Dyneema Abrasion

These tests were motivated by two incidents. In the first, I had to pull a double strand of 1/8” Amsteel after rapping down an overhanging sandstone face of “The Owl” in Red Rock (https://hwstock.org/owl/). At the very top, the Amsteel ran over some sharp projections of sandstone. There was no problem until I was down and tried to pull the cord, and it stuck firmly. I put my body weight on the cord, bounced, and gradually it freed; but when I examined it later, I found that one of 12 strands was cleanly cut. I’m guessing one strand caught on a projection, and was severed by the knife-like quartz. In the second, I had placed a 4000 lb, eye-to-eye Amsteel sling around a rock horn, and held the eyes together with a quicklink. The frictional rock made it very difficult to slide the sling around the horn. Yet every person who rapped off that section insisted on yanking the sling back and forth, apparently afraid that if the forces were not “equalized” on each eye, there would be some catastrophic result. All this fruitless tugging abraded the sling. So I wanted to quantify how strong Amsteel remains—especially compared to other materials—with the inevitable abrasion. While we thing of Dyneema as being abrasion-resistant, equivalent-strength cords can be much thinner than PET (polyester), and the manner of weaving Dyneema into ropes can make it likely to snag.

I. Extreme Tests of Abrasion: Dyneema 12-strand (3.2 mm) vs. 6.5 mm PET Kernmantle

This very aggressive test shows a weakness of 12-strand braided Dyneema (Amsteel blue 1/8”), versus kernmantle construction PET cords. The setup was as shown in this video (boy I look like an old bald man):

link to test video

A very coarse part-round wood file/rasp is affixed to the top of a pullup bar, and the test section is pulled over it against a 10 lb dumbbell 6x, back and forth. I tried to be consistent with the length of each stroke for all samples. Here is the file:

The fiile was cleaned with a wire brush between tests, and the PET and Amsteel samples were alternated.

Below are the samples after the tests.

The previous photo makes the Dyneema appear better than it was, because…

The trouble with braids

In a 12-strand hollow braid, any strand that is “inside” is “outside" within a few diameters of the cord, and each strand winds completely around the rope within ~16 diameters (depending on the lay angle). Thus a pull that seems to abrade only one side of the rope, really compromises all strands.

In the setup above, kernmantle ropes abrade quite differently. The 90 degree intersections between yarns in the sheath quickly catch and rip on projections of the file, and the rope gets desheathed, exposing the core. But the parallel strands of the core have no such intersections to catch, and further pulls back and forth fuzz the core strands somewhat; but the core strands then erode slowly.

This plot shows the results of subsequent pull tests:

Comments: The kernmantle cord did “better,” essentially because in the two samples that desheathed, the core fibers, which were nearly parallel, just pulled back and forth between the file projections. I have no idea why one section of kernmantle was relatively less affected.

II. Amsteel Pulled over a Rough Brick for Different Lengths of Time

This test was conducted similarly to the previous, but with just Amsteel sections pulled back and forth 0, 4, 8, 16, and 32 times over a sawn rough brick (40mmx48mm). This brick is much more “rock-like” than the coarse wood files used above. As before, the brick was Gorilla-taped to the top of my pullup bar.

The brick is mostly made of 1-10 mm clasts that easily scratch glass, and by appearance seem to be pieces of ground igneous or metamorphic rocks. A small amount of mortar, probably dyed with iron oxide-hydroxide, holds the clasts together:

The brick was scrubbed clean with a fine wire brush in between tests.

The initial pull over the brick requires a force of about 50 lbs (measured with linescale 3), and subsequent pulls, each 1 second long with and about 3’ throw, required about 40 lbs. In between forward pulls, the cord is allowed to fall back over the brick under the weight the dumbbell (10 lbs) for 3’.

The results are shown below, with subsequent discussion to explain the differences.

IIA: Scoping Test

All pieces were pulled from scrap Amsteel I had saved from decommissioned ropes, and the 4-pull sections showed the most fuzzy-strand wear before the test. One of the 4-pull tests was caught in a crack for 2 pulls and showed unexpected wear (the cracks was sealed before the second set of tests.) Both the 32- and 16-pull sections were quite hot after the test, and both showing signs of glazing (partial melting). Notably the Samthane coating (the carrier for the dye, and said to increase wear-resistance) had partly rubbed off all these samples before the tests. I have found that the Amsteel becomes more slippery as the Samthane wears off. All pieces were terminated with bury eye splices, sewn loosely with 100 lb braided spectra fishing line (~20 stitches). After the pulls, each section was pulled to breaking behind my Jeep, and the force to breaking was measured and recorded on a Linescale 3 sampling 640 times per second.

IIB: Tests with New Amsteel

These tests used new black 1/8” Amsteel purchased from SBG Knots; the cord came in a plastic bag with a Samson label and literature. The cord was cut into 7.5’ sections which were then fashioned to have a class 2 bury eye splice on each end, as described above. The first series of 0, 4, 8, 16, and 32 pulls was completed, then the 2^nd set of 0, 4, 8, 16, and 32 pulls was performed. All samples were then pulled to breaking as described above.

Comments: The most obvious conclusion is that: over a rock-like substrate, Amsteel can retain much of its strength, until the continued rubbing/abrasion brings it near the melting point. The number of samples is too small for a definitive conclusion about the role of Samthane; but I suspect that Samthane increases the coefficient of friction enough, that fewer pulls are needed to bring the Amsteel near melting. The two 8-pull samples broke at fairly different forces; not surprisingly, the lower-breaking sample was visually more abraded. As seen in the plot below, the first sample had two separate “snapping” episodes.