Maintain input impedance when power removed

[ord]

Hi,

I have an application where a low level signal from a chemical sensor is being buffered by a high impedance op amp. I need to find a way to maintain the high impedance when the power is removed, otherwise it shorts the sensor and causes it to take a long time to recover when power is restored. As it is now, when the op amp is not powered current leaks into the inputs. I need to keep the leakage below 1nA, preferably near the 1pA level.

I was looking for a mosfet that I could use to switch one of the leads, but can't find any with low enough leakage at 0V on the gate.

I can't use a relay because I don't have enough power available to keep it energized.

Hoping for some ideas!

Thanks!

 

[barry]

you can't find enough power to drive a reed relay? You only need several milliamps. Or you might consider a latching reed relay, so you only need power for the initial switching

 

[ord]

you can't find enough power to drive a reed relay? You only need several milliamps. Or you might consider a latching reed relay, so you only need power for the initial switching

No, unfortunately it is an energy harvesting type application, and I have only about 500uA at 3V available for a relay. So far I haven't found any that can be switched at this current.

 

[lamoun]

Hello,
changing the opamp to one with low input current isn't an option?

For example this one https://www.ti.com/lit/ds/symlink/lmc6442.pdf has an input bias current of 4pA max.

 

[ord]

Hello,
changing the opamp to one with low input current isn't an option?

For example this one https://www.ti.com/lit/ds/symlink/lmc6442.pdf has an input bias current of 4pA max.

Thanks for your suggestion.

The 6442 is one of the op amps I have tried in this application. The problem is that the 4pA input bias only applies when the device is powered. When the power is off, much higher current flows through the ESD diodes to the supply pins.

 

[pstuckey]

How about something like the attached?

 

[ord]

How about something like the attached?

Yes, something like that, but with different specs. The leakage with power off is still too high on that ADG4612, but it is a step in the right direction. It still shows 30nA leakage with power off, but maybe there are others that would work. Thanks!

 

[pstuckey]

The datasheet doesn't specify if the primary leakage path is across the switch or through ground. If it's across the switch you could route the signal through multiple sections.