Forum Post: RE: TMS320F28335: ADC reading saturation, all reading values will gradually go to 2048

Hi Tommy, There is no other load between OPAMP and DSP ADC input. I haven't change the MC3403 OPAMP to check whether the bandwidth of OPAMP matters. I took an experiment on the C169 capacitor's voltage. There is some interesting phenomenon. When the ACQ_PS is not set to zero. There will be a spike every DSP ADC sample time. Picture 1 shows Voltage on C169. To check this phenomenon, I use different ACQ_PS to check the sample and hold time. The overall disturbing time is 8.8us and 15.6us for the two cases. Which are 220 and 390 ADC time(25MHz). I think it is a reasonable phenomenon for ADC sample and hold to charge the inner capacitance. However, I don't know whether the voltage differences 45.4mV and 79.25mV is reasonable? To check this problem, first I check the disturbance from the beginning of the DSP operation. I found that, in the beginning, there was no disturbance on the C169 waveform. The disturbance gradually goes up during the operation. Secondly, I took place C169 capacitance from 103 (10nF) to 105 (1uF). The waveform is shown in the followed picture in probe 1. The probe 1 is in AC mode, so I ignore the DC value. It seems like there is no recognizable disturbance in the waveform. However, the ADC misreading problem is still existing. The conclusions of today's experiment are shown below. 1. There is a disturbance on the C169 each ADC sample and hold time when the ACQ_PS is not equal to zero. 2. When the capacitance of C169 is changed to a large capacitance, the disturbance disappears. However, the voltage on C169 dropped to a wrong voltage below the expected voltage.