When you add a ferrite bead to a CAN or LIN bus line, it is not there to filter the signal. Its real job is to suppress common-mode noise and radiated emissions. CAN_H and CAN_L form a differential pair. In an ideal world, the common-mode currents on the two lines are equal in magnitude and flow in the same direction, and the differential receiver ignores them. But in a vehicle, those common-mode currents turn the cable into an antenna and radiate in the 30 MHz to 100 MHz range and above, or couple external interference back into the board. A ferrite bead at the connector port adds an impedance that rises with frequency to this common-mode path, blocking high-frequency common-mode current inside the board while leaving the low-frequency differential data almost untouched.
Impedance and tolerance
Because of that role, the first parameter for a CAN bus ferrite bead is its impedance at a specified frequency, typically 100 MHz. For passenger-car CAN, around 600 Ω is a common recommendation, usually in an 0603 or 0805 package. Classic CAN at 500 kbps has its relevant frequency band in the tens of megahertz, and 600 Ω at 100 MHz already provides enough impedance there. Pushing to 1 kΩ or more brings limited benefit and increases DC resistance and parasitic capacitance due to more turns. Impedance tolerance is typically ±25%. Batch-to-batch consistency matters more than the absolute impedance value, because the two lines must be strictly symmetrical.
Current rating and DC resistance
The second parameter is current rating and DC resistance (DCR). Many engineers look only at impedance and ignore DCR, which is a common mistake. A CAN transceiver has limited common-mode drive capability, and the bus common-mode voltage must stay within a window of about -2 V to +7 V. Placing a bead with high DCR on each line adds a significant DC drop to the common-mode loop. Under bus short-circuit or common-mode offset conditions, the transceiver can be pushed outside its normal common-mode range and communication fails. Therefore, CAN ferrite beads usually require DCR in the range of 0.1 Ω to 0.5 Ω and a current rating of at least 200 mA, preferably with more than double margin. Also check the DC bias curve in the datasheet: ferrite permeability drops under DC bias, so a device rated 600 Ω may retain only half that impedance at a few hundred milliamps of bias. This derating must be verified in advance.
Self-resonant frequency and parasitic capacitance
The third often-overlooked parameter is self-resonant frequency (SRF) and parasitic capacitance. A ferrite bead is essentially an inductor with loss. Above its self-resonant point it becomes capacitive and its impedance falls with frequency. If the SRF falls in the low hundreds of megahertz and the vehicle's radiated emissions test focuses on that band, the bead fails exactly where suppression is needed most. Specify an SRF at least two to three times higher than the frequency band of concern, and keep parasitic capacitance as low as possible, typically below 1 pF.
Automotive grade and matching
The fourth point is qualification grade. Automotive parts must meet AEC-Q200, cover an operating temperature range of -40°C to 125°C or even 150°C, and have a clear temperature derating curve. The two beads used for CAN_H and CAN_L on the same board must be the same part number, same batch, and same package. Never use an 0603 on one line and an 0805 on the other. Otherwise the impedance of the two lines is asymmetric, and differential-mode signal is converted into common-mode, creating a new radiation source. This is the so-called mode conversion problem.
When a common-mode choke is better
If the vehicle's radiated emissions or BCI immunity requirements are stringent, two ferrite beads are often not enough. A more common engineering approach is to upgrade to a common-mode choke: two windings on the same core, where the magnetic flux from differential currents cancels, making it almost transparent to the differential signal, while common-mode impedance can reach several hundred ohms or even over a thousand ohms at tens of megahertz. Combined with a device that integrates TVS, one component can handle both common-mode suppression and ESD protection. The trade-off is size, cost, and the difficulty of controlling differential-mode parasitic parameters. This is acceptable for low-speed CAN, but for CAN FD, carefully evaluate the effect of the choke's differential inductance on rise times.
CAN FD considerations
CAN FD deserves a separate note. With data rates of 2 Mbps or even 5 Mbps, the high-frequency content of the signal shifts much higher. The bead's parasitic capacitance and series impedance begin to form a low-pass network with the bus capacitance and termination resistor, slowing rise times and visibly narrowing the eye diagram. In this case, push the bead's impedance peak to a higher frequency, choose a part with lower parasitic capacitance, or abandon the bead approach entirely and use a common-mode choke designed for CAN FD. The 120 Ω termination resistor and split capacitor are mandatory; a ferrite bead can never replace termination.
LIN bus: a different set of rules
LIN bus follows a different logic. It is a single-wire, 12 V, slow bus at up to 20 kbps, using a pull-up resistor of about 1 kΩ and the transceiver's internal low-side driver to generate the dominant level. Nodes have limited tolerance for bus capacitance, and the specification sets an upper limit per node. This means that adding a ferrite bead on LIN requires extra care in two areas. First, DC drop: the recessive level is pulled up to battery voltage, and the bead's DCR directly eats into the level margin. The current rating must cover the pull-up current plus worst-case load current, and DCR should be below 0.1 Ω. Second, capacitive load: the bead's parasitic capacitance plus trace capacitance slows edges that the transceiver has already deliberately slowed, affecting the sampling point. Therefore, LIN typically uses small-size beads of 100 Ω to 600 Ω at 100 MHz in 0805 or 1206 packages. The priority is not high impedance but clean DC characteristics and low capacitance.
From the perspective of what is being suppressed, LIN interference is more concentrated below a few megahertz and in ESD events. A ferrite bead has almost no effect at low frequencies, so expecting it to solve low-frequency common-mode problems is a wrong approach. Suppression in that range relies on capacitance to ground, TVS, and proper harness routing. The bead only blocks high-frequency residue at the port.
Layout and verification
Finally, layout and verification. Place the bead as close as possible to the connector, keep traces short, make the return paths to ground symmetrical for both lines, and do not put slots in the ground plane under the bead or cross a split. Otherwise, the loop formed by the detouring common-mode current becomes an antenna itself. For verification, do not rely only on the impedance table. Check radiated emissions per CISPR 25, BCI bulk current injection per ISO 11452-4, and actual eye diagrams and bit error rates together. It is common to see a bead change that reduces emissions by 3 dB but cuts eye margin by more than ten percent. At that point, re-evaluate whether to switch to a common-mode choke, or add nothing at all and solve the problem with termination and shielding.
What this means for your build
- Select CAN ferrite beads by impedance at 100 MHz (typically 600 Ω), but verify DCR (0.1–0.5 Ω), current rating (≥200 mA), and DC bias derating.
- Ensure the two beads on CAN_H and CAN_L are identical in part number, batch, and package to avoid mode conversion.
- For CAN FD above 2 Mbps, consider common-mode chokes or low-capacitance beads, and always keep the 120 Ω termination.
- On LIN, prioritize low DCR (<0.1 Ω) and low parasitic capacitance over high impedance; use beads only for high-frequency residue.
- Place beads close to connectors with symmetrical ground returns, and validate with CISPR 25, ISO 11452-4, and eye diagrams.
Source: compiled from public web sources.