
| PROCESSING YOUR DATA |
| Signal-to-noise ratio (S/N) |
|
You generally want to maximize the intensity of the signals you record relative to the baseline noise. One way of achieving this is to simply record data for a longer period of time, but there is an obvious time cost to doing this. For example, to double the signal-to-noise ratio, or S/N, you must record four times as many scans. To quadruple S/N, you need to increase the acquisition time sixteen-fold, and to achieve a 10x increase in S/N, the acquisition time must be lengthened 100-fold. Obviously, increased signal quickly becomes very costly in terms of time. |
|
FID's are composed of coherent signal due to the the resonance of the of the nuclei in your sample, and random noise arising from the electronics. To compare an FID of poor signal-to-noise ratio (S/N's) to a better one, move the mouse into and out of the figure areas below. To view the effect on a spectrum, scroll down or click here. |
| The spectrum below is a noisy one; move the mouse over the figure to see the S/N improve by a factor of 10. Remember that the better spectrum took 100 times as long to acquire. You should try to achieve the best S/N possible given your time constraints, because noisy spectra can lead to several problems: peak area measurements are subject to larger errors, and small peaks may be missed altogether. |