I see quite large input impedance variation specified in the datasheet. Now I
wonder, is the input impedance varying with the gain? Or is it just varying per
device and constant over gain? This is important for my design because I need
to keep the group delay of my circuit constant over gain. I am using the part
in an ultrasound flow meter.
The input impedance of the AD605 is that of the R/1.5R ladder as seen in figure
6 of the data sheet. The 40Ω variation (about 25%), represents process
tolerance for passive components when the part was released. The 3pF is usually
quite constant, being dependent on geometry of the layout mostly. These values
don’t change with gain, they are just the best available passive component
tolerances for that process at that time.
The AD605 gain-control architecture is the X-amp, where the attenuation is all
due to passive network components ahead of a fixed gain amplifier. We’ve
measured group delay on the AD604, which uses the same VGA section as the AD605
and attached a copy of the results for you. You will see the GD is quite
uniform across the whole sample lot. We’ve also found the same results with
other measurements, although all our VGA products have not been tested. I would
expect they would all be uniform in terms of phase shift and group delay vs
frequency, with the possible exception of the AD8330, and any other device that
uses a translinear gain control approach.
Please also see attached material for our new AD8338, 3mA, 20MHz, 80dB gain
range VGA. This is a new low power, high performance VGA that we think will be
well suited to a wide range of ultrasonic flow metering applications. We have
samples now and are targeting release for early 2013 (within the next 6