At what height from the ground are you proposing to mount the pressure sensor? i.e. does it have to withstand a 4m head of water or just the head due to the depth of the tank?
So you will have a 2-3m head of water acting on the sensor. Its full-scale range, as per the datasheet, corresponds to ~10m head (100kPa) and outputs 0-40mV over that range. Your tank head varies by only 1m, giving only a 4mV total output variation and you want a resolution of 1% of that, i.e. 40uV. That's going to take some sophisticated instrumentation-grade (i.e. expensive) amplification to process the sensor signal
IMHO the simplest solution would be to use something like a car fuel gauge sender, i.e. a float varying a resistance. The micro could do any necessary calibration to allow for non-linear effects.
Thank you so much out of the box solutions.
Dear alec_t please look in to the attached application pdf from preescale.com. what do you think about this amplification cct.
Its full-scale range, as per the datasheet, corresponds to ~10m
Can you please tell me where did you get this ~10m ( from your post #4)
It certainly is showing an opamp, but a 741 is unlikely to give you the precision and stability you need to resolve down to 40uV. Google for 'instrumentation opamp'.
Thus a 10mV change in your sampled signal.
Assuming a 5V supply is also the Vref for the ADC and you have 10bit sampling => 5cm is equivalent to 2 out of 1024. Thus it is barely doable ( with simple sampling) without any differential amplification.
If you use oversampling I feel you can do it adequately without any instrumentation amps.