
Hi Scardig,
It looks like your signal is actually clipping - the fact that it's "squishy" clipping (rather than being cleanly shopped off) indicates that it's the amplifier (although similar effects can happen with ADC inputs, especially if they include onboard buffer amplifiers.)
The EVAL-AD7768 user guide and schematic is a little hard to follow, but it looks like the inputs are by default routed through the ADA4896-2 amplifiers, with a default negative rail of ground. This means you must observe the amplifier's input and output headroom requirements - refer to its datasheet.
But it looks like an easy solution is to insert a large (1 uF or 10 uF) in series with your input signal. The inputs are biased to VCOM (which is set by the AD7768) through 10 k resistors. (I know this works, I've tried it before, I am pretty sure it was an unmodified EVAL-AD7768.)
I'm moving this to the Precision ADCs forum, Reference Designs is mainly for Circuits from the Lab, the EVAL-AD7768 is a standard eval board.
-Mark
Hi Scardig,
It looks like your signal is actually clipping - the fact that it's "squishy" clipping (rather than being cleanly shopped off) indicates that it's the amplifier (although similar effects can happen with ADC inputs, especially if they include onboard buffer amplifiers.)
The EVAL-AD7768 user guide and schematic is a little hard to follow, but it looks like the inputs are by default routed through the ADA4896-2 amplifiers, with a default negative rail of ground. This means you must observe the amplifier's input and output headroom requirements - refer to its datasheet.
But it looks like an easy solution is to insert a large (1 uF or 10 uF) in series with your input signal. The inputs are biased to VCOM (which is set by the AD7768) through 10 k resistors. (I know this works, I've tried it before, I am pretty sure it was an unmodified EVAL-AD7768.)
I'm moving this to the Precision ADCs forum, Reference Designs is mainly for Circuits from the Lab, the EVAL-AD7768 is a standard eval board.
-Mark