[Guide]In this article, we will discuss two test methods for input bias current. Which method to choose depends on the magnitude of the bias current. We will introduce various error sources that need to be considered during device testing. The next article in this series will introduce a configurable test circuit that can help you complete all the measurements described in this article.
The product manual usually provides a list of bias currents for the non-inverting input and inverting input (iB+ and iB-) of the operational amplifier. The difference between these two inputs is the input offset current IOS. On the workbench, you may be tempted to use the circuit in Figure 1a to test the positive input bias current, because the amplifier in this configuration is very stable and this method is effective.
Figure 1. Use the circuit in Figure (a) to measure the input bias current at the non-inverting input of an operational amplifier. In Figure (b), a loop amplifier is added to maintain the stability of the operational amplifier when measuring against the inverting input. The circuit in Figure (c) can measure the bias current of any input. The relay can determine the circuit configuration.
Unfortunately, there is no simple way to maintain amplifier stability when measuring negative input bias currents. However, a loop amplifier can be added to maintain the stability of the device under test, so that an electrometer can be used to measure the bias current, as shown in the circuit in Figure 1b. This circuit is the two-amplifier test loop we used to test VOS in Part 1, but it has a different connection.
We reversed the inputs of the two amplifiers to maintain the stability of the device under test. Although this method works well for bench testing, the speed of the electrometer is too slow to be suitable for high-speed production testing. The method we use in production testing is a modified scheme of VOS testing. To test the input bias current (IB), we added a relay and resistor (or capacitor) to the circuit. Please see resistor RB in Figure 1c.
For the sake of discussion, we use a dual op amp test loop to describe the test. However, this technique is also suitable for the two test loops introduced in Part 1. We have added a relay and resistor for each input of the device under test in Figure 1c.
When relays K2 and K3 are closed, we can use the VOS measurement technique introduced in Part 1 to measure and save the VOUT value. Equation 1 defines VOUT based on RIN, RF, and VOS.
After transforming the formula 1, the formula 2 used to calculate the VOS can be obtained.
Next, we turn on K2, take another measurement, and get VOUT(IB-). The measured voltage is caused by the offset voltage of the device under test and the input bias current flowing through the resistor RB, which can be expressed as Equation 3.
We can now solve for iB-, and divide both sides of the equation by (RIN+RF)/RIN to get formula 4.
Then, subtract the offset voltage of the device under test from both sides of Equation 4 to obtain Equation 5.
Finally, divide RB on both sides of Equation 5 to calculate the value of IB-.
Figure 2. When measuring a bias current of less than a few hundred picoamps, capacitors should be used in the circuit and a multimeter should be used to measure a series of samples.
A similar method can be used to measure IB+. When measuring IB-, turn off K3 and turn on K2. When measuring IB+, turn off K2 and turn on K3. Since we have measured the VOS of the op amp, the next step is just mathematical calculations. The results are easy to draw, and all you need is a good digital multimeter (DMM).
Note that using a resistor to generate a voltage difference to measure IB is only valid for bias currents as low as a few hundred picoamps. We can use another measurement technique for lower bias currents.
For IB values less than a few hundred picoamps, we use a capacitor to replace the RB resistor. Once the short-circuit relay is opened, the bias current will cause the loop to combine at the speed IC = C(dV/dt) * loop gain. You can calculate the bias current by measuring in a known time interval. This method can measure the bias current less than 1pA.
PCB layout is very important for these truly low IB currents. Pay attention to reducing stray capacitance, because stray capacitance may consume some IB current. Leakage of the input pins of the device under test on the PCB can also cause errors, so a guard ring should be created around the input pins and connected to the ground. This will reduce any leakage from high voltage nodes. From a topological point of view, a temperature-stable low-leakage capacitor should be used to replace the RB resistor in Figure 1c.
The capacitive method requires a good clock. This is because the input bias current measurement requires not only the opening of the relays between various capacitors (connected to the input terminals of the device under test), but also the measurement of voltage changes at known intervals. We can calculate the input bias current by measuring the loop output voltage change in a precisely determined time period.
When the capacitor’s relay opens at the t0 position, the output starts to combine in the positive or negative direction according to the polarity of the bias current (Figure 2). The programmed delay allows the circuit to stabilize. Then, at the t1 position, the DMM samples at a known sampling rate. Next at t2, there will be another delay. Finally, at the t3 position, DMM will extract more samples.
Keep the sampling measurement time constant, so that you can know the value of dt. Obtain the average value of the second group of samples and subtract the average value of the first group of samples to get the dV value or the voltage change within dt. We can calculate the current through the capacitor, such as:
Then, the bias current is calculated by the following equation:
Typical error source
If you do not discuss the sources of error encountered during the measurement process, then the discussion of VOS measurement is incomplete. Obvious errors are those caused by the resolution of the DMM and the value (noise and tolerance) of the selected components (especially resistors). More subtle errors can be divided into the following three types:
A. Thermally generated electromotive force (emf), caused by relay contact
● Welding point
● Pin connection between boards
● Automatically test the contact points and sockets of the processor
B. Leakage current caused by the following factors:
● Relay control and power trace
● Properties of PCB material
● The device under test itself
The typical error sources in all the DUT configurations discussed here are thermally generated electromotive force and leakage current. Leakage current mainly affects the bias current measurement, while the thermally generated electromotive force can affect all low-level offset voltage measurements. Minimizing these effects is a necessary condition to ensure system function and measurement accuracy.
Leakage current is caused by surface contamination and resistive paths through components or in PCB materials. Surface contamination can usually be controlled by thoroughly cleaning the circuit board, but humidity may change the surface leakage current. Other resistive paths can be set by the isolation resistance of the material. Leakage current may also occur when power lines or relay-controlled power lines are connected to resistive paths. The use of guard rings and latching relays that support high-level effective drivers can alleviate some of the effects of such leakage paths.
The thermoelectromotive force can be generated in relay contacts, solder joints, inter-board pin connections, and all other test processor contacts and sockets. For example, consider the dual amplifier VOS measurement circuit in Figure 1c. Leakage current will not significantly affect the measurement. But this circuit cannot reflect multiple sources of thermoelectromotive force.
Figure 3 is the error source of the thermoelectromotive force, marked as VT. When measuring at room temperature, the gradient gradient is normal. However, when testing in a cold or hot environment, the thermal gradient from the device under test to the resistor and relay will be obvious.
Figure 3. Thermal EMF error (shown as VT) can affect measurement results.
Dostal, Jiri, “Thermoelectric Voltage”, excerpted from section 9.3.1 of the op amp 2nd edition. Butterworth-Heinermann, 1993, p. 266.
David R. Baum is an analog IC design engineer at Texas Instruments (TI), responsible for the development of product designs for LCD and AMOLED TVs. David has more than 27 years of experience in analog design and at least 7 patents. He graduated from the University of Arizona in Tucson, Arizona with a bachelor’s degree in electrical engineering, an MBA, and a master’s degree in German literature. Email address: [email protected]
Daryl Hiser is a senior test engineer in TI’s high-precision operational amplifier product department. He is responsible for formulating and executing new product testing and characterization schemes. He holds two patents. He graduated from Northern Arizona University in Flagstaff, Arizona with a Bachelor of Science in Zoology. Email address: [email protected]