Suppose that an input signal is very low in voltage, and one wishes to use quantizer such as 1-bit ADC. Thus, one wishes to have quantized values to be very low, also.
What is the low-voltage limit of ADC here? (theoretical)
2
-
\$\begingroup\$ Are you asking for a theoretical lower limit? \$\endgroup\$– ArsenalCommented Nov 9, 2015 at 16:53
-
\$\begingroup\$ Yes I am looking for theoretical limit. \$\endgroup\$– NCLCommented Nov 9, 2015 at 16:58
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
3
With a sigma delta converter, which is a 1 bit ADC, with noise shaping feedback, and DSP filtering, you can achieve any arbitrary resolution.
That means your actual usable accuracy, and repeatability, are limited by noise, and long term drift. Once you have a few LSBs of noise, or of drift from minute to minute, then there is little point in increasing the resolution further.
If you measure in a 1Hz bandwidth, then even in theory, at room temperature, your noise power will be at least -174dBm. In practice, it will be much, much more. You can convert from noise power to noise voltage through the input impedance of your chosen 1 bit ADC.
-
\$\begingroup\$ Where is that -174 dBm coming from? \$\endgroup\$– ArsenalCommented Nov 9, 2015 at 18:26
-
\$\begingroup\$ @Arsenal Johnson resistor noise, the OP did ask about theoretical \$\endgroup\$– Neil_UKCommented Nov 9, 2015 at 18:33
-
\$\begingroup\$ Yeah, but maybe he would be interested where the limiting factor is coming from instead of just a number - well at least I was. \$\endgroup\$– ArsenalCommented Nov 9, 2015 at 18:37