Post Go back to editing

# AD8139 wrong noise model

Category: Hardware

Hello,

this question has been asked 4 years ago, but since then no answer came up. I'm trying to simulate the noise at the output of this simple circuit

I've imported the model download from the Analog Device website and included in TINA spice. I'm getting the same results also with LT-Spice, so it is not a problem of the simulator.

The noise at the output of the circuit is the following:

Which makes no sense. First of all, I don't see any 1/f noise. Second, the flat part is 18.68 nV/sqrt(Hz).which is wrong. Indeed, if I do the calculations on my own and consider only the white noise, I get the following contributions:

Consider RG1=RG2=RG and RF1=RF2=RF

1) RG1 and RG2

Sout = 2*4*KT*RG*(RF/RG)^2  = 2*4*KT*RG = (2.5 nV/sqrt(Hz))^2

2) RF1 and RF2

Sout = 2*4*KT*RF  = (2.5 nV/sqrt(Hz))^2

Sout = Sv*(1+RF/RG)^2 + 2*Si*RF^2 = 2*Sv + 2*Si*RF^2 = 2*(2.25 nV/sqrt(Hz))^2 + 2*(2.1pA/sqrt(Hz))^2*(200)^2 = (3.2 nV/sqrt(Hz))^2

By summing all the individual contribution I get

Sout =  (2.5 nV/sqrt(Hz))^2 +  (2.5 nV/sqrt(Hz))^2 + (3.2 nV/sqrt(Hz))^2 = (4.8 nV/sqrt(Hz))^2

which is way different.

Can you help me?

## Top Replies

• Apologies for any inconvenience this has caused. I will raise this to the team so we could verify it and have it fix. Thanks for raising this issue.

Regards,

Mae

• Actually I did a mistake in the calculation of the output noise due to the AD8139:

Sout = Sv*(1+RF/RG)^2 + 2*Si*RF^2 = 4*Sv + 2*Si*RF^2 = 4*(2.25 nV/sqrt(Hz))^2 + 2*(2.1pA/sqrt(Hz))^2*(200)^2 = (4.5 nV/sqrt(Hz))^2

So the total noise is:

Sout =  (2.5 nV/sqrt(Hz))^2 +  (2.5 nV/sqrt(Hz))^2 + (4.5 nV/sqrt(Hz))^2 = (5.72 nV/sqrt(Hz))^2

Anyhow the noise model is still not correct.

• Hi  ,

Apologies for any inconvenience this has caused. I will raise this to the team so we could verify it and have it fix. Thanks for raising this issue.

Regards,

Mae