Testing passive devices for thermals and noise effects

The use of passive devices in low-voltage applications is becoming increasingly common. For example, laptops and cell phones require batteries with ultra low internal (ionic) resistance. So manufacturers and users of passive components must now pay greater attention to the materials and processes used to build them, as well as new rules about using them in low-voltage applications, in order to prevent problems previously unseen.

Thermoelectric EMFs are most common in passive components constructed of dissimilar materials, such as resistors. Design and manufacturing activities, such as testing batteries with ionic resistance, demand greater focus on the impact of noise and thermoelectric EMFs on measurement accuracy. This can be particularly challenging because very low test currents are required to measure these single-milliohm resistances.

Understanding the challenges of using passive components in low-voltage circuits, the sources of error, and proper means for characterizing them is crucial to the proper design of low-voltage devices. The test for susceptibility described here can help engineers identify which passive devices are most suitable for such applications.

Mechanics of noise

Though often lumped together, the mechanisms of thermal noise, low-frequency noise, and thermoelectric EMFs (or thermals) aren’t identical.

Thermal or Johnson noise develops when the trajectories of conduction electrons in a material are altered by the thermal vibrations of its molecules. These scattered electrons develop noise voltages when they travel through the resistance of their pathway. The spectral content of this noise is white, dependant only on the material’s temperature and resistance. All components with non-zero resistance and non-zero temperature produce some level of thermal noise.

Depending on the material or device involved, at lower frequencies, additional 1/f noise terms may begin to emerge from within the white noise floor. This kind of noise is frequently the result of material defects that behave as an electron (or hole) trap. These noise sources have the annoying property of increasing in magnitude as the bandwidth decreases, so digital filters (which increase signal observation time) aren’t helpful.

Mechanics of thermals

Thermoelectric EMFs develop wherever dissimilar materials come together in the presence of a temperature gradient. Three different effects play a role: the Seebeck, Thomson, and Peltier effects.

Seebeck effect produces a voltage that is observed when a uniformly conductive material experiences an end-to-end temperature gradient. The voltage is present whether the circuit is closed (non-zero current) or open (zero current). In practice, this voltage is difficult to measure in a single conductor, but when different kinds of conductors or materials (each with a different Seebeck coefficient) are paired to form a thermocouple, it’s much simpler to measure the difference.

When the Seebeck effect is observed within a uniformly conductive material, if the circuit is closed and a current is allowed to flow, the Thomson effect can also be observed. The Thomson effect describes how these charge carriers absorb or release heat as they pass through a temperature gradient.

The Peltier effect occurs at the junction of two dissimilar materials. Charge carriers (current) must be flowing to observe a temperature difference generated by the current passing between the dissimilar materials. Although the Peltier effect doesn’t generate a voltage, the resulting temperature changes can affect the Thomson and Seebeck voltages.

Measurement impacts

Dissimilar materials, currents, and temperature gradients are all characteristics or consequences of using passive components, so these EMFs can have a significant impact on measurement accuracy. The measurement bandwidth determines the magnitude of the inaccuracies each contributes when passives are used in real applications.

Generally, the total signal observation time determines the low-frequency bandwidth limit; the width (aperture) of the A/D conversion’s integration period determines the high-frequency bandwidth limit. With thermoelectric EMFs, these offsets persist even at dc, as long as there’s a dc temperature gradient created by a point heat source, either within the passive component or somewhere on the circuit board.

However, for passives, a change in power dissipated within or around the passive causes a temporary gradient, which equalizes according to the thermal time constants of the component as situated on the printed circuit board (PCB). In other words, the EMF can be time dependent, diminishing to near zero when temperatures are in equilibrium.

The selection process for resistors usually revolves around cost and such application requirements as temperature coefficient, voltage ratings, voltage coefficient, power requirements, and physical size. The resistor’s noise and thermoelectric properties are rarely considered.

Applications across the board have been affected by the rapid regression of supply voltage across all semiconductor industries from microprocessors to high-performance analog. Lower supply voltages require less silicon (and less cost) and can be assembled into smaller packages. However, lower volatges also mean that noise will be more significant on a percentage basis. If power/currents are lowered in concert with voltage, the same can be said for noise currents. Fortunately, smaller components can usually be PCB-mounted in such a way that minimizes thermal time constants, leading to a desirable thermoelectric footprint.

Testing

Precision and low-voltage applications like battery testing sometimes require characterizing passive components (in addition to their PCB connections). This can be tricky because the magnitude of all three thermoelectric EMF contributors depends on the temperature gradients of the connecting wires and within the components.

This last term is often overlooked. If the passive component isn’t built correctly from the correct materials, a step change in temperature of the entire component will cause temperature gradients within the component, generating an EMF. Measuring the same EMF twice can be difficult due to their dependence upon the temperature gradient and the rate of change of this gradient.

Many passive components are manufactured using different kinds of materials needed to elicit other desirable properties (for example, low temperature coefficient of resistance or a stable capacitance with applied voltage), so it’s impossible to avoid generating thermoelectric voltages within them completely. However, avoiding the unwanted temperature gradient(s) also required to generate thermoelectric EMFs is possible. A passive component’s design an

Previous: Provide four times faster transmission speeds than standard wireless

Next: Infineon IPG20N04 / IPG20N06 OptiMOS / OptiMOS-?T2 Power Transistors