Composite forestay attachment point.

[DGreenwood]

Originally Posted by John Perry

A few years ago we made some test pieces to represent chain plates made in this way. The test pieces were made by wrapping unidirectional carbon pre-preg round and round a former made from MDF covered in parcel tape then heat curing. The ends of the former were full radiused so we ended up with what looked like oval chain links but square section. We trimmed the rough edges then tested them to destruction by pulling them appart in a frame to which was attached a 200 tonne nominal capacity hydraulic jack. The attachment to each end of each link was made through a close fitting pin machined from Nitronic 50. I wish I could remember the exact results of the tests but I cannot and the firm I was working for at the time has now gone bankrupt. However, I can tell you that the breaking strength of the links was less than you would predict from the cross section of carbon fibre in the two straight sides of the oval. It seemed from the way that failure occured, that load was not equally shared between the fibres around the ends of the links, so these started to fail progressively with a series of loud cracks. Failure was always at the rounded ends, not the straight sides of the oval. The breaking load was certainly less than a simple calculation for the straight sides would indicate, unfortunately I cant remember how much less, I am not a lot of help am I!

I am certainly not saying that you should not make chainplates this way, quite a few builders of composite hulls are now doing it. But I would say that you should allow for the above effect by picking a reasonably generous factor of safety/ stress concentration factor (whatever you want to call it).

When a typical stainless steel chainplate is loaded close to breaking point there is also uneven stress distribution arround the holes used for fasteners and clevis pin, however SS316 is lovely and soft (relatively!) and yields to mitigate the stress concentration at least a little bit. You need to allow for a stress concentration factor round the holes of a stainless steel chain plate (the ESDU engineering data books are one source of informaton on this) but I suspect that you may need a rather larger factor for a unidirectional carbon construction.

The failure described here can be mitigated by distributing the fibers across a broader thimble or using two broad thimbles with a gap between them and a pin spanning them. Stacking fibers on a small radius curve is to be avoided. What your proposing is common stuff now and most designers that use carbon could give you guidance that would make a safe attachment point.