How to calculate effective resolution of ciruit

[embedded.world@aol.com]

I am using 24 bit adc ADS1247 to measure voltage drop across thermsistor by passing current through it.
http://www.digikey.com/product-detail/en/PR103J2/615-1069-ND/2651611

Since thermistor value gets very large so I connected a 20K fixed resistor in parallel with it.
1. Now my current source is 50uA with initial error is +-6% with temperature drift +-100ppm/C.

2. Voltage reference is 2.048V with +-10mV & drift is +-15ppm/C.

3. Resistor = 20K +-0.01% with drift +-5ppm/C.

4. PGA = 2, operating range = -40C to +60C.

5. Minimum voltage step need to measure = 75uV.


Questions:
1. How to calculate effective resolution I will get finally.

2. Do I need any other factors to consider besides I mentioned above.

3. Do final resolution is equal to ENOB?

4. Will I be able to achieve minimum voltage step = 75uV with this adc



Crosspost:
http://forum.allaboutcircuits.com/showthread.php?p=736107#post736107

[Dago]

Start here: http://en.wikipedia.org/wiki/Propagation_of_uncertainty

[woodchips]

Provided the maximum voltage across the thermistor is less than about 600V, then yes. This is simply 75uV times 24 bit ADC resolution, + or -.

Is this useful? No.

What is more useful is what accuracy you can achieve, not to mention precision and repeatability. Resolution on its own is meaningless, just replacing a 16 bit ADC with a 24 bit ADC increases the resolution, but not necessarily accuracy, precision or repeatability. This is where you have to take into consideration the individual errors of each component.