Hi, When I use AD7705 with 2.4576M Crystal, the convertion result is very stable and repeatable after the device was reset.
Since 49U 2.4576M is hard to order, 49S 4.9152M is the best choice. AD7705 can divide it to 2.4576M by setting the clock register, from the datasheet.
When I use 4.9152M Crystal, the convertion result is also very stable, but not repeatable after reset. When the input voltage remains the same, after reset by hold reset pin low for 1ms (after reset, the selected channal was self-calibrated, PGA=4, buffered and bipolar), the convertion result has about 0.1% error from the result before reset. Reset again, the error may dissapear, or remain. The same thing also occurs after power on.
I have change the capacitor near crystal from 22pF to 39pF, the error remains. When I replace 4.9152M to 2.4576M, the error dissapears.
I want to know if the PCB has some problem to make 4.9152M not working well, or AD7705 with 4.9152M cannot work well at all.