Signal to Noise ratio and Data Collection Sampling Rate
Posted: Tue May 29, 2007 2:42 pm
by Rob Burgess
Should decreasing the sampling rate in the data collection for a standard UV detector improve my signal to noise ratio?
If this does occur, why does it occur (is there a scientific/mathematic reasoning to this)?
Posted: Tue May 29, 2007 4:13 pm
by tom jupille
The answer depends on the dominant source of noise and on how you're measuring that noise.
If the dominant source of noise is electronic ("buzz"), and if you are measuring over a fairly short time (e.g., a couple of peak widths on either side of your analyte), then a slower sampling rate should improve S/N. On the other hand, if your dominant source of noise is temperature or flow variations (which tend to produce more long-term noise) and/or you are measuring noise over several minutes, then I would expect S/N to remain about the same.
Decreasing the sampling rate is comparable to increasing the time constant (in other words, it has the effect of "averaging" the signal over some period of time). As such, it has a major effect on short-term noise, but (unless you are sampling veeeerrrrryyyy ssssllllloooowwwwlllllyyyy) probably doesn't do much to attenuate either long-term noise or drift.