The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Charge Amplifier
https://designers-guide.org/forum/YaBB.pl?num=1242305359

Message started by MarcoC on May 14^th, 2009, 5:49am

Title: Charge Amplifier
Post by MarcoC on May 14^th, 2009, 5:49am

Hello there,
I study the realization of a low noise transresistance amplifier using a charge amplifier.
I've seen that the periodical reset in a charge amplifier cause a variation on the output noise. The reset cause a modulation of the noise with a square wave, thus it generates a folding effect of the wideband white noise. Thus the output noise is white shaped and is higher than the expected one.
Do you know some design tip to overcome this problem and reduce the output noise?

Thank you,
MarcoC

Title: Re: Charge Amplifier
Post by MarcoC on May 27^th, 2009, 6:19am

so I saw that nobody can help me; nevermind, I try alone.
Thank you ;D
Now I have another big problem. My system is strongly dependent by the input stray capacitance, and I want to overcome this addiction. Anyone know some circuit or technique to decouple the system from the input capacitance?

Thanks a lot

Title: Re: Charge Amplifier
Post by subgold on May 27^th, 2009, 10:15am

hi,

the noise folding effect is the basic property of the switched capacitor circuit. i dont think you can escape from that problem. you can only reduce it by reducing the noise bandwidth of your noise sources (using larger capacitor, slower amplifier...)

as for the input stray capacitance, it shouldnt be a problem if you have good virtual ground at input. maybe you can show your schematic.

Title: Re: Charge Amplifier
Post by MarcoC on Jul 8^th, 2009, 6:04am

Hi Subgold,
sorry for this latency, I was warried by some personal problems last month.
You're right, the folding effect is a basic property and I can't overcome it. But I have seen a strange thing and I'd like to have a confirm.
Simulating only the CA the noise folding is due to reset; but adding a sampler after the CA, the noise folding is due to the sampling operation.
I think in the first case the reset introduces a shape in time, thus a correlation in frequency, showing a folding effect. However, in the second case, the sampler samples the noise out of the reset time and the reset effect vanishes (but the folding due to sampling is introduced).
What do you think about this?

As for the input stray cap I show you my schematics. If you study the noise of this circuit, you will see that the equivalent input noise power is amplified about by (Cpar/Cf)^2, thus the stray capacitance become a key parameter.

Title: Re: Charge Amplifier
Post by subgold on Jul 9^th, 2009, 3:21am

MarcoC wrote on Jul 8^th, 2009, 6:04am:

hi,
i'm not sure how you simulate the circuit, but adding a sampler will certainly gives different result. anyway, the folding effect doesn't vanish in either case. the sampler shows the noise at a certain time point, but the noise is still folded.

as for the input stray cap, i was talking about the correct functionality, i didn't think of the noise performance. but this noise power amplification is confirmed by correct simulation or just your calculation? i would think the input stray cap doesn't amplify the noise power. instead, it adds one more noise source to the circuit, whose noise power is something like 2kTCpar/Cf^2. it sounds like your Cpar is too large compared with Cf.

Title: Re: Charge Amplifier
Post by MarcoC on Jul 9^th, 2009, 3:48am

hi,
i simulated the noise performance in time, doing a transient analysis and then computing the spectrum. I'm agree with you, the folding don't vanishes but without the sampler is due to reset and with the sampler is due to sampling operation. A confirm came yesterday by a friend, a math expert. He said that a cyclostationary process with period T, sampled with period T and offset phi, create a discrete process that is stationary. Thus the folding isn't more an effect of the cyclostationary introduced by the reset, but, now, is due to the sampler. This sounds good with my simulations.

As for the stray cap, you're right, the Cpar is too large compared with Cf, but i can't use a larger Cf so i try to reduce the dependency on Cpar.
The simulations confirm that output noise power depend on stray capacitance (this time AC noise simulations neglecting the effect of the reset).
Aniway, there is some technique to reduce the noise effect of the stray cap?