I have an optically-couple "supervisory port" on my
> devices. A design requirement is that the outer surfaces
> of the devices must be "wipe clean" -- no ridges or
> grooves.
>
> I mechanically attach the mating "cable" to the port
> magnetically (an idea I stole from designing electric
> power meters -- but, they aren't "wipe clean"!)
>
> As I can't rely on any significant mechanical structures
> to help maintain the attachment's position, that must
> be guaranteed solely by the strength of the magnetic
> bond.
>
> I have a 1/4" dia glass "jewel" that is just barely convex
> to protect the optical port and support light passage.
> The slight bump helps locate the port as well as some small
> assistance in keeping the mated "cable" positioned properly.
>
> In no case can the needs of the port exceed a concentric
> diameter of 1/2". I plan for the magnet to only be present
> in the mating cable so its adhesion will be determined by
> the metallic ring surrounding the jewel.
>
> I figure I need to specify a magnet strong enough to
> "support" (against the force of gravity exerted on the
> mass of the cable) the cable in a horizontal position
> (i.e., "mated to the ceiling"). And, to support the
> cantilevered weight of the cable when mounted
> vertically.
>
> I may opt to implement the "cable" as a wireless dongle
> if the weight of the cable assembly starts to compromise that
> magnetic adherence.
>
> I guess I also need to consider the magnetic force exerted
> through the friction between the mating surfaces when
> mounted vertically (so the cable doesn't "slide down the wall") >
> Anything else I should consider? I realize that the cable end
> now will have a tendency to want to grab onto metallic objects
> so have to factor that into my dongle vs. cable decision
> (it's a lot easier to control the position of a dongle in
> space than the end of a possibly *dangling* cable!)
>
> I suspect the adhering "metal" piece has to be on the skin
> of the device as the magnetic field falls off quickly with
> distance -- diluting the ability of the cable to remain
> EXACTLY where desired. A "decal" over the surface likely
> would be tolerable...
Would a rounded bump on the device be acceptable? A matching hollow in the cable end could then locate over it.
Something like this:
formatting link