Removing Offset Errors from Passive, Power-On, In-Circuit Resistance Measurements Tutorial

Automated circuit testing is used in industries ranging from semiconductors to communications (cellular phones) to automotive (brake systems). Most circuit testing is done with active (power-on) measurements, such as voltage and current. The only measurement that is passive (power-off) is resistance. This has traditionally been the case because all accurate in-circuit resistance measurements required the power to be off. However, there are a number of benefits to being able to make all measurements with the power on including:

• Simplified test sequences and test system architecture
• Increased automated test throughput and system utilization, resulting in cost reductions
• Increased accuracy, resulting in more correct pass/fail final system verifications

This document will describe this traditional method and it will describe how to use the NI 4070 FlexDMM Offset Compensated Ohms (OCO) feature to make more accurate low-level, passive, power-on, in-circuit resistance measurements. You can also view an interactive presentation that takes you through this tutorial's material at your own pace.

For additional information and more interactive tutorials on the NI 4070, visit the Digital Multimeter Measurement Tutorial.

Low-Level Measurement Setup

The best setup for low resistance measurements is the 4-wire setup. A 4-wire setup utilizes four leads, two leads to drive current through a resistance and two leads to measure the voltage drop across that resistance. Traditional 4-wire measurements require that the power is off in the circuit so that the measured voltage drop is only due to the supplied current. Thus traditionally, the power to a circuit must be turned off before any measurements are taken. Figure 1 below illustrates a 4-wire setup. The black-colored leads measure the voltage drop. The blue-colored leads connect the test current, which is 1 mA in the 100 ohm range.


In every resistance measurement, the digital multimeter must supply a test current, 1 mA in this setup, and then measure the voltage drop across the unit under test. The measured voltage drop is then used in the calculation for resistance, thus its accuracy is very important. This accuracy can be negatively affected in passive, power-on, in-circuit resistance measurements, such as power line resistance. This tutorial discusses possible sources of error in four measurement methods.

The four comparison measurement methods are:

1. No Compensation
2. Approximated Offset
3. Measured Offset
4. Offset Compensated Ohms (OCO) on the NI 4070 FlexDMM

View the interactive presentation

Power Line In-Circuit Resistance

In-circuit measurements can have offset voltages since they often involve measuring resistance in the presence of a large voltage, for example, power supply bus resistance with the power on as shown in Figure 2. With a traditional digital multimeter, these types of measurements must be done with the power off or very large errors will be introduced into the system. For this example, the power line resistance (R_{power line)} is assumed to be 10 milliohms.


View the interactive presentation

Case 1: No Compensation

The measured voltage is equal to the line current times the power line resistance plus the test current times the power line resistance. If there is 100 mA of current in the supply bus, the measured voltage would be 100 mA times the power line resistance plus 1 mA test current times the power line resistance. With a power line resistance of 10 milliohms, the measured voltage would be 1.01 mV.

V_meas = Line Current x R_{power line} + Test Current x R_{power line}
V_meas = 100 mA x R_{power line} + 1 mA x R_{power line}
V_meas = 1.01 mV

The calculated or measured resistance then, is equal to the measured voltage divided by the test current. This measured resistance is 100 times greater than the actual value of 10 milliohms.


R_meas = V_meas / Test Current
R_meas = 1.01 mV / 1 mA
R_meas = 1.01 ohms

This is a measurement that cannot be done without some kind of offset compensation.

View the interactive presentation

Case 2: Approximated Offset

To get an accurate resistance measurement, the effect of the 100 mA power supply must be removed from the measurement. One way to do this is to approximate the error by estimating the current. There are two drawbacks to this type of adjustment. First, an external current excitation source must be used since voltage, not resistance, must be measured. The digital multimeter setup is shown in Figure 4 below.


Second, additional calculations must be done by hand to obtain resistance. The measured voltage would still include the error due to the power supply, but when calculating the resistance, an estimated voltage could be removed. The major disadvantage to this method is that the actual error will not be known.

V_meas = Line Current x R_{power line} + Test Current x R_{power line}
R_calc = (V_meas - Estimated Power Line Error) / Test Current

View the interactive presentation

Case 3: Measured Offset

Another method overcomes this disadvantage, by measuring the actual power line error voltage and then taking it out of the resistance calculation. With a traditional digital multimeter the drawbacks include:

• Multiple voltage measurements
• An external current excitation source
• Additional calculations that must be done by hand
  

The same in-circuit power-line resistance application using the measured offset technique would require the following steps.

1. Take a voltage measurement with an external current supply (as shown in Figure 4)
2. Take a second voltage measurement with no current excitation as shown in Figure 5 below. This measurement will effectively measure the power line error in the system.
3. Use the difference of the voltages to calculate resistance

The calculations for this method are as follows. The voltage drop across the power line is again 100 mA times the power line resistance plus 1 mA test current times the power line resistance The measured voltage of 1.01 mV is the same as in the previous traditional non-compensated measurement.

V_{meas #1} = Line Current x R_{power line} + Test Current x R_{power line}
V_{meas #1} = 100 mA x R_{power line} + 1 mA x R_{power line}
V_{meas #1} = 1.01 mV

With the second measurement using the measured offset technique, the 1 mA test current is not used, and so the measured voltage changes. The voltage drop across the power line is only due to the 100 mA power supply and so only that value is measured.

V_{meas #2} = Line Current x R_{power line}
V_{meas #2} = 100 mA x R_{power line}
V_{meas #2} = 1.00 mV

The difference in the two measured voltages represents the voltage drop across the power line due to the test current. This voltage is divided by the test current to calculate an accurate resistance for the power line. Thus any error due to the power supply current is removed.

V_calc = V_{meas #1} - V_{meas #2}
V_calc = 1.01 mV - 1.00 mV
V_calc = 0.01 mV

R_meas = V_calc / Test Current
R_meas = 0.01 mV / 1 mA
R_meas = 0.01 ohms

View the interactive presentation

Case 4: Offset Compensated Ohms (OCO)

A final method incorporates the accuracy of the measured offset method but maintains the simplicity of the no compensation technique. This method, called Offset Compensated Ohms (OCO) is available on the NI 4070 FlexDMM. It overcomes the requirements of multiple voltage measurements, external current excitation sources, and additional calculations, yet is still able to measure the actual error and take it out of the resistance measurement.

Once enabled and without any additional user interaction, Offset Compensated Ohms invokes the NI 4070 FlexDMM to take two resistance measurements, but with the test current source off during the second measurement. The first measurement includes the voltage drop across the power line including that from the supply current. During the second measurement the current source is turned off so the voltage drop across the power line is only due to the power supply as shown in Figure 6. The NI FlexDMM then subtracts the second measurement from the first to determine the Offset Compensated voltage (V_OCO). This voltage is then used to determine the correct resistance value.

V_OCO = V_{meas #1} - V_{meas #2}
V_OCO = 1.01 mV - 1.00 mV
V_OCO = 0.01 mV

R_meas = V_OCO / Test Current
R_meas = 0.01 mV / 1 mA
R_meas = 0.01 ohms

View the interactive presentation

Summary

The chart below is a comparison of the methods to improve the accuracy of low-level, in-circuit resistance measurements including the Offset Compensated Ohms technique on the NI 4070 FlexDMM.

	Case 1: No Compensation	Case 2: Approximated Offset	Case 3: Measured Offset	Case 4: Offset Compensated Ohms (OCO)
Number of Measurements	1	1	2	1
External Current Supply	No	Yes	Yes	No
Accuracy	OK	Better	Best	Best
Possible Error	100 X	less than 100 X	Negligible	Negligible
Sources of Error	No compensation	Bad estimate	No procedural error	No procedural error
Complexity	Easy	Moderate	Difficult	Easy (no setup changes)

View the interactive presentation

Reader Comments | Submit a comment »