We'll be honest, Treetools cannot fully explain how the physics worked in the climbing configuration test where the DMM Sentinel screw gate carabiner failed at 21kN.
But we do know… on paper at least, the DMM carabiner is technically the weak link in the chain. It has a strength of 24kN or approximately 2400kg.
Donahys 8mm Armor-Prus has a tensile strength of 2800kg and there were two legs to the French prussic (thereby doubling its strength at the bar). The tying of the friction hitch could possibly lesson its strength to some degree and the eye splice on each leg may also have some weakening effect on hitch cord strength.
Regardless of the technical specifications of each element in the configuration, once the load went on, both connection points on the carabiner bar (rope and hitch cord) splayed in opposite directions to the lateral pull, side loading the carabiner and producing the break.
Treetools has conducted over 50 break tests including climbing configurations, splices and knots. The tests are carried out using a heavy duty hydraulic ram to pull the items apart until they break. The ram is calibrated and certified and all load information generated throughout the test is recorded on computer.
However, this type of static test does not accurately replicate how climbing equipment is likely to fail in the field, where the load force is dynamic, but it is considered an appropriate industry standard for testing.
Throughout all the static break tests conducted by Treetools the friction hitch, regardless of style, starts to slip somewhere between 2.5 and 5kN. To complete the test we often have to stop the hydraulic ram to reset the hitch or place a stopper knot below it. Without the stopper knot the hitch will often continue to slip thereby making the 'break' impossible.
Real world use of friction hitches and knots, where load stress is dynamic (say in the event of a fall due to slack in the climbing line), the friction hitch is likely to slip on the climbing line (as it does in the static tests) thereby reducing the full force on hardware and other components (except possibly the anchor depending on configuration).
If anyone can provide us with a relatively simple explanation of the physics involved in the graphic below please be our guest.
The Yale Arrow Frog and Donahys Armor-Prus kN ratings stated in the graphic above are an approximation only, to demonstrate relative component strengths on paper.