Four slivers of magnet (bits of speaker magnet work well; not Neodymium!), two on the body, two on the ends of the wires from the heater.
Test the pull-away forces with small weights. If the pull-away force is too high, either reduce the cross-section of the slivers; or place one or more pieces of aluminium foil between each pair to reduce it.
Use a little Jena (http://www.jenalabs.com/contact-fluid/contact-fluid.html) or similar to ensure good electrical contact.
It is possible to get the pull-away forces down to a few 10s of grams whilst retaining good vibration resistance of the connection.
(Still not sure why the launch rod has to be titanium -- or any metal? A carbon fibre rod would be lighter, stiffer and less prone to icing.)