Rigging question

[Chads93GT]

I never thought of that. Thanks for pointing that out. is the inline 8 not a 3 way directional knot?

[Tim White]

is the inline 8 not a 3 way directional knot?

Nope. Hence the name In-Line 8.

[Chads93GT]

heh touche ;)

[Kendalcaver]

some people rappel short distances with it right?

Yes, I've rappelled short distances on a Munter hitch (maybe up to 50 feet); it kinks up the rope quite badly, though, so I wouldn't recommend it as a standard technique!

[chh]

I'll second the butterfly as a midline knot or a rebelay knot. Super nice knot.

As the munter goes. If rappelling on it keep the brake and load strands parallel. Control the rate of descent with your left hand and keep your right hand on the brake line as a backup and to deal with snarls. Works a lot better this way I think.

That said, there's allways the old addage, "if you can't tie a knot, tie a lot!"

[Chads93GT]

Yeah ive used the tie a knot tie a lot. I've had the chance to untie some lines off of a ladder in a project cave here that was "rigged" before I moved to the area. OMG it was a cluster. lol Thanks for the tip about the butterfly as a rebelay knot. I didn't know it could be used like that.

[AlanfromOz]

Re butterfly / alpine butterfly (henceforth ABK)...

Many people have love the ABK, but there is some debate about its' usefulness.

Some (generally older) sources claim it is "one of the strongest knots", but testing shows as little as 60% residual strength, with very variable results. The problem is that it is used in any different ways (eg end loop, mid rope loop, 3-way loading) and different test results have used different modes and not always said how they tested it!

I found some test results where it failed before the bowline in the other end of a 11mm rope (~900kg at breaking from memory?) Shock loading performance is also not great according to some tests. This might be because the first loop of rope in each tail has a quite small bend radius?

It is far from ideal for a rebelay, because you're using it for an end loop knot (something is wasn't meant for) and if your rebelay anchor fails, you're shock loading it end to end!

Whilst it looks like the best knot for a 3 way load, there is usually a better / as good knot. Many people use the ABK in the centre of a cowstail lanyard, but a fig 8 or overhand is better (when do you load both lanyards in different directions?).

Even for a 3way load, a figure 9 knot will perform as well or better, even with the tails pulled apart (yes, it looks wrong!)

(ref. "Vertical" by Al Warild http://www.cavediggers.com/vertical/3knots.pdf )


Alan

[chh]

900kgs seems awful low for the breaking strength of the butterfly. I've been using it because I was under the impression that it WAS as strong as a bowline or even an eight. Maybe that's just my "generally older" knowledge. It might make be do things differently if that was indeed the case. 900kgs is about 3000 pounds I think, and doesn't 11mil have a mbs of about 6ooo pounds? That's only about half. But I'm sure the usage does make some difference. I've used the butterfly at rebelays because it used very little rope and is easy to untie. The figure 9 would as well I'm sure. I've also used clove and polish hitches but a lot of cavers seem to balk at having a hitch instead of an actual loop to use at the rebelay for some reason. Presumably because the loop is easier to clip if the biner at rebelay is really against the wall, though I don't know that it really matters that much. The double bowline also works well as far as untying, but also uses a tad more rope and is bulkier, but that may not matter. I'd be interested in seeing those tests if you can dig them up alan.

-caleb

[AlanfromOz]

There is some information on knot strenght in Vertical. I know from speaking to Al that he performed drop tests on a lot of knots before the first version of vertical. Some people have disputed his results, but never presented the claimed test results to back it up!

The link to the slow pull testing is:
http://oberon.ses.nsw.gov.au/resources/ ... ington.pdf

Last couple of pages - It was actually 1000 / 1100 kg. Unfortunately no one seems to do 'proper' studies, with repeated experiments, controlling variables, etc. Guess it just takes too long and is too expensive.

If anyone has any more information, I'd love to see it.


I do have a soft spot for the ABK, but I think the data shows enough problems to justify using other knots by preference - especially when they are AT LEAST as good, probably better.

The big one i have a problem with is the ABK in the middle of a cowstail - I can't see any advantages here.


Alan

[knudeNoggin]

I never thought of that. Thanks for pointing that out. is the inline 8 not a 3 way directional knot?

That depends on the "3-way" aspect: you didn't state what sort of angle
obtained between eye and tail of the directional eyeknot. I suspect that
it was something around 45degrees, or less; that wouldn't concern me
with the directional Fig.8. (You might have also used the Butterfly or
a bowline variation or some other like-a-Butterfly knot built up from
a Slip-knot.) (And I'd have not bothered with the Tensionless Hitch
shenanigans at the start -- just a waste of rope! Btw, you don't say
how you tied off the webbing to tree#2.)

- - - - - - - - - - - - - - - - - - - -

As for the Butterfly / Lineman's Loop doubts, this is going into academic
realms: we have a lone person rapping, not the lowering of a Hummer.

Firstly, you should find inconsistent testing -- and to whatever unfavorable
results you see there are the Dave Richards results which have the knot in
one of the trio of ropes as strongest (!), and consistently strong throughout.
"YMMV" !? Seldom will you see a clear image of exactly what knot geometry
was tested -- and tying guidance & practice lead to different geometries:
with the Butterfly, do the eye legs cross on exiting the knot (as one pair
of originators --Wright&Magowan-- stipulated) ? Do you know which end
of the knot was loaded? (It is asymmetric, so this makes a geometric difference.)
... and so on. Moreover, in practice, one will see the knot loaded in some
variety of ways --on all four parts, or maybe on ends only-- before some
exceptional event might occur: that's a lot of variability to toss into some
testing regimen!

Secondly, one should consider that slow-pull testing doesn't so accurately
model in-practice loading: when rope partially breaks under slow-pull
testing, there is often a momentary reduction of tension by some release
of material, until the steady-rate-of-movement hydraulic pulling catches
up -- not so in practical circumstances, where e.g. a falling mass will
keep on keeping on (albeit with some reduction in force).

Thirdly, in the given example here, there's little chance of any significant
"shock loading" -- the knot will begin under tension on all parts, which I don't
know of any test showing.

Now, for that ULRinked-to page of pretty photos (yea!) of loaded & sometimes
ruptured knots. There are some obvious errors in this reporting, which suggests
some inattention or knowledge about testin/reporting/knots (as the disclaimer
allows).
Let's see what can be found.

1) 1st knot, basic "double-wrap" Prusik : "failure" is left to the imagination,
but it must be slippage, as the knot's intact. One can see from the photo how
the away-coil of the knot elongates (like e.g. a Klemheist or Blake's hitch),
and so its mainline presses into the near-coil (which is like a Rolling Hitch),
and thus effectively pushes it along, defeating its grip. (This is the problem
one can find in the upside-down mistaken images on On Rope, 2nd ed. for the
so-called N-Coil Prusiks -- where the single wrap of the one leg will defeat
the N wraps of the other, pressing into rather than (correctly) pulling away
from them.)

2) 6mm Triple-wrap Prusik : "both legs failed simultaneously" !!! Wow, who'd
have guessed this; unless the load-end of these legs was itself knotted and not
freely clipped in to a pin or 'biner. Simultaneous failure ... ?!
But the rupture points are not equal, one apparently somewhere within the
knot, and one apparently in two points (kern & mantle) not quite so!?

3) Fig.8 knot after loading to 1083kgs. : Notice the geometry: this knot
is symmetrically formed and loaded on its interior part for
the end/mainline (lower right); the exterior part is way loose in a big
loop at top. This might be the more common orientation. (But see
2nd following note here.)

4) Fig.8 v. 9 showing Fig.9 @2074kgs. : This Fig.9 was similarly loaded
vis-a-vis its parallel ends (interior). (Lyon Equip.'s testers felt
this was preferable for Fig.9, indifferent for Fig.8.)

5) Comments & Load at "Failure" <blank> ; Fig.8 pictured : here
we are to suppose this is the knot discussed above (my 3); but here
notice that it was the exterior strand that was loaded --there
is no big loose loop of the end's strand, which here is compressed
wide, adjacent the bits of ruptured loaded strand -- which end
has pulled free of the knot (so, not breaking at the collar/entry).

6) Fig.8 v Fig.9 ; 2190kgs. : The FIg.9's interior strand shows the
great tension, while the exterior one is loose; this is less clear for
the Fig.8, but must be the same if the knot was symmetrically,
evenly dressed (no crossing strands), and this stands in contrast
to both the ruptured Fig.8 of comment 4 and the quite loose-stranded
one of comment 3 (though we're led to believe that one of the breaks
is for the #3 knot -- not what I'm seeing).
One can see here a relatively sharp/all-at-once break that must've
come at about where the strand bends around the interior collar;
the rest has pulled free of the knot, allowing this end to un-turn.

7) Tape sling ... ; 551 kgs. : "noted that the outside bend of the
tape knot failed" NO, rather, the exterior positioned end failed,
but not at the obvious "bend" --which one can see is still in place
(solid yellow side vs. face w/3 blue stripes). The failure occurred
at the point of tight compression on the heavily loaded strand,
the opposed and interior-positioned one having the protection
against such compression of the broken strand end. (I've seen a
case where rupture had begun --but a break came elsewhere--
at the center point of this knot (here, upper right point).)

8) Tape sling, single strand, >10 Y.O. ; 785 kgs. : Once again,
look at the hash marker threads (blue), and see that the broken
strand remains in the knot, the break at high-compression entry
seeing the end pulled free.

...
[skip to next KNOT]
9) 25mm tube tape both knots tied ... Compare with ... :
Well, what to deduce? That here the test specimen was two,
opposed, EYEknots; but from "sling" in prior descriptions (above),
I figured we're talking about the usual instance of a ring sling
-- so, a knot here loaded on 2-vs-1 strand vs. above 1-vs-1.
The difference in effect is that there should be about half the
tension on the strand pressing against the singly-loaded
strand (mainline) of the eyeknot vs. the end-2-end knot,
and that should have some ameliorating benefit to cutting
(there will be behind this half-loaded eye-leg strand though
the twin eyeleg strand).
There IS a way to tie a Water knot /Ring bend so that it is
symmetric, and both loaded ends interior or both exterior;
it was tested in the former case, and found to be about as
strong as the usual case (1" tubular nylon).

[... skipping down a few to ... ]
10) 4-wrap "Klemheist Knot in 8mm cord ... ; 1744 kgs. : note how
contorted the hitch+object-rope have become; this would not be
the case were the object-rope (11mm) under some tension, as
in the case of the OP where a friction hitch used vice eyeknot.

11) "Bloodknot" -- single marlon spike / double... [sic] ; 893 kgs.
Okay, this is a bit comical -- "marlon spike", not written by one
w/maritime knotting background (it is variously "marlinespike",
"marlinspike", even "marlingspike" ! prefer the first).
But the photos don't match -- note red marker yarns in lower
photo (which happens to match description, btw). The photo
of "The knot before loading... " is of something else (!); I'm
having a heckuva time discerning the righthand knot (the
left is the expected Dbl.Overhand / Strangle knot of the
Grapevine bend, as is the righthand/surviving knot of
the lower photos broken bit).
To names: "blood knot" might have originated for knots
put in single strands of a multi-strand whip, to draw blood
-- and here they'd be dbl./triple/ overhands (a Strangle
w/o anything within it being strangled, one could say);
but currently the most common application of the moniker
is by anglers to an end-2-end knot where tails emerge either
in parallel or in opposition from the center, at a right angle
to the axis of tension.


Whew, enough for now.
The Butterfly's broken bits make me wish for a pre-break
image; it's a rather small part of the initial turn that remains,
and original orientation is unclear to me, now.

*knudeNoggin*

[Chads93GT]

The webbing was tied off with a tensionless hitch too actually. I didnt feel like dealing with a horribly set knot to undo to get my webbing off. The angle on the inline 8 was around 45*. I ran it out closer and closer to the lip to get the angle down as small as possible.

Now.........I shall read the rest of the post, thanks for the info

[AlanfromOz]

As for the Butterfly / Lineman's Loop doubts, this is going into academic
realms: we have a lone person rapping, not the lowering of a Hummer.

True. Anything a caver is doing with rope is unlikely to break it with a static load. Where is becomes important is dynamic loading, particularly when you start talking rescue loads. You could probably rig any given cave with thumb knots and live... but I wouldn't recommend it!

Seldom will you see a clear image of exactly what knot geometry
was tested -- and tying guidance & practice lead to different geometries:
with the Butterfly, do the eye legs cross on exiting the knot (as one pair
of originators --Wright&Magowan-- stipulated) ? Do you know which end
of the knot was loaded? (It is asymmetric, so this makes a geometric difference.)
... and so on. Moreover, in practice, one will see the knot loaded in some
variety of ways --on all four parts, or maybe on ends only-- before some
exceptional event might occur: that's a lot of variability to toss into some
testing regimen!

Secondly, one should consider that slow-pull testing doesn't so accurately
model in-practice loading: when rope partially breaks under slow-pull
testing, there is often a momentary reduction of tension by some release
of material, until the steady-rate-of-movement hydraulic pulling catches
up -- not so in practical circumstances, where e.g. a falling mass will
keep on keeping on (albeit with some reduction in force).

Thirdly, in the given example here, there's little chance of any significant
"shock loading" -- the knot will begin under tension on all parts, which I don't
know of any test showing.

I think this is the heart of the problem - knot tests don't usually give us enough information to know how it was tested. Also, I'm yet to see any 3-way load testing.

You are also right about the shockloading - that's the only way we'll ever get near breaking a rope. The only drop testing information on the ABK I've been able to find (Warild) rated the ABK rather poorly.

[NZcaver]

Many people have love the ABK, but there is some debate about its' usefulness.

Some (generally older) sources claim it is "one of the strongest knots", but testing shows as little as 60% residual strength, with very variable results. The problem is that it is used in any different ways (eg end loop, mid rope loop, 3-way loading) and different test results have used different modes and not always said how they tested it!

I found some test results where it failed before the bowline in the other end of a 11mm rope (~900kg at breaking from memory?) Shock loading performance is also not great according to some tests. This might be because the first loop of rope in each tail has a quite small bend radius?

It is far from ideal for a rebelay, because you're using it for an end loop knot (something is wasn't meant for) and if your rebelay anchor fails, you're shock loading it end to end!

Whilst it looks like the best knot for a 3 way load, there is usually a better / as good knot. Many people use the ABK in the centre of a cowstail lanyard, but a fig 8 or overhand is better (when do you load both lanyards in different directions?).

Even for a 3way load, a figure 9 knot will perform as well or better, even with the tails pulled apart (yes, it looks wrong!)

There is some information on knot strenght in Vertical. I know from speaking to Al that he performed drop tests on a lot of knots before the first version of vertical. Some people have disputed his results, but never presented the claimed test results to back it up!

Interesting point about inconsistencies caused by quoting older sources of information (perhaps older ropes used for testing?). I've found this a bit frustrating too, but ironically Warild's earlier material turned out to be the cause of the frustration not the solution! The 1988 edition of Vertical was the first reference I had to rope testing figures. I believed it - until about 8 years ago. By then I was seeing results of more recent testing by a variety of different sources (On Rope, Life on a Line, Nylon Highway, Lyon, etc), and witnessing some testing myself.

It soon became apparent that many of the arbitrary figures in Vertical (mostly attributed to G. Marbach and J-L. Rocourt, 1980 and Courbis, 1984) were simply not consistent with more recent testing results. Of course rope testing results can vary greatly, but some of those published in Vertical were well beyond any standard deviation. In turn, this formulated preconceptions which weren't necessarily accurate. Knots declared as "non recommended" by Warild were and are regularly used by cavers and rescuers, and there's no trails of dead bodies in their wake. Knots like the clove hitch, the bowline and the butterfly.

I had a brief look for more recent testing results for the butterfly, and all I found was this Nylon Highway article which doesn't indicate anything too upsetting (although test parameters are for slow pull, not shock load). I'm wondering if Life on a Line or Lyon or anybody else has done recent testing with the butterfly. Surely somebody has?

Bottom line, it's great that Warild has updated his source material and is continuing to move forward. I like the butterfly in certain applications and I think it can safely be relied upon, but I'll keep an open mind.

[AlanfromOz]

Knots declared as "non recommended" by Warild were and are regularly used by cavers and rescuers, and there's no trails of dead bodies in their wake. Knots like the clove hitch, the bowline and the butterfly.

In defence of Al, the knots you listed are under the 'other knots' section and have the foot note " 'Other rigging knots' are knots that work, and are often used, but there are better knots for the same purpose". I suppose he would argue you don't *need* any of them when there are more foolproof alternatives. Then again I know he uses some of them on a regular basis!

Maybe it's a case of giving anyone reading a book a basic set of knots that will work. Let those who know the potential pitfalls use other knots where it matters (eg bowline on a bight rather than double fig 8 where saving rope or being easily able to untie it outweigh the lower strength)

Surely somebody has?

I hope so, although the lack of dead bodies probably makes it a low priority!

[knudeNoggin]

Many people have love the ABK, but there is some debate about its' usefulness.

It is far from ideal for a rebelay, because you're using it for an end loop knot (something is wasn't meant for) and if your rebelay anchor fails, you're shock loading it end to end!

?! Wright & Magowan, who presented the knot in a 1928 article, envisioned it to be used for a middleman's connection; I should expect in such cases it would at times of need see such end-knot loading. I'm unaware of what (electrical) linemen intended it for.

Even for a 3way load, a figure 9 knot will perform as well or better, even with the tails pulled apart (yes, it looks wrong!)

Hmmm, YMMV. It is interesting to note, though, in this vein, that the CMC test results shows for "rope with a loop in it" (i.e., pulled end-2-end, eye unloaded) the Directional Fig.8 to be 6 %pt.s weaker than a plane-ol' Fig.8, with a Butterfly 4 %pt.s stronger. Here, again, we have no idea of the respective knot geometries that met the load: keep reminding yourself of this!

Interesting point about inconsistencies caused by quoting older sources of information (perhaps older ropes used for testing?).

Well, necessarily older ropes in many cases (decade & more ago) !
And here one must remark that rather than thinking/asking How strong is Knot_K1? one should ask How strong is Rope_R1 knotted with Knot_K1? -- material matters.
AND BEYOND that then comes the almost never-specified exact geometry of what "Knot_K1" denotes. Again, look at the aforementioned knots-photos URLink and that upper Fig.8 eye knot presented devoid of any information: its broken strand lay exterior, unlike those of the other Fig.8s presented -- that is a significant difference in such tied-with-a-bight knots which is typically ignored (Lyon Equipment testing however did try ...). So, one really has a lot of different structures going under a common moniker and producing varied results, any one of which might be what gets attention and ... echoes.

Knots declared as "non recommended" by Warild were and are regularly used by cavers and rescuers, and there's no trails of dead bodies in their wake. Knots like the clove hitch, the bowline and the butterfly.

Recall Bob Thrun's apt remark that if knot choice seems to depend upon break strength that he should be using a stronger material. One can see non-strength reasons for avoiding the Bowline; to its vulnerability to loosen there are however ample recourses, such as re-tucking the tail or tying it off.

I had a brief look for more recent testing results for the butterfly, and all I found was this <Dave Richards's testing> which doesn't indicate anything too upsetting ...

Here, the Butterfly --OF WHO-KNOWS-WHAT GEOMETRY-- comes out as tops, near tops, and neck & neck with the Fig.8 eyeknot(s).
(NB: Would ANY of you think to test a "Fig.8 re-woven" AND a "Fig.8 on a bight" as distinct knots???! Richards did, and so, too, Jim Frank's CMC Rope Rescue Manual (3rd ed.) !? I could see doing some study of these tying methods to see if they produced similar geometries (and then test if different); but I'd want to point out the reason for the apparent (name-wise) duplication. For, otherwise, the knots are supposed to be "properly" (where is THIS detailed?!) dressed & set, and so should be equal irrespective of tying method. In the Dynamic rope testing by Richards, there is a puzzling big difference in Stnd.Dev., though Means are similar; in the low-elongation rope (Rhino 12.7mm?) testing by CMC there was a 1%pt difference (=).)

I'm wondering if Life on a Line or Lyon or anybody else has done recent testing with the butterfly. Surely somebody has?

Although one can question the significance of time in "recent" vs. other differences (Rope-X, geometry-G, figuring of % reported (i.e., based on tested tensile or vendor-rated tensile ...)) , the 2nd ed. LoaL gives "60-70%" vs. "65-75" for Fig.8 (but one should note that a quite uncommon geometry is presented for the latter, and who-knows for the former). An interesting tidbit: this reference states that the eye of the Butterfly "can be loaded in expansion only if the main line is also under tension." !? I don't see the emphasized term used in the presentation of any other (eye)knot, so one must wonder at what sort of evaluation others got re this, and why it's mentioned only here. I presume it to mean "ring-loading" --i.e., treating the eye qua ring sling-- ; and what is interesting is that it is possible that such loading would capsize the knot into . . . a Butterfly (bend) ! -- but I think it's also possible, and probable with imbalanced loading of the ends, to have a partial capsizing, so don't count on the nice result (I discovered this for some similar construct). IIRC, Lyon's 2001(?) testing put the Butterfly in good standing.
(And there are many dubious statements to be found in LoaL , to beg questions.)

I like the butterfly in certain applications and I think it can safely be relied upon, but I'll keep an open mind.

But one should question how any of this strength testing should influence the use of the knot; rather, one would I think want to know about stability, ease of adjustment / untying, & ... , and where are (what are...) the tests re that?

*kN*