SPICE Simulation File SBOC525: 120k Ohm series resistor in noise analysis

not seeing that 120k in this document anywhere? 

Hi Seeran,

the 120k resistor gives a resistor noise of SQRT(4 x k x T x R x B) = 6.3µVeff. Compared to the sensitivity of micro of -35dBV this corresponds to a signal-to-noise-ratio (SNR) of micro of 20 x log (17.8mV / 6.3µV) = 69dB. This would be a very good electret micro, because most such micros have a somewhat lower SNR of 60dB or even less.

So, yes, the 120k resistor could simulate the inherent noise of mirco.

But what about the input noise current of TLV6741? Would the 120k resistor create an additional error?

No. The input noise current of TLV6741 is 23fA / SQRT(Hz). Together with the 120k resistor it will see an equivalent resisance of 5.3k to signal ground. This will give a 20kHz broad band noise of about 23fA / SQRT(Hz) x SQRT(20kHz) x 5.3k = 17nVeff, which is much smaller compared to the inherent noise of micro.

Kai

Good morning Michael,

there's a link in the appnote:

sboa290.pdf

Kai

Oh I found it now, it is in the sim file at the bottom, 

That must be an assumption on the microphone source impedance - however, for noise and signal gain you need to consider the 5.9kohm biasing resistor - if the 220kohm is real, that is a big signal attenuation with that biasing R. 

I actually think that is probably a mistake, spec table from app note, 

\

In the other sim file it has an model for the microphone, opening that shows this?

Hi Kai,

thank you very much for your reply ! Should the mid-supply voltage bias (voltage divider with two 200k Ohm resistors) also be considered in the source impedance equation? Since VCC is 5V, the equivalent circuit for the voltage bias would be a 2,5V voltage source in series with a 100k Ohm (R2 in parallel with R3) connected to the non-inverting input, so it adds 2,5V to the voltage sigal from the mic. Does this not increase the source impedance by 100k Ohm? 

If yes, the current noise of an Op-Amp because more relevant.

Seeran

Hi Seeran,

the series circuit of C1 and R1 is in parallel to this 100k impedance and C1 is a short circuit a higher frequencies.

But the dynamic source impedance of FET-amplifier within the electret microphone modul should be much lower than 1k. So, at the usual audio frequencies the source impedance seen by the OPAmp should be in the 100R range. It mainly depends on the drain source current of FET within the electret modul. So, by decreasing R1 the drain source current can be increased and by this the noise of electret microphone modul can be lowered. It's worth to experiment a bit with R1, if lowest noise is desired.

Kai

Seeran,

Do you need any more help? If not please mark one the the answers as the solution.

Hi Kai,

thank you for the information. If I create an equivalent circuit for the mid-supply voltage bias it would result in a circuit as seen in the picture below? Would the source impedance not be R1 + R2 II R3 ? If I am wrong, please correct me. I am considering to use the INA1620 as a pre-amp because of its low voltage noise density but if the source impedance is much higher I would consider an Op-amp with lower current noise density.

Seeran 

Hi Seeran

this is the equivalent noise impedance of the input:

Kai

You are welcome, Seeran :-)

Kai