an Very High Temperature stability Pin Programmable Instrumentation amplifier
with gains x1 (input signals /-12.5V min), x10 and x100. It is very important,
when overrange occured in ranges x10 and x100 there not to be any leakage
(current)from inputs. Is it possible to use AD621 and AD620 with inputs in
parrallel? What influence between AD621 and AD620 is expected?
Q2.We used Burr-Brown's PGA 204 before. It has handy logical inputs, HOWEVER,
when overrange occurs, there is significant (for our application) input current
caused by input protection circuits, we suppose. It happens when inputs are
still within supply rails...ADI's pin-programmable instrumentation amplifiers
rely on an external switches or relays to switch the gains? We can't use
relays, because of the space limitations in our design and their switching
time. Semiconductor switches, however, are too nonlinear and temperature
dependent... Do you have a suggestion for gain switching of ADI's
pin-programable instrumentation amplifiers?
Q1.The lowest drift pin programmable InAmp available from AD is the AD624 C
grade (2001). The input bias current will not increase significantly at gains
of x10 and x100 provided the inputs are kept within the supply rails. However,
the settling time will increase when the inputs exceed the nominal input
levels, that is there will be some recovery time associated with the amplifier
coming out of saturation. You should characterize this recovery time in your
own circuit if you believe it will be an issue in your design.
There should be no problem, in theory with using the inputs of two InAmps in
parallel - again providing the inputs never exceed the supply rails and the
absolute maximum ratings are respected. Also remember to a path for the input
Q2.It's certainly possible that the over-voltage protection circuitry on the
PGA 204 is causing the unwanted increase in input bias current. There is always
a balance to struck between providing input protection and minimizing the
additional leakage current.
The AD624 does has input ESD protection diodes but it does not have
over-voltage protection like the PGA 204 to protect to +/-40V. As I stated
previously I would not expect the bias current to increase significantly when
the input exceeds the nominal input range provided the input remains within the
power supply range. The main effect will be saturated output of the internal
amplifiers and there will be a recovery time associated with outputs coming out
of saturation. However, do note that we do not specifically test the input bias
current under these conditions therefore you should perform some additional
testing for yourself to confirm that performance will meet your requirements.
Regarding switching the gains, you are correct that the linearity and
temperature stability of the switching element is critical in ensuring accurate
operation and semiconductor switches have not been good at either of these.
However we now have a range of switches which have inherently low on-resistance
as low as 7ohm with +/-1ohm flatness over input range.
For switching the gain I recommend our Low Ron switches with dual supplies
ADG621/2/3. The dual supplies should help keep the Ron in the flat linear
region. Note that the ADG621 requires a +/-5V supply.
We don't have any stand alone digitally controlled precision PGA, but we do
include this function in many of our sigma delta converters. The AD7705 for
example allows 16 bit accuracy from an input from +/-70mV to +/-2.5V and the
AD7714 achieves up to 16 bit accuracy from an input of +/-10mV and 18 bits with
a 2.5V input. . Of course this is only of interest if you ultimately plan to
digitize the input signal.
For my own information, could you tell something about the application you are
working on, expected volume and information on the products that XPECT build.