Does the Ferrite Core on a Power Cable Actually Do Anything? Measuring Conducted Noise
イチケン / ICHIKEN
Many power cables for appliances, switching power supplies, and computers have a ferrite core attached. Most people use these cables without thinking about the lump. In this video, Ichiken asks how much that ferrite core actually reduces noise, and tests it with the measuring equipment they have available. They also compare it against other countermeasures: a larger ferrite core, a dedicated AC line filter, a noise-cut transformer, and power supplies that produce less noise to begin with.
What the ferrite core is supposed to do
Ichiken first compares a normal power cable with one that has a ferrite core. Inside the molded section, the ferrite is split into halves that clamp shut around the wires. This forms a magnetic path that acts as a filter. Because the core surrounds all the conductors together, its main effect is on common-mode noise.
They then separate the two main kinds of noise. Conducted noise comes from a source such as a switching power supply and travels out through the power cord connected to it. Radiated noise is emitted into space from the source. According to Ichiken, a ferrite core on a power cord mainly attenuates conducted noise, and it does this by converting the noise into heat in the ferrite.
The practical purpose is to keep noise from spreading. If conducted noise is attenuated at the cord, less of it reaches other products plugged into the same power strip. Strong noise could otherwise make that other equipment malfunction. The core also works in the other direction, reducing conducted noise that comes in from outside.
Ichiken adds a caution. A ferrite core on the supply line does not reduce the noise the source actually generates. It only stops some of that noise from getting out, or from getting in. To reduce the noise itself, the root cause has to be removed, although they acknowledge that is sometimes difficult.
The measurement setup
The test chain has four parts. A noise-cut transformer connects to the AC mains and supplies relatively clean power by removing noise from other equipment. Next comes a LISN (line impedance stabilization network, described in the video as a pseudo power supply network). The LISN removes conducted noise arriving from its input again. It also has a terminal that outputs only the conducted noise coming from the device under test. The device under test, the noise source, plugs into the LISN, and the measuring instrument connects to that terminal.
Conducted noise is usually measured with a spectrum analyzer. Here Ichiken used the spectrum analysis function of an oscilloscope. They also note a limitation. For reproducibility, conducted noise tests are normally done with the setup placed on a grounded metal plane, but they could not do that because of equipment constraints. This point comes up again when they interpret the results.
A dubious switching power supply as the noise source
The first device under test is a switching power supply Ichiken had lying around. It has no PSE mark, and from the circuit they could see no noise filter on the power input. They describe it as a bad supply that looks likely to spread all kinds of noise.
Before connecting anything, the FFT display shows only background noise. The horizontal axis spans 150 kHz to 30 MHz, which Ichiken notes is the same range used in conducted noise standards. The vertical axis shows noise intensity, so a higher trace means stronger conducted noise.
With a plain power cable, the noise rose sharply across the band. Ichiken's comment was that a lot of noise was coming out.
The bundled ferrite-core cable: barely any difference
Next they used the ferrite-core cable that came in the box. On first look the result seemed almost the same. Comparing the traces directly, conducted noise appeared slightly lower below about 5 MHz. Ichiken hedges this: the difference could be caused by the poor reproducibility of the setup. Since there was a small decrease, the core might have some effect, but they were not confident about it.
A large ferrite core with a loop
For a stronger test, Ichiken looped the power cable and passed it through a large ferrite core. In principle, a bigger core has a larger cross-section, and looping the cable adds turns to the resulting inductor, so it should filter noise better. In practice, the spectrum again changed very little. Ichiken felt that measurement reproducibility was a bigger factor than the core itself.
To explain this, they measured the common-mode inductance of each configuration:
- Cable without ferrite core: 4.0 μH
- Cable with the bundled ferrite core: 5.5 μH
- Cable looped through the large core: 150 μH
A core that the cable passes through only once barely changed the inductance compared with a bare cable, which matches the lack of noise reduction. Yet even the large core, with far higher inductance, left the conducted noise largely unchanged. Ichiken's conclusion for this supply was that adding a ferrite core alone does not seem to be effective.
A dedicated AC line noise filter
Ichiken then tried a power line filter containing capacitors and a 0.3 mH common-mode inductor. The low-frequency part of the spectrum stayed about the same. Above 500 kHz, however, the noise dropped substantially, although a peak around 20 MHz remained. Ichiken judged this a clear improvement and concluded that a dedicated AC line filter is a better choice than a ferrite core alone. They add that this type of filter reduces both outgoing noise and noise coming in from outside, and they call it quite essential in noisy environments.
Using the noise-cut transformer as the filter
The next experiment used the noise-cut transformer itself as the filter, with the circuit rearranged accordingly. For this setup, the power came from an AC power supply made by NF Corporation, and some background noise was visible even with nothing connected. Ichiken treats that as unavoidable.
With the noisy supply connected, low-frequency conducted noise was heavily attenuated, and the higher frequencies were close to background level. Ichiken called the result excellent, as expected from a noise-cut transformer. They also note that putting such a transformer in front of every device is unrealistic outside special situations, so in normal use a dedicated AC line filter is the practical option.
A properly made switching power supply
Ichiken points out that the supply tested so far spreads an unusually bad amount of noise. For comparison, they tested an ordinary switching power supply made by a reputable company, with no added filter. Its conducted noise was far lower from the start, and Ichiken says this is how a power supply's conducted noise should normally look.
Opening the supply showed why. Near the AC input there are capacitors, common-mode chokes, and other filter components. The power conversion stage still generates noise, but because the filter is built into the circuit, little conducted noise gets out.
A low-noise power supply
Last, Ichiken measured a low-noise power supply from Daitron, which the channel had covered in an earlier video. They wanted to see whether a supply marketed as low-noise also produces little conducted noise. The measured conducted noise was very low, which Ichiken called very high noise performance. They attribute this to the internal circuit, which is designed for soft switching. Because the supply generates little noise internally, less conducted noise comes out.
Conclusion
For the power supply tested in this video, Ichiken found that the ferrite core on the power cord could not attenuate conducted noise much, even with a large core. Adding a filter circuit afterward, such as a dedicated AC line filter, worked much better. They describe that approach as a coping measure, however, because it does not reduce the noise the source produces. Their final point is that serious noise control means designing the system so the noise source itself emits less, as the properly filtered and soft-switching supplies in the video show.
This video is brought to you courtesy of DigiKey. Hello. Today, we will examine whether the ferrite core on a power cord is effective.
If you look at the power cables for appliances, switching power supplies, and computers, they may have a ferrite core attached. This ferrite core is there to keep noise generated inside electrical appliances from traveling out through the power cord. Most people probably use it without paying much attention to the ferrite core. We will verify how effective it really is using the measuring instruments here.
But first, let me introduce today's sponsor, DigiKey. Go ahead.
DigiKey is fully supporting an electronics contest! This will be DigiKey's first contest of its kind in Japan. It is the first contest, so please try making and submitting electronic projects related to the number "1". I, Ichiken, will be participating as a special judge this time! I see here that there is also an Ichiken Award.
The winning entries will be exhibited at Maker Faire Tokyo 2026. Ichiken will also interview you on the day of the exhibition. For details on how to apply and the judging criteria, please check the link in the description. Applications will be accepted until June 22, 2026, so if you have created a project, please remember to submit it!
First, let's look at the power cables. On the left is a normal power cable, and on the right is a power cable with a ferrite core. This part is the ferrite core. Let's look at the structure of the ferrite core section. It contains a split ferrite core like this. It closes around the wires, forming a magnetic path and acting as a filter. This ferrite core clamps around all the wires together, so its main effect is on common-mode noise.
There are many kinds of noise, mainly conducted and radiated noise. Conducted noise is electromagnetic noise transmitted through a power cord connected to a noise source, such as a switching power supply. Radiated noise is electromagnetic noise radiated into space from a noise source.
The noise that this ferrite core can attenuate is mainly conducted noise. In the case of this power cord, there is a ferrite core here. A ferrite core can be attached here, here, or at the bottom, for example. The conducted noise transmitted from the noise source can then be attenuated here. When we say it attenuates noise, the noise is actually converted to heat at the ferrite core.
If conducted noise can be attenuated here, it is less likely to be transmitted to other products connected by power strips and so on. If the noise is too strong, it could cause other equipment to malfunction.
The role of attaching a ferrite core is to reduce conducted noise traveling out from noise sources. Even if conducted noise is transmitted from outside, a ferrite core reduces the incoming noise. The effect of conducted noise on other connected devices can be minimized as much as possible.
A word of caution here: the ferrite core in the power supply line does not reduce the conducted noise generated by the noise source itself. It only keeps it from going out or coming in. To reduce conducted noise emitted by the noise source itself, the root cause of noise generation must be eliminated. However, it is sometimes difficult to eliminate the noise source itself.
Here is the measurement setup for this time. On the right is a noise-cut transformer. In the middle is a LISN, a line impedance stabilization network. Here is the specimen, which in this case is the noise source. Last are the measuring instruments.
Here is the connection diagram. Connect the noise-cut transformer to the AC power supply, then connect the LISN to it. Connect a product that produces conducted noise, such as a switching power supply, to the LISN. The LISN also has a terminal that allows us to measure only the conducted noise coming from the noise source. Connect the measuring instrument to it.
Here is an explanation of their respective roles. The noise-cut transformer removes noise from other products, providing a relatively clean power supply with low noise content. The LISN has power input and output terminals and a signal measurement terminal. Conducted noise coming in from the input is removed here again. In addition, only the conducted noise component transmitted from the noise source being measured can be extracted as a signal.
For measuring instruments, we mostly use spectrum analyzers. This time, I decided to use the spectrum analyzer function on the oscilloscope. To ensure reproducibility, conducted noise measurements are normally performed on a grounded metal conductor. That is not being done this time due to equipment limitations.
Let's start with an experiment. This is a switching power supply I had lying around. Looking at the circuit, it doesn't seem to have a noise filter in the power supply line. It doesn't even have a PSE mark. It looks like a bad power supply that spreads all kinds of noise.
What you see on the oscilloscope now is the FFT result showing how much noise is being produced. Nothing is connected to the LISN, so what you are seeing now is background noise.
First, let's try a normal power cable without a ferrite core. When connected, this is the waveform. It's pretty amazing. There is quite a lot of conducted noise. The horizontal axis covers a range of 150 kHz to 30 MHz. This is the same as the standard for conducted noise measurement. The vertical axis shows the intensity of conducted noise. The larger this is, the stronger the conducted noise. There is quite a bit.
Next, I'll try this power cable with a ferrite core. When connected, it looks almost the same. Comparing the results with and without the ferrite core, conducted noise seems to be slightly reduced at frequencies below 5 MHz. However, this result could be due to the reproducibility of the experimental setup. There is a slight decrease, so perhaps there is some effect.
I wasn't too sure about its effectiveness, so I'll try an additional experiment. I'll make a loop with the power cable here and put a large ferrite core through it. In principle, a large ferrite core increases the cross-sectional area. Looping the power cable also increases the number of inductor turns, which should be advantageous as a noise filter.
I tried it and got this result. With or without a large ferrite core, not much has changed. If anything, the reproducibility of the experiment seems to be the bigger issue. The ferrite cores we have experimented with so far don't seem to have any significant effect.
Let's check how much common-mode inductance these power cables have. This ferrite core surrounds the wires together, so it acts on common-mode noise. With the ferrite core, it was 5.5 μH. Without a core, it was 4.0 μH. With the large core, it was 150 μH.
At least with a ferrite core that the cable passes through only once, the inductance value changed very little compared with the cable without a ferrite core. In fact, the conducted noise was not reduced that much. On the other hand, even with such a large ferrite core attached, the conducted noise remained largely unchanged. For this switching power supply, simply adding a ferrite core does not seem to be effective.
It wouldn't be fun to end here, so I'll try other noise filters. This is a filter for the power supply line. It contains capacitors and a 0.3 mH common-mode inductor. When used, this is the result. The low-frequency part of the spectrum remains largely unchanged, but above 500 kHz, the noise is much lower. The peak around 20 MHz remains the same. We are seeing quite a bit of effectiveness.
Rather than using a ferrite core by itself, it looks like we should use a dedicated noise filter for the AC line. Power line filters like this reduce outgoing conducted noise and are also effective in reducing conducted noise coming in from outside. It is a rather essential filter in noisy environments.
For the next experiment, I'll use the noise-cut transformer itself as a noise filter. I changed the configuration a bit to make the circuit look like this. First, look at the background noise with nothing connected. In this case, we are using an AC power supply from NF Circuit Design. Some background noise is visible. This part is unavoidable.
I'll connect a noise source to it. The low-frequency conducted noise has been attenuated considerably. The higher-frequency parts are also almost at the background noise level. As I expected from a noise-cut transformer, it is quite excellent. However, providing such a noise-cut transformer for each device is not realistic except under special circumstances. You would have to use a dedicated AC line noise filter like the one we used earlier.
The switching power supply I have used so far is a pretty bad one that spreads a lot of noise. Let's also look at a typical switching power supply made by a proper company. I'll turn it on.
I think it's pretty good. I don't have a noise-cut filter connected right now. The switching power supply itself produces less noise, so the conducted noise starts at a much lower level. In fact, this is how the conducted noise of a power supply normally has to be.
If you look inside the switching power supply, this is the AC line input. There are capacitors, common-mode chokes, and various other components. Noise is produced in these power conversion parts, but because the filter is already in the circuit, the outgoing conducted noise is low.
There are also low-noise power supplies. Since it is low-noise, you might think its conducted noise would also be low. Let's measure it. This is a Daitron product. We've covered it in a video before. If you are interested, please take a look at the link in the description! I'll turn on the power.
That's great. The conducted noise itself is at a fairly low level. This result shows very good noise performance. There is a good reason for that: the internal circuitry of this power supply is designed for soft switching. The noise emitted by the power supply itself is therefore low, resulting in less conducted noise.
Today, starting with how much conducted noise a ferrite core can attenuate, I tested how much conducted noise various filters can reduce. For the power supply tested this time, we found that the ferrite core on the power cord could not attenuate the conducted noise very much.
Retrofitting a filter circuit as we did this time is a way of dealing with noise after it is generated. It does not reduce the noise emitted by the noise source itself. If you are serious about dealing with noise, you need to design the system to reduce the noise emitted by the source itself.
If you found this video helpful, please give it a like and subscribe to the channel! If you have any concerns, please leave a comment as well. Thank you for watching the video until the end.
Article published · Updated