2
\$\begingroup\$

I joined a project after some of the initial design decisions were made, so it’s not quite how I’d like it, which now presents me with a bit of a problem.

I have two power cables that are emitting noise. I’m going to work on a few solutions to get rid of the noise at source, but time restraints may mean I have to get a temporary solution in place first. One option is to change the cable to a shielded one, but I don’t have a chassis (at either end) to ‘ground’ to (enclosures at both ends are plastic).

One end is a battery box, the other a motor controller. I presume that as I can’t do 360° shielding, my next best bet is to connect the shield to the battery negative rail at only the battery box end.

Does anyone have any better solutions or am I going in the right direction?

\$\endgroup\$
14
  • 1
    \$\begingroup\$ What is the frequency (range) of the noise that you see? \$\endgroup\$ Feb 10 at 4:14
  • 1
    \$\begingroup\$ I doubt the shield connection at the battery box end will make much difference, but the (-) terminal is as good as anything would be. On the motor controller end, shield should be soldered to the PCB's clean ground plane. \$\endgroup\$ Feb 10 at 5:03
  • 1
    \$\begingroup\$ You're dealing with a problem generated by someone else - a patch-up job. Better to find the noise source and fix it there - likely at the motor-controller. \$\endgroup\$
    – glen_geek
    Feb 10 at 14:27
  • 1
    \$\begingroup\$ @Sylvester, Are you sure the noise is radiated? If the noise is conducted (and not radiated), the cable shield will have no effect. \$\endgroup\$
    – Vincent
    Feb 10 at 15:18
  • 1
    \$\begingroup\$ Note that in many weight sensitive applications like satellites, power cables are not shielded in order to save weight. The plus and minus wires are twisted. \$\endgroup\$
    – SteveSh
    Feb 11 at 0:54

1 Answer 1

0
\$\begingroup\$

The noise current in the cable needs to be returned back to the source. If you are only dealing with 30MHz plus, you may just be able to capacitively couple The shield back to the source voltage rail or ground.

Ideally you would probably want two couple this to PCB ground. You need a low impedance pathway back to ground so keep the inductance low by keeping the shield near the PCB.

You could also try and do some experiments with the near field probe and see if you can come back on some of the radiation by grounding it to different points.

\$\endgroup\$
3
  • \$\begingroup\$ All our problems are actually below 30MHz. Particularly in the 150KHz to 600KHz range. \$\endgroup\$
    – Sylvester
    Feb 11 at 4:33
  • \$\begingroup\$ Capacitive coupling should be enough, however, you will need to limit inductance \$\endgroup\$
    – Voltage Spike
    Feb 11 at 17:39
  • \$\begingroup\$ Yes, I did find some noise was better if I added caps over the power lines at the motor controller, but some worse. Once I get some shielded cables made up, I’ll add it to the list of options. \$\endgroup\$
    – Sylvester
    Feb 12 at 9:39

Your Answer

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

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