As I pointed out the inductance and resistance of the cartridge create a zero at R/L in the transfer function. The negative resistance input amplifier lowers the R to reduce the frequency of the zero and extends the low frequency response. With conventional EQ you are simply trying to amplify the falling response and you amplify the noise by the same amount.
There will be no difference in noise level (the simulation confirms this). Active emulation of negative resistance is also gain.
Apply the same noise transformation on figure 16 as was used on figure 14 and you will see that the equivalent input noise is the same, except for the extra noise voltage from the resistors in the positive feedback path. As the circuits have to be equalized to get the same transfer from input to output, the same holds at the outputs.
internal installation of my phono stage v.5. Details are here [ https://www.patreon.com/posts/insaidnyi-ot-vip-72574505 ]
You cannot use noise figure to choose the bias current, because noise figure does not take into account the noise voltage generated by the shot noise of the base current on the inductive part of the cartridge impedance (which is much larger than the resistive part). As a result, the bias current, which really provides the minimum RMS output noise, turns out to be much less than that which seems to be optimal according to the noise figure optimization criterion. Detailed mathematics see in my book [ freely @ https://www.patreon.com/posts/razdaiu-i-chto-67628599 ] page 67 [ for MM with 700 mH inductance real optimal bias current is 4,25 μA not 100! but with zero inductance, the optimal current is really 153 μA -see page 68]
Last edited:
Op-amp circuits with positive feedback are rare, so their analysis may not be obvious. When positive feedback is introduced, the noise gain increases. Thus, all sources of noise will be amplified, as well as the useful signal. The situation is no different from using conventional EQ. Netzer's article says that a circuit with a negative input resistance "performs better than a high-input impedance amplifier and equalization network". But when compared with a transimpedance amplifier and equalization network, parity is obtained.
Attachments
Try and see (measurement is confirmation). But paralleling bjt is paralleling not only emf (nice) but also a shot base noise (bad), so in reality it is good only for lowering equivalent Rbase noise. The better is using jfets, as I said earlier repeatedly.
Last edited:
The input trz collector resistor in ne5532/4 is 12...13k...that's around 1mA collector current which I was told is ideal for shot noise...
While I don't doubt your capabilities, it seems that you choose to say whatever helps you narative...
I hope Leoniv will get his topic out of here cause I don't want to polute this heavenly topic once more.
While I don't doubt your capabilities, it seems that you choose to say whatever helps you narative...
I hope Leoniv will get his topic out of here cause I don't want to polute this heavenly topic once more.
" MM with 700 mH inductance real optimal bias current is 4,25 μA not 100!"
So, is every other preamp designer wrong? With that small collector current you would have severe loss of beta (Hfe). If you look at a chart of beta vs collector current you see that as the collector current decreases the beta also decreases and at zero collector current you have zero beta. With 100uA collector current the transistor beta may be down 20-30% from the value at 10 mA. but at 4.25 uA it would be way down. Marcel had some measured values of cartridge impedance and it didn't exceed 50 kOhm. In the actual preamp the cartridge is shunted by the load resistor which will reduce this value further.
So, is every other preamp designer wrong? With that small collector current you would have severe loss of beta (Hfe). If you look at a chart of beta vs collector current you see that as the collector current decreases the beta also decreases and at zero collector current you have zero beta. With 100uA collector current the transistor beta may be down 20-30% from the value at 10 mA. but at 4.25 uA it would be way down. Marcel had some measured values of cartridge impedance and it didn't exceed 50 kOhm. In the actual preamp the cartridge is shunted by the load resistor which will reduce this value further.
You are wrong with it, imho you do not understand bjt noise sources. Simply measure or simulate in Spice/Microcap (but with standard IEC 268-15 input MM equivalent 500 mH inductance (16k @ 5 kHz), not input shorted).
Last edited:
You are wrong (or live in a hopelessly frightened past}. Prooflink attached.
Attachments
You forgot you're talking to a guy that knows everything and that he is also smarter than everybody else on the Planet Earth...
https://www.electronics-notes.com/a...mponents/transistor/current-gain-hfe-beta.php
Attachments
The whole discussion about optimal collector bias currents doesn't make much sense to me without a specified hFE and psychoacoustic weighting or spot noise frequency. To give you some examples of calculated theoretically optimal collector currents, all for 700 mH, 1 kohm cartridges (the 700 mH doing much more than the 1 kohm) and negligible base resistance and 1/f noise:
RIAA + A weighting:
hFE = 100: IC ~= 15.1 uA
hFE = 600: IC ~= 37.1 uA
hFE = 1200: IC ~= 52.5 uA
RIAA + ITU-R 468 weighting:
hFE = 100: IC ~= 11.3 uA
hFE = 600: IC ~= 27.6 uA
hFE = 1200: IC ~= 39 uA
Spot noise at 20 kHz:
hFE = 100: IC ~= 2.92 uA
hFE = 600: IC ~= 7.15 uA
hFE = 1200: IC ~= 10.1 uA
The optimal current is inversely proportional to the source impedance, so you can multiply everything by 1.4 for a 500 mH cartridge. The values for 20 kHz spot noise are probably too low when you use a 47 kohm passive termination resistor.
RIAA + A weighting:
hFE = 100: IC ~= 15.1 uA
hFE = 600: IC ~= 37.1 uA
hFE = 1200: IC ~= 52.5 uA
RIAA + ITU-R 468 weighting:
hFE = 100: IC ~= 11.3 uA
hFE = 600: IC ~= 27.6 uA
hFE = 1200: IC ~= 39 uA
Spot noise at 20 kHz:
hFE = 100: IC ~= 2.92 uA
hFE = 600: IC ~= 7.15 uA
hFE = 1200: IC ~= 10.1 uA
The optimal current is inversely proportional to the source impedance, so you can multiply everything by 1.4 for a 500 mH cartridge. The values for 20 kHz spot noise are probably too low when you use a 47 kohm passive termination resistor.
Last edited:
