Signal Characteristics: The IF (I/Q) signal consists of a DC component and an AC component.

DC Component: 2.52V; AC Component: 1.72 – 2.04 V (V P-P); BW: 20Hz - 500kHz; Power: 20dB

Requirement: Now I want to digitize this signal for further processing. I prefer a sampling at 1 MSPS with ADC and then porting digital data to PC using a FPGA over SPI interface.

Problem: The problem is, the signal with DC and AC component ranges from 3.5V to -1.5V with 2.5V from DC and almost 1V from positive cycle or negative cycle of the AC signal. Now I can use a coupled capacitor to block the Dcand use the AC for processing. My choice of ADC reference voltage is almost 2V-2.5V (for good resolution usability of ADC) so the signal can be amplified. But this signal have a negative part which I think is not good for a linear behavious for the ADC, as it has no negative reference voltage level. ADC reference is from 1V-5V. Using a rectifier is not a good choice from my view.

Question: How can I digitize this signal? Also comment on my views if not correct as I have only theoretical knowledge and never worked on a practical project.

  • $\begingroup$ It would be useful if you specified the ADC model you're planning to use. In any case, you're right that you want to make the input signal match the converter's input range. That is always possible, as you say, by using couplers, amplifiers, voltage dividers, etc. $\endgroup$ – MBaz Nov 11 '14 at 12:36
  • $\begingroup$ Your numbers don't seem to add up. Your DC component is ~2.5 V and the AC component is only ~2 V peak to peak. Why do you expect negative voltage swings? $\endgroup$ – Jason R Nov 11 '14 at 14:14
  • 1
    $\begingroup$ yeah, i am wondering the same as @JasonR. if the peak-to-peak is, at most, 2.04 V and there is DC bias of +2.52 V, i cannot see how the voltage ever gets more negative than +0.48 V. that said, if there are voltage swings into negative territory and that is bad for your unipolar-power-supply ADC, then the answer to that is biasing the signal (with a resistor) sufficiently (and perhaps, a known and precision value, if knowledge of the instantaneous DC value is critical) so that it is never negative going into the A/D converter. you can block the DC in the software, if you want. $\endgroup$ – robert bristow-johnson Nov 11 '14 at 15:50
  • $\begingroup$ @JasonR and robert bristow-johnson I thought of using a coupled capacitor to block the DC and use the RF, as I'm only interested insampling the RF, so when I remove the DC, the output is from a negative to a positive cycle. But today, I tried the same and found the coupled capacitor causes a phase shift which is absolutely undesireable. So Now I think to keep the DC offset and change the signal from 0.5V -> 4.5V cycle.... $\endgroup$ – dDebug Nov 12 '14 at 13:46

Browse other questions tagged or ask your own question.