The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> SNR: AC VERSUS TRAN
https://designers-guide.org/forum/YaBB.pl?num=1207565626

Message started by superpfff on Apr 7^th, 2008, 3:53am

Title: SNR: AC VERSUS TRAN
Post by superpfff on Apr 7^th, 2008, 3:53am

Hi.

I designed a CTIA (Capacitance TransImpedance Amplifier) based on a Common source (CS) amplifier using cascoded devices. It is a classical cascoded common source biased by a low voltage current mirror. Theory shows:

The weakness, compared to the classical CS, is the voltage excursion. We lose 2 times VDSsat. It means that the capacitance will get less "charges" with the cascoded device than with the simple CS. To me, it implies that SNR has been damaged because of the lack of excursion.

What's your opinion? Some tips to put that on equations?

Regards

Title: Re: SNR: AC VERSUS TRAN
Post by thechopper on Apr 7^th, 2008, 7:15pm

Hi,
I think you are just limiting the output dynamic range while improving (or at least keeping) the SNR. Assuming your cascode does not add any noise, then both the input noise and signal will get amplified by the same number (disregarding frequency domain considerations) so SNR is the same. It is just you can accomodate less signal (and less noise) with the cascoded version.

Hope this helps
Tosei

Title: Re: SNR: AC VERSUS TRAN
Post by superpfff on Apr 8^th, 2008, 7:13am

thechopper wrote on Apr 7^th, 2008, 7:15pm:

Hi,
I think you are just limiting the output dynamic range while improving (or at least keeping) the SNR.
Assuming your cascode does not add any noise, then both the input noise and signal will get amplified by the same number (disregarding frequency domain considerations) so SNR is the same. It is just you can accomodate less signal (and less noise) with the cascoded version.
Hope this helps
Tosei

Hi Tosei. thx for the answer.
Input signal and input noise aren't amplified by the same number in cascoded devices. Because of quadratic relationships, noise progression is of squareroot type. So that at the end, cascoded devices improves SNR.
So my question is: is the output dynamic range improves or not the SNR? I can't prove it, but it seems to me that, in large signal, the more dynamic, you have the better the SNR is.