The LM4562 in a phono stage.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Maybe it was the way I was attempting to make a comparative noise measurement by reading the level at a specific frequency on a spectrum display. Perhaps a screen shot of the measured noise will help:

For RIAA EQ circuit noise measurements, all you need to do is add in the source resistance (about 1k ohm) of a phono cartridge to get a more accurate idea of actual system noise. You should find that your circuits have a SNR of around 90 dB this way, which is basically silent since the record surface noise (plus TT noise) is at best -70dB. Since the MM source impedance is around 1k ohm, then generally speaking an FET input op amp will give better results for noise. I like the OPA627 myself.
 
For RIAA EQ circuit noise measurements, all you need to do is add in the source resistance (about 1k ohm) of a phono cartridge to get a more accurate idea of actual system noise. You should find that your circuits have a SNR of around 90 dB this way, which is basically silent since the record surface noise (plus TT noise) is at best -70dB. Since the MM source impedance is around 1k ohm, then generally speaking an FET input op amp will give better results for noise. I like the OPA627 myself.

You are implicitly neglecting the cartridge inductance, but the impact of the cartridge inductance is anything but negligible. If you want to minimize the integrated RIAA- and A-weighted noise, you can take the cartridge impedance at 3852 Hz as a good estimate of the effective source impedance. For a 1 kohm and 500 mH cartridge, that boils down to about 12 kohm. Modern FET op-amps and the NE5534A have a good noise match to impedances of that order.
 
(The vast majority of) MM cartridges are massively inductive, about 1/2 a henry, that dominates their source impedance above a few 100Hz.


To measure system noise, just plug the cartridge in. The RIAA curve partially cancels the rise of noise with frequency from the current noise into the inductance. The voltage noise from the preamp falls with frequency (as seen at the output) as it follows the RIAA curve, so is unlikely to make much hiss by comparison to the inductance/current noise, although the flicker noise portion ought to be prominent in the noise spectrum.
Its even more complex than this, as the 47k load resistance voltage noise will be more prominent at high frequencies too, as its effectively in parallel with the rising cartridge impedance. Synthesized load circuitry can address this latter issue.
 
Member
Joined 2014
Paid Member
(The vast majority of) MM cartridges are massively inductive, about 1/2 a henry, that dominates their source impedance above a few 100Hz.


.


I'm am edge case as mine go from 40mH to over 1000mH. Scott of course has a Grado which is so low you wonder how it works.



Very glad we are all being honest that there is only one sensible way to measure noise for MM stage :)
 
Why not just measure the noise with the cartridge as a source impedance?

That's the best you can do when you want to compare different phono amplifiers for a given cartridge. It can be inconvenient when you want to compare a phono amplifier's measured performance with calculated or simulated data, because you have to take into account the frequency-dependent ESR of the cartridge somehow.
 
Why not just measure the noise with the cartridge as a source impedance?

The main problem is that cartridges are very sensitive to ambient magnetic fields. (Not electrical fields, except in one notorious case) As related in Electronics for Vinyl, I found a mumetal screening can to be the minimum shielding for reliable noise measurements. Two nested cans (as sometimes used for MC step-up transformers) is better.
 
Don't forget the thermal noise of the 47 kohm termination resistor, it is usually of the same order as or greater than the cartridge's thermal noise.

If you want a good example of an incompetent design, look up the moving-magnet version of the Elektor Supra 2.0:

Supra 2.0 | Elektor Magazine

It uses a fortune's worth of ultra low noise op-amps to arrive at the highest noise moving-magnet amplifier ever, all because the designers didn't take the effect of current noise and cartridge inductance into account.
 
Don't forget the thermal noise of the 47 kohm termination resistor, it is usually of the same order as or greater than the cartridge's thermal noise.

Unless the circuit is an inverter, the termination impedance will be in parallel with the source, and will thus serve to reduce the effective source impedance and thus its Johnson noise, not increase it.
 
Why don't you short it and get rid of the noise altogether then?

Seriously, what matters is what happens to the signal to noise ratio. The termination resistor attenuates the signal and the thermal noise of the cartridge and it adds its own thermal noise. If it only attenuated the signal and the thermal noise of the cartridge by an equal amount, the SNR would stay the same, but since it also creates its own thermal noise, the signal to noise ratio degrades.

When you compare the total noise of the cartridge and the termination resistor to the noise you would have had with a cartridge and a noise-free termination resistor, you see a degradation of about 3 dB to 6 dB A- and RIAA-weighted. This is assuming that the cartridge only produces thermal noise, which means you don't play any records, otherwise record surface noise usually dominates. A nearly noise-free termination resistance can be made by cryogenic cooling or by using combinations of series and shunt feedback (as is very often done in modern radio receiver LNAs and almost never in RIAA amplifiers).
 
Other than headroom limitations, anyone want to comment on this design? using the LSK489 as the input in front of a LM4562? I am not much into LP's anymore but there is a resurgence for what ever reason. It used to be cool getting stoned and listening to LP's back in the day but to me now LP's are just a PITA to baby sit and change every ~20 mins.
 

Attachments

  • LSK489_Application_Note_p10.pdf
    55 KB · Views: 88
By the way, the real part of the impedance might well get larger when you connect a resistor in parallel to a lossy inductor. Depends on the losses and the frequency - for a lossless inductor, the real part would always go up.

Not sure that I follow - how would paralleling anything with anything else result in an increase of the real component of impedance? Real conductance always increases when you parallel more conductance. Conductance can't go down by adding more in parallel.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.