We often take resistors for granted. They are one of the fundamental building blocks of electronics design but are often given little thought when choosing them. You will select a value and the tolerance, and probably briefly consider the power dissipation and provided that isn’t very high you will most likely be picking all your resistors from the same range, for example, all 1%, 0603, 100mW. However, make sure you don’t forget other characteristics which could be relevant. Here are just a few parameters to consider.
Resistors have a voltage rating. An 0603 resistor might be limited to 50V. This is nothing to do with the power dissipation — if it was a 10ohm resistor, then you wouldn’t be able to put 50V across it. However, if you have a high-value resistive divider in a power supply and there will be more than 50V across it, you will need to rethink the choice. In general, it means increasing the physical size and using 0805 or 1206 resistors. Another option is to use two or more equal values in series to reduce the voltage requirement for the individual resistors.
Temperature coefficient of resistance (TCR)
1% resistors are not expensive, but you also need to consider the temperature coefficient if you are looking for accuracy. A 1% resistor might have a 100ppm/C temperature coefficient so over 100C could vary by 1% in addition to your initial tolerance. To reduce the temperature coefficient, you will most likely find that you have to buy a more accurate resistor than 1% e.g. 0.1% to get 25ppm/C. Once you start looking at 0.1% resistors, you will find tolerances as low as 10ppm/C such as the Panasonic ERA3ARB series or even 2ppm/C for 0.01%.
All resistors have a thermal noise (Johnson noise) due to their value which is sqrt(4kTRB) where k is Boltzmann’s constant, T is absolute temperature, R is the resistance and B is the bandwidth. However, there will also be additional noise due to the current passing through the resistor. This usually isn’t quoted by resistor manufacturers but can be important in some applications.
This graph from the Vishay D/CRCW datasheet shows the effect (although Vishay has dropped that graph from their recent data sheets). If you had a high-value resistor in a transimpedance amplifier for amplifying a photodiode current, then if there was a DC current due to ambient light then you could see an effect on your noise levels due to resistor current noise. From the graph above you can see that the problem is worse with physically smaller resistors so by using 1206 resistors instead of 0402 ones could reduce current noise by a factor of up to 8. In a lot of applications, current noise will not be a problem, but you need to check if you are designing a low noise circuit. In particular, if the voltage across a resistor is zero then there will be no current and hence no current noise (but there will still be thermal noise). As with VCR below, most resistors have no current noise quoted so if you think it might be a problem in your low noise design you may have to make your own measurements or contact the resistor manufacturers for more information.
Voltage coefficient of resistance (VCR)
Resistors can have a small variation in resistance based on the voltage applied across them. Resistor manufacturers rarely quote this variation — maybe it is considerably smaller than the absolute tolerance— but the effect is different. For example, a resistor with a resistance that varies with applied voltage can cause distortion in the signal depending on the exact purpose of the resistor. In some cases, the data sheet quotes the effect as a third harmonic distortion figure for the resistor, but in most cases, it is simply not mentioned. If you are designing very low distortion circuits (and low noise circuits), then you may need to search for resistors aimed at precision, low noise applications. Even so, you may struggle to find a figure for VCR or third harmonic distortion in the datasheets. If you think your low distortion design may require a very low VCR you may have to make your own measurements to compare resistors or contact the manufacturers for more information.
Pulse power dissipation
Some applications require a high pulse power capability but low average power dissipation, such as with current limiting or protection in pulsed circuits. You cannot simply use the average power to choose the resistor without checking the peak capability — you could end up blowing the resistor like a fuse. Some resistors such as the Bourns CRS series are anti-surge resistors and so have a graph to show the surge rating as shown below.
So, an 0805 CRS series resistor can withstand 100W if the duration is less than 100µs.
This is another tricky parameter because manufacturers don’t seem to quote it. Data sheets list some resistors as “low EMF” but then don’t give a value. Most resistors simply don’t mention it. As a thermal EMF arises due to dissimilar metals, the thermal voltages in a resistor caused by the change from the solder terminals to the resistive material should cancel out as there should be two equal and opposite voltages. If the two resistor terminals are at different temperatures, then the voltages won’t cancel out. Imperfections in the resistor could cause a residual thermal EMF. Again, for very high precision circuits you need to contact the resistor manufacturer or make your own measurements.