OPA189: How to measure input bias current

Toshi,

The input bias current, IB, in CMOS op amp cannot be measured using resistor because of its tiny magnitude but instead must be integrated as shown below and calculated using Ic=C*dV/dt equation.  In the simulation below, this would result in Ic=1nF*slope=1e-9*35n/500n=70pA

I have attached below Tina-TI schematic so may may do your own simulation.

OPA189 IB measurement.TSC

Hello Toshi,

To add to Marek's comments, the LMC660 datasheet (and several of the LMC60xx family datasheets)  has a blurb in the Applications Section about measuring the fA input bias currents using integration (page 11).

The preferred way is to use the inverting integrating configuration (cap in the feedback path, S2 and C2 in the datasheet figure).

The inverting integrator has a fixed common mode point, whereas the non-inverting (cap on the non-inverting input) will change the common mode as the capacitor charges up.

This is not a big problem at mid-common mode, but as you approach within a volt of the rails, the bias current changes as the ESD diodes start to contribute. So you end up with an inflection in what should be a linear integration.

The best way is to use a separate integrator to measure the current. This also allows "empty socket" measurements to subtract out fixture leakages.

Waaaay back in 1993, Bob Pease wrote an EDN article on measuring fA input bias currents, "Whats all this Femtoamp stuff, Anyways", and the follow-on Teflon article. Unfortunately, the offical EDN site is messed-up and does not have the images.

Below is an archive of the National Semi Pease site, which contains several of the "What's all this..." articles.

 https://www.ti.com/ww/en/bobpease/assets/www-national-com_rap.pdf

The Femtoamp article is on page 45, and the Teflon article is on page 41. What was true back then is still true now - the principals still apply - just the packages have changed.

And, yes, his Technician "Paul" was me...