The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Transient noise contribution list
https://designers-guide.org/forum/YaBB.pl?num=1340278043

Message started by michaelFCT on Jun 21^st, 2012, 4:27am

Title: Transient noise contribution list
Post by michaelFCT on Jun 21^st, 2012, 4:27am

Hello everyone,

I have an issue regarding transient noise.

I'm simulating a simple active RC integrator (transistor level amplifier). I apply a sine wave at the input and get a sine wave at the output. Perform an FFT on the output and I get a given THD.
In the spectrum the harmonics are clearly visible.

I then activate tran noise (only Fmax specified, Fmin left blank), set noise contribution to off and select all devices in my testbench. So, basically what I expect is the same result as before. But now I get a much improved THD (+15dB).

I then deactivate the noise contribution list, i.e., I perform a normal tran noise simulation with all noisy elements active. Now I get a worse SNR (expected) but my THD is still better than with tran noise off.

What is tran noise doing if all noisy devices are off?
May I be misinterpreting the noise contribution list?

One final question: This question was posted in another message but nobody answered. Is there any way to find what who are the noise contributors and their contribution during a tran noise simulation?

Thanks.

Title: Re: Transient noise contribution list
Post by Ken Kundert on Jun 21^st, 2012, 10:06am

Can I ask why you are using transient noise. Wouldn't conventional noise analysis be faster, more accurate, and naturally generate the list of individual noise contributors?

-Ken

Title: Re: Transient noise contribution list
Post by Geoffrey_Coram on Jun 21^st, 2012, 11:33am

I second Ken's question.

But I'll take a stab at your original question by suggesting that perhaps enabling transient noise causes the simulator to decrease tolerances, because it expects to have to resolve the tiny noise voltages added. Or perhaps the Fmax specification constrains the timesteps, because the simulator expects to be generating a new random number every 1/Fmax seconds (for every noisy device).