Transimpedance Ampifier - Input Referred Noise Current Density

Also, not sure if I should run my noise simulation up to the 3dB bandwidth of the TIA or 2X the 3dB Bandwidth as some have suggested.  I ran it to my 3dB Bandwidth.

Thanks.

What the noise analysis normally outputs is the output noise in V/sqrt(Hz) or A/sqrt(Hz) - this is essentially the sqrt of the noise power spectral density in a 1Hz bandwidth, so I assume this correspond to what you mean by RMS noise (looking at this wikipedia description, I suspect it is). Unfortunately if you have an input current source and an output voltage, both the Direct Plot form and the Print Noise Summary get a bit confused about the units. If use use Direct Plot->Main Form and pick input noise, and the output in V, it's actually plotted in A/sqrt(Hz) - the axes are correctly labelled. If you request the output in V**2, then it actually labels the axes in V**2/Hz, but it's really in A**2/Hz (unfortunately).

If you pick Direct Plot->Equivalent Input Noise or Direct Plot->Squared Input Noise, these do end up with the right units on the graph.

When using the Noise Summary, if you pick V, then the input referred noise is in A/sqrt(Hz) - (even though it says V/sqrt(Hz). If you integrate it over a frequency range, then it will be in Amps in that bandwidth.

As for the correct frequency range, I can't comment - that depends on the specification of your system.

Andrew.

Hi Andrew,

Sorry for the bump on an old thread, but did this ever get fixed in Noise Summary or do I still need to do a mental conversion?

Thank you,

PB1956

Hi PB1956,

No (I just quickly re-checked). I didn't report it to R&D at the time. Given that I answer things here in my spare time, I cannot always get the time to package up a test case and file it with R&D to be fixed; generally anything that is a bug should be reported to customer support so that we can prioritise it as a real customer issue.

In this case I'll go ahead and file a CCR to get the units sorted. (Now done: CCR 2114281 for the record)

Regards,

Andrew.

Hello all,

I  have the same question. I'm designing a chopper amplifier .My design input is current and output is voltage. I have to  use PNOISE  analysis. For PNOISE analysis i use PORT instead  common current source in input and In output. Pnoise analysis give the input referred noise in V/sqrt(Hz) while input is in current and output is in voltage. can someone please guide me on how to compute the input refered noise in A/sqrt(Hz) in Pnoise analysis? 

OK, this is rather similar to your other post, and is also breaking the forum guidelines.

The Noise Summary form still shows the results rather ambiguously - input referred noise is actually in A^2/Hz (if you picked V^2 on the noise summary form), or in A/sqrt(Hz) (if you picked V on the noise summary form). The Direct Plot form (as I mentioned in the other post) has been fixed (I didn't check precisely when it was fixed). So it looks as if part of my CCR (2114281) hasn't been fixed yet (the noise summary doesn't clearly show the correct units for input referred noise if the input was a current and output a voltage, but the numbers are right).

As I mentioned in your other post, asking this in a new thread with sufficient information would rather help pinpoint precisely what you're running and which version.

Andrew