Hello,
Im experiencing some problems with an ADC setup based on TI ADS5271, and Im hoping you can help me to solve it.
The differential inputs on the ADCs are accommodated using an analog stage which is designed to be an anti-aliasing filter with a gain of 0.5 to adjust our signals to the ADC input range of [+1,-1] V. This analog stage has been implemented using a a TI THS4151 fully differential I/O amplifier. I attach a picture of the analog stage:
The differential input to the analog stage is that on the sides of resistor R15 (105 ohms). The differential output is the one between the capacitors C4 and C6. These traces go to the ADC inputs. Our traces have been designed to match 50 ohms. R1, R2 and C3 set the gain and the first pole of the filter. The 22 ohm resistor R3 and the 320 pF capacitor C3 set the position of the second pole.
If I inject a differential signal, I am able to observe the digitized output of the ADC, which is correct. The problem comes when I don't inject any signal on the ADC input, but the wires are still present on the connector that drives the differential traces on the PCB. In that case I observe the ADC digital values to oscillate from the MSB to LSB (that is, from 4095 to 0), and some lower values. As soon as I inject a signal, I see the correct values. If the wires are removed from the connector, the ADC outputs the digital code for the analog zero (around 2047).
Am I having a noise problem? Could be an issue of the chracteristic impedanze not terminated? Maybe something different?
I will be grateful to any clue about this problem.
Thanks.
