Hi Iolanda,
On the failure: The AFE0064 will not get damaged as long as the input voltage at the pins are within the Abs max ratings mentioned in the datasheet. Also, try to limit the current onto any IN pin within 10mA, as sustained values of current more than that could damage the part.
On the linearity of the device: The AFE0064 is meant to integrate charge packets, (example current pulses or charge from photo-diodes from a flat panel detector). The DDC series (eg DDC264) is more easily used for systems with continuous currents like the case you have mentioned. Nevertheless, you could use the AFE0064 part with the following information (these are not obvious in the datasheet because the part was not meant to take a continuous current):
(1) The reset sample should be within 10% of the FS: The integrator enters feedback (ie exits reset) at 9th CLK rise (for simult mode) and hence immediately starts integrating any input current causing a non zero reset sample. Hence the input current should be small enough so that this charge integrated from 9th CLK rise to SHR rise edge must be < 10% of the range. For eg. with CLK = 1MHz, the input current is integrated for (64-8)*CLK+t3 = approx 56 CLKs = 56us before the reset sample is taken. Hence the input current for good linearity in range7 must be limited to < 9.6pC*0.1/56e-6 = 17nA. Linearity may be compromised if it exceeds 10% and in certain cases, the output may behave non-monotonically wrt input.
(2) The device output corresponds to signal sample - reset sample: Since the part has a non-zero reset sample due to the constant input current between 9th CLK and SHR rise, the output will be lesser to that extent.
Please recheck the linearity with the above in mind.
Regards,
Nagesh