In data acquisition experimenters are often confronted with the unfavorable situation of detecting a signal too small to be recognizable - the signal is embedded into noise which is almost as high as the signal itself, or even higher than it.

In order to cope with this situation one has to increase the signal to noise ratio (SNR) to a minimum level of 3. In principle, this can be accomplished by reducing the noise (using better equipment, or a different experimental set-up). However, the reduction of noise has principal limits which are given by physical laws (i.e. the thermal noise cannot be removed without cooling the entire measuring device to absolute zero).

Another approach to increase the SNR is to repeat a measurement many times and add all measurements. Given that the signal is stable enough over a longer period, this will finally cancel out the random noise, while the signal itself accumulates - thus increasing the SNR.

This method is called time averaging and is a strategy commonly applied when signal levels are (too) low and the signal is stable over the interval of repeated measurements. One example is the data acquisition of NMR (nuclear magnetic resonance) signals.

Download
English version	[528 kB]
German version	[528 kB]
After downloading please unpack all files of the zipped packages and start the executable.

The program Accumulation of Spectra shows how the SNR of NMR (nuclear magnetic resonance) signals can be increased by accumulating many measurements.