Well the test rig was built, the picture shows the second set up, the first was the servo connected to the prop, being 1.5mm approx. (my vernier is a manual and made by draper, not bad for the price but you get what you pay for), the flight linkage connector is 2mm i/dia, so it has some play, tried to rough it up a little for the solder, when soldered the connector to the shaft it took a little bit of aligning but seemed reasonable.
I had built the propshaft tube last night using Araldite, so i connected the propshaft to the connector using Araldite, it was out of shape by a small amount, not a problem for the water test, went ok and moved water, but after assessing how much i do not think the boat would move any faster than 10 knots at scale speed - ok in harbour not very good for a simulated attack run at 30-45 knots or even 20 knots cruise speed.
After the test, the solder connection came loose and it would require a serious amount of roughing up to make it work, one of those sleeve connectors prieviously mentioned would be better.
The next test was one of the small cheap motors from maplins, I wired the suppression parts and the servo controller to the motor it ran a lot faster with the forward and revers facility as expected, then i built it into the test rig, note i call it a test rig - bit heath robinson don't you know.
This when tested in water did somewhere closer to the cruising speed, so i am happy with that, now the problem (there always is), malcolmfrary is quite right, the second picture shows a close up of the servo controller, the right han side of the controller shows two items a blue balloon type item and a black block item (capacitors ??? - yep clueless), this black item gets hot after test running the set up in the bathroom sink for less than a minute, concerns are that it will generate too much heat in a confined space and might fry the controller in less than the 10 minutes overall time i am looking for (5 minutes realistically).
I have an ACTion speed controller, small black box, P52 on the circuit board, will try this if anyone confirms my suspicions.