The data sheet for the AD9288 is not real clear when it comes to defining the analog input compliance. No formulas are given, and there is contradictory information.
I am driving the part with a signal whose magnitude ranges from 0-1V, on a DC bias of 1V, from an AD8138 ADC drivier . That signal is DC coupled into the 9288 Analog inputs as follows:
Ain+ receives a signal varying between 1V and 1.5V
Ain- receives a signal varying between 1V and 0.5V
Page 3 of the Rev C data sheet states: that Input Voltage Range of Ain+ with respect to Ain- is +/- 512mV. In Table 3 (p.8), it mentions that the Ain- is complementary (to Ain+, presumably).
A complementary signal set would have the folowing characteristics (insofar as is permissible by a text editor):
Ain+ __ ___ __
Where Ain+ and Ain- both range from ~0.75V to ~1.25V but are complementary). Thus, (Ain+ - Ain-) ranges from -512mV to +512mV.
In my circuit, the drive is 'opposing', not complementary:
Ain+ __/ \__
Ain- \ /
Ain- ranges from 0.5V to 1.0V, and Ain+ ranges from 1V to 1.5V. (Ain+ - Ain-) ranges from 0 to 1V.
In each case, each member carries half of the signal range.
In each case, the DC level they vary about is 1V.
Another way to look at it is that the differences between the two approaches has to do with bias. In the former case, the DC bias of each input is identical. In the latter case, each signal has a different DC bias. It's sort of akin to numerical representation of an integer vs. an unsigned value of the same size. The magnitude of the range is the same, but the zero-point or bias is different.
The "Analog Input" section Page 14 states that "...The nominal input range is 1.024 Vpp centered at Vd * 0.3..." My driving circuit is completely compliant with that statement. I interpret is as meaning the nominal range of (Ain+ - Ain-) is 1.024Vpp.
The difference between the two presentations is essentially DC bias of the input signal. My input signal has a positive DC bias (it is entirely positive). If I remove the DC bias, allowing the input to vary about zero, the
AD8138 outputs presented to the ADC will assume a range per that stated on p. 3 (Ain+ with respect to Ain- is +/- 512mV) The data sheets claims broad compatibility with both DC and AC coupled inputs. However in order to comply
with the requirement that the input range of Ain+ - Ain- being +/-512mV, I must AC-couple.
In fact, a "Nominal Input Range" of 1.024 Vpp centered at VD * 0.3 sounds like (Ain+ - Ain-) = 1.024Vpp, centered about 1V. THat is what I have.
Which is correct? Can the inputs tolerate more common-mode bias that what is implied by the initial range remarks on p. 3? Can both the schemes shown above work? If not what is the easiest way to modify my DC-coupled circuit to generate an (Ain+ - Ain-) range of +/- 512mV.