I'm planning to have short (~1m) RS485 runs @<500k baud to transmit some data between PCBs. These PCBs also have a few addressable LEDs (WS2812B) that send data at 800kHz.

Would it be problematic to run the 800kHz signal in the same Cat5 cable as the RS485? (Using an unused twisted pair as Olin mentioned)

What precautions do I have to take? I could convert the 800kHz signal to differential with another RS485 driver as well but I'm not sure if this is overkill.

p.s.: The reason I'm using RS485 is that one of these boards will have a long cable and I'm using the same hardware for all of them.

  • \$\begingroup\$ I was with you thru the first paragraph. The second makes no sense. What "either" signals? What Ethernet cable? Nothing of Ethernet was previously mentioned. Are you asking about using unused twisted pair? \$\endgroup\$ Jan 27 '17 at 18:22
  • \$\begingroup\$ @OlinLathrop -- By Ethernet I meant Cat5, I corrected that. I am asking about using an unused twisted pair, and if this would probably be detrimental to the RS485 signal and/or the 800kHz signal at short runs. \$\endgroup\$
    – Wesley Lee
    Jan 27 '17 at 18:25
  • \$\begingroup\$ Crosstalk is possible with an unbalanced pair adjacent to a balanced pair, sharing of grounds and LED ground shift may also affect results. Bench Test simulation is advised since we dont have details. \$\endgroup\$ Jan 27 '17 at 18:27

What you suggest will probably be OK. CAT5 cable contains four twisted pair, each with a different twist pitch. This causes almost all coupling from one pair to another to be common mode. It's really hard to inject a signal into a CAT5 cable that couples in differential mode to a different twisted pair.

RS-485 is a differential signal going from up to +5 V to -5 V when transmitted. Receivers have a couple volts or so hysteresis typically. It would take a lot to exceed the common mode range, or introduce enough differential mode noise to prevent correct reception.

One thing to keep in mind is that while RS-485 is differential, these two lines still need a ground reference. At a minimum, this requires 3 wires in the cable. You probably want to dedicate a whole pair to ground in reality.

800 kHz is really not that high. Unless the voltage swings are very large, it's simply not going to cause problems for RS-485. Since 800 kHz is your highest frequency of interest, it might be a good idea to add a filter in front of the cable to attenuate any higher frequency components, which you don't want anyway.

If you can send the 800 kHz differentially using its own twisted pair, even better. Even just a balun right before injection into the cable will help, although I really think you'll be fine even if this is sent as signal and ground on its own twisted pair.

I really don't expect any problems with what you are propsing.


You should also use a LVDS or RS485 driver for the 800kHz signal, too.

The reason is not the other RS485 transmission line, because this symmetrical signals can't produce interferences (almost). You should be afraid of signals from outside the cable, which can easily disturb your 800kHz transmission. The connected LEDs will flicker in case of transmission errors, this should be avoided.

  • \$\begingroup\$ This would also avoid crosstalk between the signals to some extent, correct? \$\endgroup\$
    – Wesley Lee
    Jan 27 '17 at 18:31
  • \$\begingroup\$ Of course, cross-talk can be avoided by symmetrical signals. \$\endgroup\$
    – auoa
    Jan 27 '17 at 18:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.