Forums

Caleb,

A differential analog MUX can be used to switch the inputs of an instrumentation amplifier as long as the charge injection that will occur when switching channels doesn't not cause too much of an issue with your design. I should point out that unless you need the 3 fA input bias currents of the INA116, several other FET input instrumentation amplifiers may be easier to implement in your system (INA111, INA121, etc.). The INA116 needs careful consideration of layout to ensure the leakage paths to the input pins are properly handled. Here is a simplified example schematic:

R2 and R3 will be necessary to provide a bias current pathway for the instrumentation amplifiers inputs. Their values should be MUCH larger than the on resistance of the mux so that mismatch between the two inputs does not cause a large degradation in common mode rejection. I also recommend that you look for muxs with very low channel-to-channel leakage to avoid the crosstalk effects you previously saw.

John,

Because normal devices have ESD protection which provides a reverse biased diode, and this results in a gate bias current of serveral nA due to reverse diode leakage current. And INA 116 has properietary ESD protection that results in a very high input impedance and very low input bias current, so to minimise the measurement on the open circuit potential in the solution.

So, I still need to add in the R2 and R3 for open circuit potential measurements? would this cause any problem? Hope you can let me know if I need a multiplexer how should I be doing it, a multiplexers which I can share the INA116 with more than one solution samples. A setup of one solution is as below:

And now I'm only building it on breadboard, does this affect it much?

Here is a PCB layout:

Does this looks fine?
Should those jumpers be using BNC connectors instead?

And after the INA116, I would need an ADC before passing it to a microchip, I'm thinking of using MAX11040K, as the resolution is good and think will be easy to impement. Hope you can give some suggestions.

Thanks,

Caleb

Caleb,

Although the INA116 features extremely low input bias currents, they are still present and without a suitable input bias current pathway, the output of the INA116 will drift to one of the supply rails. When the INA116 is connected to the solution you are measuring, bias current will pass through the solution which may be acceptable, however, if you are using a multiplexor on the front end of your measurement, there is the potential for this bias current pathway to be interrupted as the MUX switches, which will cause the output of the INA116 to change in between measurements. A more suitable solution might be to multiplex the outputs of multiple INA116s to the ADC (place the multiplexor after the INA116s, not before).

As for the construction of this circuit, achieving extremely low leakage current requires extremely care to be taken. It is unlikely that constructing the circuit on a breadboard will achieve the performance that the INA116 is capable of because of the presence of leakage pathways and surface contamination. The layout of the pcb looks suitable but I recommend that you thoroughly clean the board after assembly to eliminate the effects of residues left on the surface.