By Stephen Oxley, Senior Engineer, Applications & Marketing (resistors), Sensors and Specialist Components
Everyone knows that resistors are relatively simple components — except that they’re not. Certainly, they are passive two-terminal devices with their basic behaviour defined by Ohm’s Law serving many functions: current limiting, current sensing, capacitor bleeders, voltage balancing, signal attenuation, and rise/fall time control, to cite a few roles.
Nonetheless, it may seem that once you have calculated the needed resistance value, most of the work associated with resistor selection is done. However, there’s much more to selection than identifying just a few basic parameters. Reality is that there are many other factors that go into choosing the right resistor type for the application, and it’s easy to overlook them.
Among the factors that are easy to neglect:
● Power rating: by their nature and the basic P = I2R relationship, resistors are self-heating components. This means that the resistor must be able to handle the dissipation and still meet specifications (Figure 1) and may need a heat sink in extreme cases. At the same time, simply over-specifying and choosing a resistor with excessive safety margin adds unnecessary cost and real estate. If operating in high temperature environments, look carefully at the derating curves on the vendor datasheet. In particular, note whether the temperature axis relates to ambient air, termination or heatsink temperature.
Figure 1: The WSMHP series of non-inductive thick-film resistors is available in TO-263 SMT packages, with ratings to 35 W.
● Pulsed loads: even if the average power the resistor dissipates is within its rating, bursts of higher-power pulses can exceed that rating by several orders of magnitude. If this is important for the application, a resistor with specified pulse performance should be chosen, remembering that thick film and wirewound are the commonest technologies for pulse resistors. A careful analysis of pulse levels and duty cycles is important, in order to establish that the mean power is within the power rating.
● Temperature Coefficient of Resistance (TCR): self-heating, as well as any other heat sources in the area, will cause a resistor’s ohmic value to change. For some situations, such as a basic pull-up resistor, the change may be of little concern. However, if that resistor is being used to measure current flow via voltage drop, the TCR-induced change may severely affect the accuracy of the reading. The same concerns apply to critical resistors in many instrumentation applications. Do the calculations and choose a resistor with low-enough TCR.
● Allowing for cooling: as they are heat sources, resistors will often need cooling, usually via passive convection but possibly using forced air. Be sure to provide a clear path for the cooling airflow, and that the resistor itself is not shadowed by larger nearby components which will block that needed flow. If you are using the PC board itself as a heat spreader, don’t let other heat-generating components near the resistor diminish the heat-absorbing potential on which you are relying.
● Voltage rating: resistors are often used in higher-voltage applications, with potentials across their leads reaching into hundreds or thousands of volts. Be sure the resistor is rated for the circuit voltage (Figure 2) independent of the current and power levels, or you may have arcing and outright failure.
Figure 2: For high-voltage applications, various resistor families are available with lead spacing and other attributes needed to support circuit functions in the tens of kilovolts.
● Resistor mounting: Many higher-dissipation resistors are larger than typical small ICs and other passives, so how and where the resistor will be placed is an issue. Is the resistor a surface-mount device, or through hole? Does it need any sort of mechanical support or at least need to be placed where the PC board is not flexing?
● Self-inductance: for DC applications, the inductance of the resistor is not an issue, but it may become one at higher frequencies. Resistors are available with very low self-inductance for situations which are sensitive to this parameter.
● Adequate connections: as current levels increase, it’s critical to plan for solid, low-resistance connections for unimpeded current flow and minimal IR drop, avoiding reliance on thin PC board tracks. Be sure the resistance of the current-carrying leads to the resistor is insignificant compared to the resistor’s value. Don’t ignore the possible need for Kelvin (four-wire) connections, if you are trying to accurately measure the voltage across the resistor, as well as leaving a place for the amplifier (often a differential or isolated unit) which senses the voltage, as it needs to be located close to the resistor itself, to minimise noise pickup.
● Finally, and perhaps most difficult: choosing an appropriate resistor type. Once you get beyond the basic low-power, small chip resistors which are used in non-critical applications, you face a world of distinctive resistor types, including fusible wirewound, steel substrate, and silicon based networks among others. Some have extremely low TCR, some are mechanically rugged (Figure 3) some are resistant to contaminants such as sulphur, some can handle pulse overloads, and some have well-defined, open-circuit failure modes.
Figure 3: The extremely rugged, high dissipation WDBR resistor series uses multilayer construction built on a stainless-steel substrate to make it highly immune to cracks and fracturing which could result from thermal extremes and vibration.
Unless your resistor application is routine, your smartest and least-risky option is to work with the application specialists at a resistor vendor such as TT Electronics, since they have the experience and expertise to deal with situations such as such as yours.