Design engineers who have been to an EMC testing laboratory must be familiar with the large selection of ferrite cores that a test lab often provides. For a product with a mains lead or a system consisting of long cables (HDMI, USB, etc.), the most common EMC failures often include conducted emissions in the frequency range of 9 kHz and 30 MHz and radiated emissions in the 30-300 MHz range. In such cases, cables often become effective antennas, and ferrite cores are generally used to locate and suppress the noise (at least during the troubleshooting stage).
The practice of using ferrite cores on cables is often performed using a trial-and-error approach. Engineers place a ferrite core on one end of a cable and measure the performance. If the noise at the frequency range of interest is reduced with the placement of the ferrite core, this means the approach works. If not, the ferrite core is then removed and placed in another cable.
In this column, I offer a brief summary of a more systematic approach for using ferrite cores on cables. Hopefully, this summary can serve as a “ferrite core checklist” for design and test engineers.
To start with, we need to understand some basics of a cable antenna. For example, a 1-meter-long cable could, depending on its connections, serve as either a half wavelength or a quarter wavelength antenna. If both ends of the cable are fastened to the equipment chassis, chances are that the characteristic impedances at both ends are low, so the cable is a half wavelength antenna. Since 1 meter is a half wavelength for 150MHz signals, this cable will radiate quite efficiently for noise around 150MHz.
However, suppose one end of the cable is fastened to the equipment chassis while the other end is in free space (at a high impedance). In that case, the cable becomes a quarter wavelength antenna (you can treat the equipment chassis as the other half of the antenna). Common mode current at around 75MHz range starts to flow from the low impendence end of the cable to the high impedance end, which translates to radiation at 75MHz.
Therefore, for a quarter wavelength cable, adding a ferrite (which effectively increases the characteristic impedance) only works if it is placed at the lower impedance end of the cable (often the equipment chassis end). Increasing the impedance at the lower impedance end breaks the boundary condition of a quarter wavelength antenna, which results in less common mode current at the frequency of interest. If ferrite cores are located at the higher impedance end rather than the lower impedance end of a cable, chances are that the ferrite cores will prove ineffective, or in some cases, make the emissions worse. (See my YouTube video at https://youtu.be/HMvpml48v4g that demonstrates the point.)
My first advice for using ferrites is to treat them as a resistive (lossy) rather than as an inductive component. Therefore, when selecting the ferrite grade, it is often more important to aim for maximum resistive loss.
Another part that engineers often overlook is that the impedance provided by a single-turn configuration ferrite core is simply not large enough. When we placed a ferrite core on a cable and saw no impact, we could easily conclude that this approach was ineffective. Checking the datasheet from ferrite manufacturers, one can see that a 1-turn configuration has only 25% of the impedance compared with a 2-turn configuration at its effective frequency range. (Note that, as frequency increases, the impedance of a multi-turn configuration ferrite drops rapidly due to the parasitic capacitance introduced by the turn-to-turn windings. Hence the multi-turn advantage is progressively lost above a certain frequency.)
This suggests that sometimes a single-turn ferrite core solution is not effective enough, and a multi-turn configuration should be tried in such cases. In cases of large diameter cables where bending radius limits a multi-turn ferrite option, we suggest using a few ferrite cores in series to increase its impedance.
Another benefit of placing a ferrite core near the equipment chassis is that it forms an R-L-C filter with the chassis/ground plane within the equipment. Because ferrite material is essentially a ceramic, it has both high permittivity and permeability. Therefore, placing a ferrite near a conductive surface will increase its capacitance. Placing a ferrite core near the equipment chassis will improve the filtering performance compared with using the ferrite in free space.
A short column like this cannot cover every aspect of using ferrites on cables. Secondary effects such as saturation, leakage resistance, etc., are also important. Hopefully, the following checklist can provide effective guidance for engineers who wish to use ferrites as a valuable aid to electromagnetic compatibility.
- For ferrite cores used as common mode suppressors, saturation is generally not possible. However, if you must place a core around a single conductor, be sure that the current does not exceed the saturation current.
- Always check the manufacturer’s datasheet and select ferrite materials for maximum resistive loss (that is, not for inductance).
- For a single-turn configuration, long sleeve ferrites are preferred as the impedance is proportional to the length of a ferrite core. The best performance is often achieved if the ferrite fits the cable snugly.
- For a multi-turn configuration, a fat toroid shape is preferred. Keeping the wires wide apart helps improve the performance at high-frequency range (for the reasons we explained before).
- Place the ferrite at the lower characteristic impedance end of a cable, e.g., equipment chassis.
- If possible, place the ferrite on a conductive surface (such as the ground plane of a system).
- If possible, always try maximum impedance (i.e., multi-turn configuration) first.
- In rare cases where placing a ferrite could increase emissions of signals at a certain frequency, do not remove the ferrite cores. Instead, place another ferrite at the other end of the cable.
Dr. Zhang –
It occurred to me when reading your article that if you have a cable going from one chassis to another it is a half wave antenna as you said but if you put a ferrite at one end you raise the impedance at that end changing it to a quarter wave antenna. Then if the chassis at the other end has a source that resonates at that quarter wave frequency the emissions spectrum can change which is why you may need to add a ferrite at each end.