Thermal EMF (seebeck effect) on PCBs

Question

Can poor PCB fabrication process / assembly, and the type of solder used cause thermal EMF (seebeck on PCB) problem? Is it affected by the type of the material used? For example, quality of plating, vias, use of different metals such as gold, tin, copper etc?

Answer 1

The Seebeck effect is always there, and has nothing to do with the quality of the PCB fabrication process. Copper is copper, and exhibits a certain Seebeck effect.

Unless you have a very sensitive low-level analog circuit, the Seebeck effect can be ignored on a normal circuit board.

First, for there to be a voltage offset due to the Seebeck effect, there has to be a thermal gradient. The whole PCB at the same temperate won't cause any offsets, regardless of what the temperature is.

Second, even with thermal gradients across the board, the offset is 0 over any loop of copper traces. Whatever offset voltage is caused along the gradient to a different temperature going out, is offset by the reverse gradient coming back to the starting temperature.

Third, the offset voltages due to the Seebeck effect are small. Copper generates about 6.5 µV/°C. Even if one side of a board is 50 °C hotter than the other, that only causes 325 µV offset. And again, you generally can't sense that even if you wanted to because this cancels out in a loop.

Thermocouples exploit the Seebeck effect by using two different materials out and back. The voltage offset seen at the electronics at room temperature is the difference between that generated by the two materials across the temperature difference.

The most common reason for considering the Seebeck effect on a circuit board is when designing thermocouple receivers. Since a thermocouple measures the difference in temperature, not the absolute temperature, you have to know the temperature of the junction where the thermocouple wires are connected to copper traces on your board. Those two junctions also need to be at the same temperature.

In high-accuracy thermocouple receiver circuits, this is usually done by keeping the two junctions physically close and clamping a copper bar across them. The copper is electrically insulated from the junctions, but thermally connected as best as possible. Since copper is a good thermal conductor, the two junctions will hopefully be very close in temperature to each other, and to the absolute temperature sensor on the board that is used as the reference temperature.

Answer 2

Yes, and it can be an issue when trying to build metrology grade instruments.

Typically you see this stuff sweated about by people like Keithly and Keysight when designing 7 digit volt meters where thermal EMFs can really matter.

Other fun things can be thermally induced stress causing oscillators to change frequency, and capacitors to acquire charge, lots of fun things to worry about when playing in that space.

Quite often you see PCBs with slots cut in them to limit leakage (probably a bigger issue with cheap boards), and I have done this myself when dealing with giga ohm impedance.

Back in the day Cadmium based solders were used for low thermal emf connections with copper, ROHS has made that more difficult then it once was....