What Is a Source Measure Unit (SMU)?

Publish Date: May 25, 2016 | 19 Ratings | 4.05 out of 5


Virtually all research, design, development, and production applications require an instrument that can source power to a device that is being developed or tested. Many of these applications also require the ability to monitor the voltage and current being consumed by the device to characterize device behavior or to test for proper operation. Often, you can meet both of these requirements with a single programmable power supply that sources either a constant voltage or constant current as well as reads back the associated current or voltage. In these applications, milliamp sensitivity in current measurement often suffices.

Other applications require sourcing and measuring with more precision than you can find on a typical programmable power supply. For instance, consider ubiquitous electronic devices for which every microamp of current drawn reduces battery life. Manufacturers often need to characterize these devices during production for power draw in a variety of states. For these situations, a high-precision power supply offering microamp-level sensitivity works well.

Certain applications are especially demanding and require even higher precision and specific features. Semiconductor validation and characterization is an example of an application that requires current sensitivity into the nanoamp range and below. In addition, the demand for more precision, higher speed, remote sensing of voltage, and four-quadrant outputs can render a traditional programmable power supply insufficient. For these situations, where precision low-level sourcing and measuring is needed, a source measure unit (SMU) is the best choice.

An SMU is a precision power sourcing instrument that provides voltage sourcing and measurement resolution at or below 1 mV as well as current sourcing and measurement resolution below 1 µA. In addition, SMUs offer remote-sense capability and a four-quadrant output that incorporates both bipolar voltages and the ability to sink power. Finally, SMUs are optimized for sweeping both current and voltage to determine the IV characteristics of a device. As a result, SMUs have been widely adopted, especially in the semiconductor industry and are a common component in many automated test systems. For more information on the specific SMU features and their associated applications, read the corresponding sections below.

Table of Contents

  1. Precision
  2. Remote Sensing of Voltage
  3. Four-Quadrant Operation (Source and Sink)
  4. Sweeping
  5. Transient Response
  6. Multifunction SMUs
  7. Conclusion
  8. Next Steps

1. Precision

The most notable characteristic that differentiates an SMU from a standard power supply is its precision. Precision is defined as repeatability or reproducibility. When considering the precision of instrumentation, keep in mind two related key features: sensitivity and accuracy.



Sensitivity is defined as the smallest detectable change that can be measured (or sourced) by an instrument. In other words, sensitivity is the smallest increment that can be set on the output of a device or detected on the input of a device. SMUs achieve greater sensitivity than standard power supplies by offering multiple ranges on which they can set and read voltage and current.



Accuracy is the maximum uncertainty of a given source or measurement. Absolute accuracy is referenced to a “true” reading represented by a standard. SMUs typically have accuracies for both sourcing and measuring that are at or below 0.1 percent of the output to which they are set.


Together, the sensitivity and accuracy of an SMU will ultimately define its performance in a given application. While some applications can be mainly focused on detecting small changes, others focus on tight certainty of a sourced value or measured response. In general, SMUs are employed when the accuracy of sourced and measured values is important, and the application requires sensitivity beyond what can be found in a typical programmable power supply.


Back to Top

2. Remote Sensing of Voltage

A challenge in accurately sourcing or measuring precise voltages is the effect that lead resistance has on the voltage that a device under test (DUT) sees. Lead resistance is always present but is most prominent when smaller wires with longer distances are involved. Although typically no larger than a few ohms, these small resistances can have a large effect on the voltage a DUT receives, especially when the internal resistance of the DUT is small.


Figure 1 shows a diagram of a generic circuit that consists of a power sourcing instrument, lead wires, and a DUT. In this case, the lead resistance is assumed to be 1 Ω for both the positive and negative lead wires connecting the power source to the DUT.


Figure 1. Example Connection Diagram for a Typical Programmable Power Supply

Assuming that the power source is set to an output of 5 V and the DUT has an impedance of 1 kΩ, you can calculate the actual voltage seen at the terminals of the DUT using the following equation:


For the initial case, the voltage seen is actually just 4.99 V. For some devices, this small change is not an issue; however, for applications that require precise characterization based on operating voltage, this error can become very important. Furthermore, for devices that have lower input impedances, you can substantially reduce the resulting change in the voltage seen at the leads. Table 1 lists the values that that the example DUT sees based on lower values of its input impedance.

DUT Impedance DUT Voltage
1 kΩ 4.99 V
100 Ω 4.9 V
10 Ω 4.16 V


Table 1. Voltages Seen by DUT Based on Input Impedance


The solution to lead-resistance-induced voltage error is remote sensing, also known as 4-wire sensing. This technique accounts for the voltage drop across the lead resistance by measuring the voltage directly at the DUT and compensating accordingly. This method is similar to the way that digital multimeters (DMMs) perform 4-wire resistance measurements to remove the effect of lead resistance from resistance measurements. Both power sources and DMMs feature two extra terminals on the output to allow for this 4-wire remote sensing technique, and these extra terminals are connected directly at the DUT. Although there is still lead resistance in the wires used for remote sensing, voltage measurements are high-impedance so no current flows through the sense wires and no voltage drop is seen.


SMUs typically include remote-sensing capability to make full use of the added voltage sensitivity they offer.


Back to Top

3. Four-Quadrant Operation (Source and Sink)

Another defining SMU characteristic is the flexibility of their outputs. SMUs have four-quadrant outputs, which can provide positive voltage and positive current (quadrant 1), negative voltage and positive current (quadrant 2), negative voltage and negative current (quadrant 3), or positive voltage and negative current (quadrant 4). Typically, an SMU data sheet has a quadrant diagram similar to the one in Figure 2, showing the maximum voltage and current that you can apply in each of the four quadrants.

Figure 2. Quadrant Diagram for an SMU



For a power supply or SMU to be classified as four-quadrant, it must be able to source both positive and negative voltage from the same terminals. This is important for characterizing the breakdown in active devices that have both forward and reverse characteristics important to their operations. You can characterize these forward and reverse characteristics using an output channel that can accommodate sweeping voltage from negative to positive values.

Figure 3. IV Curve of a Typical Zener Diode Showing Both Forward and Reverse Breakdown Characteristics



Similarly, for a power supply or SMU to be classified as four-quadrant, it also must be able to source power as well as sink power. Sourcing power refers to providing the stimulus for a circuit, and sinking power refers to dissipating power that is being applied by an external active component such as a battery, a charged capacitor, or another power source. Four-quadrant operation is an essential capability for applications that require both sourcing and sinking, such as charge cycle testing on rechargeable batteries, testing the output short-circuit currents on the pins of a digital semiconductor device, or sinking power from a power-management IC.


Back to Top

4. Sweeping

A primary application for SMUs is the characterization and classification of various electrical components, semiconductor devices, and custom chip designs. A typical way to perform this characterization is by sweeping either the voltage or current being sourced to the DUT through a list of values. A classic example of this characterization method is tracing IV curves for diodes and transistors. In both of those cases, voltage is swept across the terminals of the DUT and the resulting current is measured.


Figure 4. IV Characterization of a Transistor With NI LabVIEW and Modular SMUs


Back to Top

5. Transient Response

Another difference between an SMU and a power supply is the speed of the transient response, or how quickly and cleanly the instrument reaches its desired output. Ideally, the output control loop is tuned to provide a very fast rise time with no overshoots or oscillations; however, it’s impossible to optimize the control loop of the instrument without knowledge of the DUT’s impedance.


Traditional power supplies are designed to source very stable power to the DUT with minimal emphasis on the rise time of the instrument. By prioritizing stability over speed, power supplies generally provide very reliable power to a variety of DUTs, even highly reactive or inductive loads. Specialized, fast transient power supplies are optimized for powering mobile devices and RFICs that draw a large, instantaneous current and require the instrument to maintain its output voltage. Because these instruments are targeted at a specific group of DUTs, vendors have enough knowledge of the DUT’s impedance to tune the control loop to provide a fast transient response.


Because SMUs are used in a variety of applications, often for extensive IV sweeps and high-volume production test applications, they cannot afford to throttle the transient response speed to accommodate all types of loads. Traditional SMUs use analog control loops to provide feedback for the output, while the latest NI SMUs implement this control loop digitally. SMUs with analog control loops strike a balance between speed and stability, and often provide two or more output modes you can use based on the DUTs impedance. This provides some flexibility in dealing with reactive loads, but typically is still limited to less than 50 uF, and the transient response is still either slower than necessary or too fast, causing overshoots and oscillations. NI SMUs with digital control loops, however, give you full control of the transient response and allow you to tune the SMU for your specific DUT. When using SMUs with digital control loops, you can achieve very fast rise times with no overshoots or oscillations, ensuring you have the fastest possible response while maintaining system stability and protecting your DUT from overvoltage or overcurrent conditions. These SMUs also allow you to use the device as a fast transient power supply for powering RFICs and mobile devices.


Figure 5. Customizable Transient Response of an SMU With a Digital Control Loop


The figure above shows the SMU output of a 1 V step into a 100 nF load. Because the SMU is using a digital control loop technology called NI SourceAdapt, you can tune the output of the SMU to your specific DUT, instead of settling with a predefined transient response.


Back to Top

6. Multifunction SMUs

Some SMUs are optimized for specific applications or requirements, such as leakage testing, high-power IV sweeps, or powering mobile devices with a fast transient response. While these instruments are ideal for serving a specific function, you may need multiple SMUs to fully characterize devices that require a variety of different stimulus or functionality. Alternatively, “system” SMUs combine high-power, high-precision, and high-speed source measure capability into a single instrument. This gives you the flexibility to perform many different functions and test a broad range of devices with the same instrument. This not only simplifies connectivity, but also reduces the number of different instruments, test system footprint, and overall cost.


System SMUs have a wide DC power boundary that gives you the ability to source and measure both high voltage and high current. Additionally, many of these SMUs further extend the DC power boundary by implementing an extended pulsing range where you can source or sink high-power pulses. This additional functionality is critical for performing high-power IV sweeps without cascading the output of multiple SMUs to achieve higher current output. Combining high-power pulsing with the precision measurement capabilities of system SMUs makes them ideal for devices such as high-brightness LEDs or power transistors that require both high-power IV sweeps and low current measurements.


Figure 6. IV Boundary of a NI System SMU With an Extended Pulse Boundary


The figure above shows the IV boundary for a system SMU, with a 20 W DC power boundary and additional 200 W and 500 W pulse boundary.


Back to Top

7. Conclusion

SMUs provide improved performance over basic power supplies by providing higher precision, four-quadrant operation, and specialized features for IV characterization. This functionality makes them essential instruments for lab characterization and automated test applications, especially within the semiconductor industry. Additionally, system SMUs are designed to address a broad range of applications by combining high-power, high-precision, and high-speed source measure capabilities into a single instrument. This not only simplifies connectivity, but also reduces the number of different instruments, test system footprint, and overall cost.


Back to Top

Bookmark & Share


Rate this document

Answered Your Question?
Yes No