|
Post by gregggagliardi on Nov 19, 2013 13:02:42 GMT -8
The attached research paper at the link below casts doubt on some previously hallowed ideas about multipoint anchors, particularly the effect of the angles between anchor legs on anchor and leg strength and the importance of leg length over leg angle. The paper is quite technical, certainly beyond my limited education in math and physics. Nonetheless, the research appears to be rigorous and conducted by folks who know how to do good research of this type. The conclusions are clear and fairly easy to understand. Whoever writes the section on multi-point anchors in Freedom 9 ought to consult this paper. In the meantime, I think that you will be surprised by what you read. www.caves.org/section/vertical/nh/51/Multi-point%20pre-equal%20anchors.pdf
|
|
|
Post by dougsanders on Nov 20, 2013 7:51:31 GMT -8
This study has been out for a while. Believe it was originally presented at ITARS and there was some funding from the MRA. In a nutshell, the shorter the anchor leg the higher the force.
Regrettably, I don't think anyone has tested adjusting leg length to compensate for this. Meaning, in a multi-leg, distributed anchor the short leg would have some slack compared to the longest leg. When loaded the forces would then be better shared among the legs.
The rope rescue community has recognized for a long time (circa 1990) the fallacy of equalizing anchors and known of the huge variation in forces within distributed anchor legs. There has been quite a bit of testing with dynamometers. As their systems use low-stretch cordage and greater mass, the anchors take more force that what we see in climbing. Additionally, rescue anchor legs can be quite long. The longest leg I can remember building was 60 meters! Given 2% elongation with low stretch ropes, one can CRUDELY calculate anchor leg length adjustment which we do.
I have tried to adopt this strategy to climbing anchors with long legs, but at best, it is a guess. Better than nothing but who know how much better? The study reinforces why our anchor pro needs to be solid; pro redundant; nylon used rather than aramid cords and slings; built in energy (force) absorption (harness rather than anchor belay device); and, designed toward legs of similar length. Perhaps someday we will have load limiters (screamers) in short legs or at the device.
On the otherhand, even when aramid cordellettes were used with long legs, one just didn't hear of anchor failures. Our alpine belay systems seem to have many ways to absorb energy.
|
|
|
Post by gregggagliardi on Nov 20, 2013 12:13:06 GMT -8
Thanks for the additional history, Doug. I believe the cited paper did use ordinary 7mm perlon cord. What I found particularly fascinating was how the angles between legs tended to decrease when the cordelette was loaded and how radical angles between the legs seemed to have such relatively small effects on the force on the legs.. The authors recommend extending pro placements with static slings to keep the leg lengths on the cordelette as nearly equal as possible.
The tests that I am looking for would examine the forces exerted on each component of the system during real world high factor falls. Imagine a drop tower with a 178 lb test dummy in a harness, tied in with a rewoven figure eight who falls 10 feet on 5 feet of rope. Load cells are attached between the tie in knot and the harness, and at each junction (leg) in the multipoint anchor, including the master point. Now add an experienced belayer, tied to the anchor in the usual manner, using a typical belay device.
What forces are generated at each junction in the system? Now change various parameters of the anchor set-ups, and use an autolocking belay devices for some of the tests. Change out the cordlette using different kinds and diameters of cord. How does this affect system performance?
|
|
|
Post by Deling Ren on Feb 11, 2014 17:45:00 GMT -8
I have read through the paper. The math and physics aren't that complicated. The most advanced is probably understanding a non-linear spring (second order polynomial). The amount of data points can be a little tedious to anyone without statistics background (which I don't). The data verified the models. I knew increasing the angle doesn't have a huge impact but was still a little surprised how little it is. But if you take a close look, it all makes sense: at 60 degrees, the outer legs contribute cos 30, which is 0.866; at 90 degrees, it decreases to cos 45 = 0.707. They are not that significant since the middle leg takes the most load. Rope and knot stretch further decreases the significance.
|
|
|
Post by gregggagliardi on Feb 11, 2014 18:46:53 GMT -8
Discussion of climbing methods greatly benefits from empirical research. The days of armchair speculation about what X might or might not do, based on ad hominem arguments and appeals to received wisdom are clearly over. ;>)
|
|