Electronic – Does Shielding “Electrically” shield “Magnetically” too

First: Don't try to solve a problem that is not there to begin with :) I know this is obvious, but attempted solutions to imagined problems often cause other problems.

Second: If you are worried about EMI noise visible on the low light images, I believe the problem should be addressed at the CCD chip design level. Maybe outside the scope of your question.

With that said, you have a standard EMI question and I can provide my standard EMI advice that have worked for many many products:

• A Faraday cage (metal box) is almost always required if you have multiple boards or things sticking out from one board. In your case we haven't seen the camera, but it's likely that you need a Faraday cage. You may have a problem with the opening for the lens, depending on how that is engineered. A Faraday cage does not need to be solid, but can be a net. Do some testing to see what is sufficient if you want to optimize this.

• All wires should preferably go in and out in one side of the box and preferably the same side where you make the connection between the internal ground and the Faraday cage.

• All shielded cables exiting the Faraday cage should have the shield connected to the Faraday cage (look at the shield as an extension of the Faraday cage). Notice that getting the parasitic inductance really low is essential for high frequency operation. This is a good option for your Ethernet connection.

• All unshielded cables exiting the Faraday cage should have a low pass, low inductance filter between each wire and the Faraday cage where the cable exits the Faraday cage. This would work well for your DC-connection.

With the information given, this would be my best advice for you. Feel free to ask more questions.

When you say signal is tied to the shield, which pins? The ground pins are tied to the shield. See pinout below

I am surprised to see that the grounds of the vga cable are tied to the shield, on the cables I have some of the grounds are tied to the shield. This is a more noisy solution, but with the cable making industry its about what works not what is best. They may be cutting corners by eliminating some conductors in the cable and running the ground through the shield.

A shielding conductor is used to prevent electrical fields to capacitively couple to the conductor they are protecting.

As far as turning the cable into an antenna:

From the cable shielding wiki:

The best method to wire shielded cables for screening is to ground the shield at ONE end of the cable

This is to avoid ground loops, ground loops are a big problem because current flowing through the cable causes current through the shield conductor to couple to the signals inside the cable magnetically. So to prevent this current flowing on the outside of the cable a choke can be placed on the outside of the shield to increase the high frequency impedance of the cable and cause high frequency signals to find a different "pathway" (as high frequency currents will "prefer" the path of lowest impedance/inductance).

The cable is shield is probably not tied to ground on both ends (computer and monitor sides), and one would hope that it is isolated somehow. Otherwise DC and AC currents will flow through the cable and cause problems.

So the shield shouldn't be tied to ground, it is probably introducing noise but it is and it still works, as far as we can tell. It would be interesting to find the VGA specs and see if this is a standard practice.