0
\$\begingroup\$

I'm trying to shield 6 flat cables that give me some communication issues because I thought that there might be cross-talk or other sources of interference that cause the problem. The shielding is not connected to ground. For now, it's just an improvised aluminium foil wrapped around each cable.

Leaving aside that it's not really a good shield, I was wondering if, theoretically speaking, this shielding might affect the signal integrity of digital and/or analogue signals. The curiosity comes from the fact that the shielding changed drastically the distribution of E field around the flat cable, so it's not the one the cable would normally generate anymore.

In other words, when the distribution of E and/or H field is changed by external factors, is the signal integrity affected? And how?

Set-up description

  • 6 flat cables (24 pins) are connected on one end to the same board at a distance of approximately 2 cm to each other, at the other end are connected to different boards at a distance of approximately 50cm to each other;
  • All flats carry the same signals, the highest frequency is 10 MHz clock with a rise time of 8-9ns, which is present at just one wire. All the other wires of the flat carry low frequency signals (below 500 kHz);
  • Flat length is 1.5m;
  • Two 150W switching regulators are present and at a distance of 30cm from the board where all flats are connected.
\$\endgroup\$
  • \$\begingroup\$ Frequency of signals? \$\endgroup\$ – Dirk Bruere Apr 27 at 11:23
  • \$\begingroup\$ In my case, there is a maximum frequency of 10 MHz. However, I'm also interested in some general considerations to better understand the situation. \$\endgroup\$ – Elia Apr 27 at 11:57
  • 2
    \$\begingroup\$ What is the distribution of signals in the cable? For 10 MHz, I would have every other wire be a ground. Electrostatic shielding (alum foil) won't do much for low impedance signals (which these probably are). Low impedance signals generate and are sensitive to magnetic fields. Having a ground for each signal will minimize the loop area and therefore the magnetic field. \$\endgroup\$ – Mattman944 Apr 27 at 13:15
  • \$\begingroup\$ what is the risetime? Is a black-brick power supply anhwhere near? \$\endgroup\$ – analogsystemsrf Apr 27 at 13:28
  • 1
    \$\begingroup\$ I recommend Henry Ott's book. The one I have has been superseded by another, but I am sure it is still as good. I attended Mr. Ott's week-long EMC course in the 80s. \$\endgroup\$ – Mattman944 Apr 27 at 15:16
1
\$\begingroup\$

You asked if shielding affects signal integrity: Yes, of course it does. Shielding changes the characteristic impedance of the transmission line model.

Concerning your ungrounded shield: It might reduce external magnetic field coupling (inductive coupling) a little, but it does certainly not reduce magnetic field coupling between the cable signal lines. And an ungrounded shield does not prevent electric field coupling (capacitive coupling), too.

So you might want to ground the shield. And you might also want to follow Mattman944's advice to use an adjacent ground return for each signal to minimize loop area.

\$\endgroup\$
  • \$\begingroup\$ In the past, when I had to just make something work, I found that running a ground for every signal, and twisting the ground with the signal, improved signal integrity a lot, even though all the grounds came from one pin at the source, and terminated at one pin at the destination. It was a mess, and not my preferred way of doing it, but it worked. \$\endgroup\$ – mkeith Apr 28 at 17:25
  • \$\begingroup\$ I see what you mean, but how can I ground the shield reliably (mechanically and electrically speaking) if it was not supposed to be there? As I said in my post, now the shield is just an aluminium foil wrapped around the flat. Is there a reliable way to "add" a grounded shield to a flat cable? \$\endgroup\$ – Elia Apr 28 at 18:24
  • \$\begingroup\$ I don't know. Maybe. You could try using copper tape. Stick it to the flat cable, and ground it somehow on the source side. It depends on whether you are looking for a volume production solution, or a few units, or a field retrofit, or what. If you are just experimenting, for sure you can figure out a way to do it. Use conductive copper tape, and maybe a foil crimp. You can also buy braided copper shield tubing and slip it over the whole cable. \$\endgroup\$ – mkeith Apr 29 at 19:44

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.