I have an AD7609 powered by a single 5V supply (AVcc linked to Vdrive with a 1000R @ 100MHz, 200mA EMC Ferrite Filter), using Serial Data output. This is using an external 2.5V reference from an ADR421BRZ.
Initially this was working properly but now we have observed a strange phenomenon where the ADC outputs incorrect readings from all 8 Channels when powered by 5V. Random values for (positive and negative) were coming out on all 8 channels. Applying a voltage across V1+/V1- seemed to make a change (even though the values were still incorrect) but voltages across any other channels did not even register.
Reducing the voltage to 4.75V seems to fix the problem. AVcc of 4.75V is the lower limit of this chip's supply voltage so we are worried about stability issues.
Is there any reason why is was working quite happily at 5V and now there seems to be some internal biasing that is pushing that AD7609 to work only at 4.75V?