doenisz
Junior Member level 2
Hi all,
I recently started working on noise shaping SAR ADCs, and I'm learning the ADC starting from scratch.
I'm confused about the distinction (or lack thereof) between residue and quantization error.
In the context of pipeline ADCs, for example, the residue seems to be the amplified version of the quantization error before being fed to the next stage, while in the sigma-delta ADC and SAR ADC, my understanding is that they refer to the same thing (eventually).
More specifically, for the SAR, the quantization error is the final value of the residue, because the residue during the testing of the initial bits will be large, but as the binary search converges, the digital output will also converge to the analog input, so the residue will decrease, and its final value will be the quantization error. Am I right?
For sigma-delta, the quantization error is the error of the ADC (let's say flash) in the loop, and the residue is after the output is fed back (going through a DAC), and subtracted from the input. Am I right?
Maybe I'm overcomplicating things but the papers are very loose in referring to them, using them interchangeably etc., so I got confused.
My final question is that, I intuitively understand when they write Dout = Vin + Q for an ADC, Dout being output, Vin the input and Q being the quantization noise, but it feels like adding 2 grapes to 3 oranges to get 4 apples, they are all different things (analog vs digital). Again, this may be a minor detail but I wanted to be sure.
Please let me know and thanks in advance.
I recently started working on noise shaping SAR ADCs, and I'm learning the ADC starting from scratch.
I'm confused about the distinction (or lack thereof) between residue and quantization error.
In the context of pipeline ADCs, for example, the residue seems to be the amplified version of the quantization error before being fed to the next stage, while in the sigma-delta ADC and SAR ADC, my understanding is that they refer to the same thing (eventually).
More specifically, for the SAR, the quantization error is the final value of the residue, because the residue during the testing of the initial bits will be large, but as the binary search converges, the digital output will also converge to the analog input, so the residue will decrease, and its final value will be the quantization error. Am I right?
For sigma-delta, the quantization error is the error of the ADC (let's say flash) in the loop, and the residue is after the output is fed back (going through a DAC), and subtracted from the input. Am I right?
Maybe I'm overcomplicating things but the papers are very loose in referring to them, using them interchangeably etc., so I got confused.
My final question is that, I intuitively understand when they write Dout = Vin + Q for an ADC, Dout being output, Vin the input and Q being the quantization noise, but it feels like adding 2 grapes to 3 oranges to get 4 apples, they are all different things (analog vs digital). Again, this may be a minor detail but I wanted to be sure.
Please let me know and thanks in advance.