SN65HVD75: noise suppression component for differential lines

The ferrite bead should have a low impedance at the signal frequencies (for square waves, this is up to about ten times the bit rate), and a high impedance for the noise frequencies. (The datasheet of the bead should have a graph for the impedance vs. frequency characteristics.)

The high impedance should be much larger than the termination resistor, because the two form a voltage divider. If you combine a 120 Ω bead with a 120 Ω termination resistor, then a 100 MHz noise impulse will still drop half its voltage over the receiver pins.

If there are signals and noise at the same frequencies, you have to use common-mode chokes.

Hello Inoue-san,

To add on to the above comments from Clemens, how the bus is terminated can also impact common mode noise immunity.  In general, use of differential termination on a bus tends to improve noise immunity compared to an unterminated bus, since the low differential-mode impedance helps to keep the noise coupling equal between the inverting and non-inverting halves of the differential pair.  (This ensures that the noise is primarily "common mode" in nature, which allows it to be more easily rejected by a differential receiver.)  For a terminated bus, further improvements can be achieved by using "split" termination in which two equal resistors are placed such that their central node can be filtered via a capacitance to ground.  You can read more about this technique here:

Beyond this, common mode chokes could be used (as mentioned above) if the noise frequency is in the same range as the data.  These are coupled inductors that are specially wound to ensure that they present a high impedance to "common mode" noise while a low impedance to "differential mode" signaling.  If there is good separation between signal and noise frequencies, simpler filtering techniques are possible.  Series ferrite beads with high impedance at the noise frequency are one common technique, and you could also consider shunt capacitors on each line to serve as a filter for higher-frequency noise.  (Note that capacitance should not be too large as to significantly affect the timing of the RS-485 data.)

I hope this is clear, and please let me know if you have further questions.

Regards,
Max

Hi Clemens,

Thank you for your reply.

The ferrite besd is considered with low impedance at the signal frequencies and a high impedance for the noise frequencies

I'll share your comment with the customer.

I appreciate your great help.

Best regards,

Shinichi

Hi Max,

Thank you for your reply.

I'll send the customer your helpful information.

I appreciate your great help and cooperation.

Best regards,

Shinichi