look at the ASR site – nice!
Thanks for the compliments- I wish I could update content more frequently… 😉
I’ve found the NatSemi app note so no need to worry about answering my comment. Thanks