Thanks very much for all your responses and iI am particularly pleased so many are still about 3 years after this thread started. I will respond in order of suggestions.
Guy the pressure switch is a possibility I should check though not sure how and is a bit coincidental with two pumps in a row . I recall
trying to adjust the previous switch pressure switch but could not improve it.
Dlymn there is no accumulator though am considering installing one in desperation though water flow is good and pressure variation hasn't been a big problem. There was an accumulator on a 2009 Seawind 1170 I sailed on in the Whitsundays last month but don't think they were standard on earlier Seawinds and we have only started having this problem in the last year.
Yes I will look for air in the lines again and I did find 20 pieces of white 2 to 3 mm flat (like chips of paint) particles in the suction side filter which I cleaned and now replaced to eliminate as a possibility. Also the pump will groan on all night if I let it which might exclude the flexing pipes.
The check valve is not obvious to me on the Shurflo pump and assume it is part of the pressure switch but again bit coincidental with two pumps in a row.
Yes i can see all the joins back to the water tank and feel the positive head pressure when tank is full could eliminate them but will check again.
HopCar the fact the pump works perfectly when drawing from a bucket has to be an indicator to the problem but I just can't see how though yes, would have to be from pump back to tank. I pumped up pressure with a plumber pressure test pump on the pump delivery
side but not with the pump connected or through the pump but can't see good reason for extending this test to include them. And the tank breather would reduce water flow?...which isn't a problem
Sylfur there is a hot water system we replaced 3 years back but the draw from bucket test would not work if the HWS was a problem. I guess it could be air ingress on the suction side though of low probability with the positive head pressure.
Bruce K the bucket test probably rules out pump terminal issues as far as I can understand.
So I will try and pressure test the suction line but not sure how I can without missing out the final join to the tank or cutting the line and adding adding another T piece I would have to cap off when finished, leaving another potential leak point.
Thanks again for your efforts and let me know any other suggestions.