In this article at eetimes.com they show the signal chain for measuring an ECG.
The raw signal of an ECG contains noise and offsets at least a magnitude larger than the real signal. (Some few mV ECG, several tens of mV from power line noise and electrode offset and up to several hundred mV baseline wander due movement of the breast.)
This would intuitively make me do the filtering of the signal in front of the amplifiers, to avoid amplification of the unwanted signal components. In this article however they do the signal filtering after the input amplifier, the high frequency noise removal even after the second amplifier.
I can not really think of a reason why they would do this. The only thing coming to mind is the very high impedance of the signal source, but the filtering would not affect the signal source, because that frequency range would obviously be in the passband.
Am I missing some important reason why you would do the signal conditioning in this order?
Best Answer
Yes you are...
The front-end differential amplifier will be chosen such that it has a common-mode rejection level of many tens of dB, quite possibly in the region of 80 dB.
This diff amp converts a differential signal into a single-ended signal and any common-mode interference will be largely ignored.
If you were to put filters on both legs of the diff amp, to avoid a balance mismatch you would have to choose components (such as capacitors and resistors) that were matched to at least an equivalent level of -80 dB.
You could regard 1% capacitors as having potentially a difference in value of 2% and that, in dB terms could be regarded as -20 log(50) = -34 dB. In other words, you would never get decent differential common-mode performance with filters on each leg before the diff-amplifier.