[Need Advice] Low Noise Amplifier with source resistance 1 kohm

[Aamir M]

I am designing a low noise amplifier with 1 k source resistance with bandwidth 0-400 Hz. My amplifier goes into a transformer amplifier stage with effective impedance 0.5 ohm in series with 0.5 H, in parallel with 1.6 uF capacitor. The transistors have been chosen and biased accordingly for minimal noise output from the transistors.

(a) I used ac coupling capacitor of 24 mF. The gain is found to somewhat vary when the excitation frequency of the lock-in is varied.

(b) How would I calculate how much noise is added from the 100 kohm resistor and the 10 kohm in the emitter follower, and from the 3.3 kohm in the diff. amplifier.

 

[WimRFP]

Hello,

Just add a voltage source in series with the resistor, or a current source in parallel. For the voltge spectral noise you can use sqrt(4kTR) [V/rtHz]. When you use a 1V source (in AC simulation), and the effect at the output is 0.1V, then the voltage spectral noise from the resistor (sqrt(4kTR)) is attenuated with factor 10. When you do this for all resistors, then you know the resistor that introduces most noise. If you want to know the total noise, you need to add the effect of all resistors based on non-coherent addition.

So total spectral noise voltage is sqrt(V1^2 + V2^2....+ Vn^2).

The 10 kOhm resistor will add few noise as it looks into the emitter (has low impedance). The 100 kOhm resistor (R4) will add few noise when C2 has reactance < 3.3 kOhms. However for low frequency (where C2 has high impedance), the 100 kOhm resistor (R4) will add lots of noise, as it looks into the high input impedance of the emitter follower. Virtually all noise EMF will appear at the basis.

For high frequency, R4 and R2 are in parallel, so effect of R4 is negligible.

Of course the noise component in the base current of the emitter follower will create a separate noise voltage across R4 also.

As you are working at low frequency, you may check the noise specification of the resistors (if present). Noise production from resistors at low frequency can be far above the thermally generated noise. Same is valid for the noise components in the BJTs.

Note that the output impedance of this amplifier will be << 10kOhm as you look into the emitter.