Break Test Extravaganza(s)

[allene]

So, how is it going?

[Brion Toss]

Hello all,
My apologies for neglecting to report on the break tests. No excuse; I've had the results, but just hadn't gotten around to organizing them. So, a brief, most belated summary:
First, rate of pull. In previous tests, at New England Ropes and elsewhere, the hydraulic cylinder was set to move at about 12"/minute, while the cylinder at TreeStuff, where the current tests were done, moved more than twice as fast, at 27.5"/minute. At the slower speed, the knot has a chance to "set" before a serious load comes on. At the higher speed, we wanted to see if the set of the knot and the rate of pull had an effect on the strength of the shackles; in the real world, people might not take the time to set shackles they build, and the load will tend to come on faster than in the lab. Sure enough, the average break strengths were lower than had been reported in previous tests. The "standard" Lanyard Knot configuration averaged 106% of rope strength, instead of at least 120%, for instance.
Interestingly, the results for the button-knot-and-bury shackle were not significantly different than the slow-rate tests; I am guessing that this is because all four parts of the button version draw up under load, while only two parts of the lanyard version do.
We can't do much about bringing the load on gradually, but it seems pretty clear that setting a new shackle is an important step, at least for the lanyard version.

Next, effects of diameter. In all of our tests so far, the bigger the rope, the lower the relative break load. For instance, with 5/32" rope it is not uncommon to get breaks of 280% with the button version, but by the time you get to 11mm results are down to about 160%. The only explanation I have is that it is much harder to produce a fair, snug knot in larger materials than in smaller. This scaling factor seems to explain the range in break strengths we seen from other people's tests. For instance, the estimable Evans Starzinger was getting, as I recall, about a 150% average for the lanyard version, while New England was showing more like 120%. With larger rope than Evans could break on his machine.
If further tests confirm this scaling effect, then we will need to scale shackle strength to rope size, instead of using a blanket average number for all sizes. Further, the scaling effect argues for the use of the inherently stronger button version, because with it we can use smaller rope, with a higher efficiency, for the same load. This effect also seems to undo, to some extent, the advice to use the lanyard knot version, but with larger rope, if greater strength is required, because greater rope size does not seem to result in correspondingly greater strength.
I am trying to avoid confirmation bias here (the button version is my baby), but must also note that smaller rope simplyfits into places that larger rope can't, which matters in things like connecting rope to chain for rodes. It is also easier to get an acceptable radius for the bearing point of smaller rope. So even though the button version is more difficult to make, it seems to be worth the effort for many applications.
Next, operator skill. For a given configuration, how well the knots were formed made a huge difference. One tyer, skilled at making the lanyard version, and with a 135% break average to prove it, got a less-than 100% average for his attempt at the button version, instead of the over 200% results that skilled tyers of that knot obtained. His buttons looked malformed to my eye, but not everyone knows how to judge that kind of thing. Therefore I urge people to get destruction tests done. Or, if they are buying soft shackles, ask to see test results from the vendor.
Finally, the +300% Grail. In no previous tests have we gotten over 300% efficiency for any shackle, but a Mr. Gardner came up with the idea of a"bifocal" version, in which two pieces of rope are used, instead of one. Sadly, his version averaged only 99%, but the concept was compelling, so I played around with it. One variation averaged 328%(!). I have some further samples ready to test, and will try not to delay in getting them out. Or in getting back to you re results.
If you submitted samples for the tests, please be in touch for detailed numbers on your work.
Again, I regret taking so long in getting these results out.
Fair leads,
Brion Toss

[allene]

Brion,

Thanks for filling us in. I would like to see some details. One thing that would be very interesting is the range of results for each type of knot. I assume you had a range of skill levels so this would give a good indication of what we could expect to see for a range of users. Averages are perhaps not as interesting as the minimum strength. The average doesn't mean a thing if the shackle you are using is at minimum strength. Personally, I would rather use a shackle that had an average strength of 110% and a minimum strength of 100% than one with a average strength of 200% and a minimum strength of 50%. This example is exaggerated and I am not suggesting it represents the results. Of course, if the shackles are inspectable such that any user could judge a good shackle from a bad one then bad ones could be excluded from the results. But lacking some usable criteria, I think they need to be included.

You say that going to larger line causes a loss in efficiency but is that loss enough to erase the increase in strength of the larger line? Going up in line size gives 50 to 60% increase in strength. Is the loss in efficiency enough to negate this increase in raw line strength? That would be surprising. Perhaps the combination in added strength and lost efficiency is not enough to equal the superior button knot strength. Does this hold for the minimum strengths as well as the average?

And finally, if it is possible to give each person the results of the units they submitted that would be nice. You have my permission to publish any of my results if that would help. To tell you the truth, it has been so long I can't remember what or even if I sent any samples.

Allen
L-36.com

[MarkG]

Brion,

This is some feedback on my failed (<100%) button knot attempts. I didn't understand the importance of setting the knot carefully. I pulled it as tight as I could by hand, thinking that it would tighten on the first use. It obviously didn't.

I set the diamond knots as hard as I could with a lever on the shackles that were good (135%). The one that wasn't so good (96%) was my first ever soft shackle and was not set at all.

Are there recommendations from experienced people about what tension should be used to set the knots? Has any testing been done with controlled, measured setting? I am wondering if this would make breaking strength more repeatable.

I also had an idea that might improve my original bifocal shackle. The free ends from the button knot could each have a McDonald brummel eye spllce. This would result in three eyes going around the shackle using a single line. Would this be a trifocal shackle

Mark Gardiner

[allene]

Originally Posted by MarkG

Brion,

This is some feedback on my failed (<100%) button knot attempts. I didn't understand the importance of setting the knot carefully. I pulled it as tight as I could by hand, thinking that it would tighten on the first use. It obviously didn't.

I set the diamond knots as hard as I could with a lever on the shackles that were good (135%). The one that wasn't so good (96%) was my first ever soft shackle and was not set at all.

Are there recommendations from experienced people about what tension should be used to set the knots? Has any testing been done with controlled, measured setting? I am wondering if this would make breaking strength more repeatable.

I also had an idea that might improve my original bifocal shackle. The free ends from the button knot could each have a McDonald brummel eye spllce. This would result in three eyes going around the shackle using a single line. Would this be a trifocal shackle

Mark Gardiner

You can tell if the knot is set by banging it on a table. If it sounds like a wood ball hitting the table, it is set. If it sounds like slapping a pillow on the table, it is not set. When they are set, they are rock hard. You want thousands of pounds so basically a lot of leverage. http://L-36.com/soft_shackle_9.php

[Brion Toss]

Hi again,
Ohi, it has been way too long. Sorry to have kept you waiting. Here are some details from the TreeStuff tests:

I'll orient these notes around Allen's samples, since he so bravely offered them for public scrutiny. He sent in 2 conventional (lanyard) shackles that broke at 122 and 116% of rated strength. This puts the average at about the same 120% that we had seen at NER and elsewhere. It is less than the, as I recall, up-to-150% breaks that Evans Starzinger had recorded with the same knot with 1/8" rope, but I now more than suspect that efficiency just doesn't scale reliably with rope that is larger.
We had a total of 8 lanyard knots submitted. Some of them were clearly low-quality -- you could tell just by looking at them -- and sure enough they broke weaker than all the others. Nevertheless, we got an average of 106%, with 105% at the low end, and 135% at the top. Sure enough, the knots with the highest average were also visually superior; they were symmetrical, and drawn up well. Remember this when you are buying soft shackles. One other variable: the testing machine ran at a faster rate than in previous tests, and I had thought that might have given lower results. But I now believe that operator skill was the deciding factor.
By the way, one of this group was a variation of mine, designed to function as a hank for a stays'l. It broke at 130%.
Allen also sent in an 880 button variation, that broke at 214%, which puts it towards the high end of the samples tested. In this group, the weakest was 175%, and the strongest was 225%.Again, quality of finish resulted in higher break strengths. But I'll note that even the weakest, least handsome of the bunch tested much stronger than any of the lanyard variety. The only way I managed to get a weaker result was by making an eye that was intentionally too tight. When I did that, I got a break of 109%. For those of you who might have been making tight eyes to prevent the shackle from opening under flogging loads and such, here is a good reason not to. Another reason is that, in my experience so far, even relatively big-eyed shackles simply do not come apart until you want them to. Don't get intentionally loose-eyed, but don't strangle the poor things, either.
I also submitted three "filled" variations, in which I inserted a short bit of smaller line into the eye of the shackle, to provide more mass. One of these broke at an impressive 275%,but another only reached 232%, and the third only got to 162%. This configuration shows promise, but we need to figure out how to do it right.
One more variation, from an idea courtesy of Mr. Gardner, essentially combines two shackles into one: you make eyes in two pieces of rope, then tie their ends into a 4-strand version of Ashley's 880, and then bury them as usual. The first two samples I submitted broke at 339 and 319%, and a third one I sent as a follow-up was just yesterday reported as breaking at 343%. While this configuration does break the 300% barrier, it is rather labor-intensive to make, and awkward to button and unbutton, as you are dealing with two eyes instead of one. Still, it might be useful in applications where space is at a minimum, but loads are high.
I also submitted one more variation in which the two eyes were combined into one. This would get rid of the awkwardness. Unfortunately, this broke at a mere 230%, and it did so with a surprising failure at the button. This looks like another promising configuration that needs more work.
Short form: the button variation, properly-tied, is reliably at least 50% stronger than the lanyard version. Because previous tests tend to show that efficiency lessens with increase in diameter, it seems preferable to go with the button version in a smaller diameter when the loads go up, rather than making a lanyard version in larger rope. The smaller rope also produces a shackle that fits into much tighter spaces. This had proven valuable in joining rope anchor rodes to chain.
I'm ongoing sorry about being so slow in getting all this down. Thank you for your patience.
Fair leads,
Brion Toss