1
\$\begingroup\$

We would like to send I2C and 12v PWM for a DC motor over the same CAT6a S/UTP cable. I know this is not ideal, and "differential I2C" would be better, but for now standard I2C is what we have to work with. This answer says:

"Using [i2c over] CAT5 should be fine, as long as you select the pairing properly. At least never pair up clock and data. For I2C it means you should pair clock and ground, and supply and data. Since you have extra pairs, maybe twist clock with ground, data with ground, and then the remaining could be two supply-ground pairs. I don't recommend having two supply wires as a pair"

This makes sense for the pairing the I2C signals, and we are using an LTC4311 to help with I2C rise time over the distance. Therefore, two pairs in the CAT6a cable are available for 12v PWM to drive a standard brushed DC motor. The current will be <500mA, so 23 AWG should be fine. My question below is ultimately about how the PWM lines should be paired in the cable to minimize PWM<=>I2C crosstalk.

If we take a cross-section of a S/UTP CAT6 cable, it separates the pairs with a solid membrane:

cat6 cross-section

For the purpose of posing my question, we label the cross-sectional areas as ABC and D:

 A | B
---+---
 C | D

Question:

  1. How should they PWM motor lines be paired to minimize interference with the i2c line? Since two-pairs are available for PWM, there are a few options:

    • Twist the +/- PWM lines into the same pair, eg: orange=PWM+12v, orange-stripe=PWM-GND
    • Twist the +/- PWM lines into two pairs, eg: orange/green=PWM+12v, orange-stripe/green-stripe=PWM-GND.
    • Dedicate one pair to each PWM conductor to treat the pairs as separate conductors, eg: orange/orange-stripe=PWM+12v, green/green-stripe=PWM-GND)
    • others?

  2. In which quadrant should each pair be located (PWM, PWM, SCL/GND, SDA/VDD)?

    • Corner-to-corner: PWM=A,D and I2C=C,B
    • or side-to-side: PWM=A,B and I2C=C,D
    • others?

Other ideas are welcome, but keep in mind that ultimately the question is about PWM wire pairing and pair placement.

Thanks!

\$\endgroup\$
5
  • 2
    \$\begingroup\$ if you can, definitely try to get FTP, not UTP - no matter how you twist the pairs, you're going to get some coupling between the PWM and the I2C, and that's very bad - not only is it PWM, it's PWM for a motor, one of the noisiest things possible. 12V is a pretty huge voltage swing as well, so it wont take much coupling to cause problems. \$\endgroup\$
    – BeB00
    Commented Sep 8, 2023 at 23:11
  • 2
    \$\begingroup\$ have a look at PoE (Power over Ethernet) ... supply the 12 V as DC ... send the PWM as a low level control signal \$\endgroup\$
    – jsotola
    Commented Sep 8, 2023 at 23:34
  • 1
    \$\begingroup\$ Doesn't directly answer the question, but if you can use two different cables. \$\endgroup\$
    – user1816847
    Commented Sep 9, 2023 at 0:00
  • 2
    \$\begingroup\$ Note that the linked answer makes no consideration of EMC, particularly immunity. Putting a source of switching noise within the cable itself is a good way to guarantee nothing ever communicates over that poor I2C bus. Pairings are irrelevant. Use differential, period. \$\endgroup\$
    – Tim Williams
    Commented Sep 9, 2023 at 5:01
  • \$\begingroup\$ @TimWilliams, thanks. We're switching to 12v differential using a PCA9600 even though it is a modification we were trying to avoid... So should I delete this question or leave it here for others who might try something as crazy, so they can at least read the comments? If you or anyone would like to write a good answer for posterity on why this is just a terrible idea, then I'll mark it as accepted. \$\endgroup\$
    – KJ7LNW
    Commented Sep 10, 2023 at 0:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.