damage the device. The max ratings on the ADG712 allow the inputs to be maximum
of 300mV above VDD or 300mV below AGND and apply at all times. In general,
ADG712 is very robust, but it won't handle this type of voltage. Diodes will
starts turning on etc...etc.
If the input to any pin is likely to exceed the max ratings under a fault
condition, then you can protect the part by limiting the input current to less
than 5mA. For example if the max overvoltage condition is 2V, a 400W series
resistor will protect the input. If this is not practical in your design, then
a pair of Schottky diodes connected between the input and power supplies will
clamp the input voltage to within 300mV of the supplies at all time.
You can also use a channel protector such as the ADG466 to perform the clamping
There are a number of tricks you use when designing a hot plug module. One of
the simplest is to make the power and ground connectors slightly longer than
the signal connector on your module to ensure power is applied slightly before
signal voltages. You can use combinations of clamping diodes, current limiting
resistors and shunt capacitors to ground to protect downstream circuitry during
the power transient. Finally for a truly robust solution you may need to
consider adding relay switched to ensure the IO lines are isolated until power
is fully restored.
Regarding your second question, We do not test and cannot guarantee the off
resistance of any of our CMOS switches during power up power down. I do not
expect the resistance to be less than 100ohms during power up / power down but
it is not a design criteria for the part. We can provide information on the
typical behaviour but cannot guarantee this across all parts. CMOS switches are
not the same as solid state relays in this respect.
We use the ADG712BRU analog switch on a module that must be hot-pluggable. When
plugging in the module, one end of the switches will be connected to voltages
in the 1.0-1.5V range, while the other end as well as the power supply and the
control input will be at GND. Is any information available as to the behavior
of the switch in this situation? As long as we may expect more than a few
hundred ohms, we'll be happy.
When the power supply comes up a while later, we still expect the control input
to be at GND. Should we expect any transitional states with switch resistance
below a few hundred ohms?
Any information on this would be helpful, no guarantees are required!