The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Other Measurements >> Filtering ADC inputs
https://designers-guide.org/forum/YaBB.pl?num=1133863744

Message started by vivkr on Dec 6^th, 2005, 2:09am

Title: Filtering ADC inputs
Post by vivkr on Dec 6^th, 2005, 2:09am

Hi,

I have read several documents mentioning the importance of filtering the signal generator outputs before feeding these to the ADC inputs. I have a couple of questions here:

(1) As bandpass filters are the best in terms of removing all out-of-band harmonics and noise, they are preferred over lowpass filters. However, this introduces the problem of having to provide a DC bias to the filtered signals.
Are there any critical points to consider when choosing the signal source to provide this DC? I think its response time is of some concern even though it is common-mode.

(2) Passive filters used for filtering the inputs are very sensitive to the termination impedance. Since the ADC input presents a very rapidly switching capacitance, how does one handle this? It is tempting to assume that the effect is the same as a resistance R = (1/Fs.C) but this is not the case from a dynamic point of view, or is it?

(3) Can one use a fixed termination resistor and connect the input of the ADC to it all the same?

(4) How does one generate high-purity clocks? If the clock signals from the generator are passed through an RF transformer to create differential clocks and then filtered, then one gets sinewave signals. Is it fine to use these after "digitizing" them on-chip?

Thanks
Vivek

Title: Re: Filtering ADC inputs
Post by Paul on Dec 6^th, 2005, 1:18pm

Hi Vivek,

here some comments to your questions:
(1) You should be able to provide good DC bias using bias T's. You just have to make sure they do not attenuate your input signal at the desired frequency.
(2) Switching of the input capacitors of the ADC on a relatively high impedance filter output will generate ripple which may hurt your measurement precision, especially if the node does not settle at the desired precision in one clock cycle. Would you be able to use low-noise buffers at the filter output?
(3) I do not fully understand this question.
(4) If you need differential (CML-type) clock signals, a buffer chain with sufficient voltage gain should suffice to sharpen the clock edges. I imagine this is what you call "digitization".

I hope this helps a little

Paul