In the world of precision resistors, temperature fluctuations are the enemy. Although precision resistors are designed to be as stable as possible, temperature still has an effect on the resistors’ properties, albeit minimal. Some of the effects are widely understood, like the relationship between temperature and resistance, or resistor self-heating. But there’s another temperature driven effect that generally only analogue circuit designers know about (or care about), which actually generates a voltage from the resistor: the thermal EMF effect.
The thermal EMF effect (or more accurately, the Seebeck effect) is an interesting phenomenon. Essentially, when wires made of two different metals are joined together, if the wires are at different temperatures an EMF is produced (otherwise known as a voltage). This is down to charge carriers migrating from the hot material to the cold material. The effect happens for any type of conductive materials, provided there is a junction between two materials that are not the same, but different combinations of materials can result in more or less EMF being produced per unit temperature.
The thermal EMF effect is widely known and used extensively for measuring temperature, in the classic format of two different types of wires twisted together and used as a sensor. However, not many people know that it can also apply to resistors. It isn’t that much of a stretch of the imagination when you consider that a wirewound resistor is basically a wire lead, joined to a coil of wire, joined to a wire lead, and the leads are made from a different material than the wire in the body is. Perhaps slightly less obvious is that this applies to all the other types of resistors too, to a greater or lesser extent.
For example, ordinary (non-precision) carbon resistors can produce a few hundred microvolts per degrees C (µV/°C) while some metal film resistors can produce tens of µV/°C. Precision resistors such as Evenohm wirewound types may still produce a couple of µV/°C. This introduces a DC voltage in the circuit, seemingly out of nowhere! For most applications, the voltage produced is too small to make a noticeable difference, but for precision resistor applications like high gain, critically balanced circuits or those where low ohm values are required, it can be a significant source of error.
The best way to minimise this effect is to keep the leads and the body of the resistor at the same temperature, as far as possible. Orientation is the key word here, which may sound simple, but there are a few of things to consider. Firstly, heat rises, so consider the resistor’s orientation relative to that. Placing resistors vertically on end to save space is a no-no, except of course if your board will be used vertically, in which case, it’s essential. Secondly, consider the orientation relative to any heat generating components on the board, and place resistors perpendicular to the heat source wherever possible. Thirdly, if there’s a fan in the system, the resistors need to be perpendicular to the air flow.
Ultimately, the best solution would be to minimise thermal gradients around critical parts of the circuit, but since that isn’t always possible, considering the resistors’ orientation can help eliminate erroneous voltages that can lead to noise problems.