Rope joining knot failure - death

mobius · 9 June 2016 11:33

Thank you Dan, I have spent some time trying to work out out the video and ‘results’ (French is not my native language).

Dan and I are mostly at odds with how good the EDK is (or so I feel we are ). I do not like the knot for what I consider good reasons, check some of my recent posts, and I have not changed my mind. However, the video result (above) is nothing I want to use to support my case.

The image I show below is the best I could get from the start of the EDK video trial. As somebody who tries to trial knots with serious intent, I would not present to a discerning readership a dyneema vs dyneema EDK trial, with a skull and X-bones fail, passing that off as relevant. The EDK is quite possibly poorly tied in the trial, just to add insult to injury!

I have not had much direct experience with dyneema. I have trialed 3mm dyneema exactly once (with a bowline). That said, I do not expect it to ‘work’ with any reasonably footprint-sized offset knot, including the ones I have worked with lately. That does not mean that offset knots should be ignored in this application, unfortunately the video gives the counter impression from what I gather from it. Dyneema is a very harsh mistress when it comes to knot integrity from what I know about it.

The second image I show is a knot I did not recognize (what is it?). The fact that it sits against the the D-clip under load annoys me greatly. I wanted to see the knot perform in a bit of space.

Cheers,

Ian.

[Edit: it is suppposed to jam that way apparently]

EDK in dyneema.JPG

What is it.JPG

agent_smith · 10 June 2016 01:11

I have attached 2 photos from my upcoming paper on ‘analysis of abseil rope joining knots’.
They are reduced resolution/quality due to the size restrictions imposed on this forum.

The second (lower) image illustrates the concept of using a thinner ‘pull’ cord to retrieve the abseil ropes.

The ‘white’ rope would be the thinner/lighter ‘pull’ cord.

The blue rope is deadheaded (jammed) up against the anchorage.

Obviously, this is not a nominal loading profile for #1410 - since it is reliant upon jamming up against a small orifice - with that orifice being of a such dimensions that the mass of the knot cannot pull through.

The principal concept behind this technique is weight savings. Full length abseil descents are still possible…however many of the advantages of a double rope climbing system are sacrificed. The lead climber can only lead with a single dynamic rope (instead of 2 dynamic ropes).

The thinner pull cord is simply to allow retrieval of the rope after a full length abseil descent.

Commentary:

All of the benefits of the offset knot profile are lost (if a larger knot mass is used instead of an offset geometry)…if the knot must translate around a 90 degree edge, it may become stuck.
In this particular loading profile, it would be preferable to tie #516 (double overhand stopper knot) in the blue rope - and then ‘strangle’ the white pull cord around the blue rope (in effect #1415 Double Fishermans). #516 would create a larger and more secure mass to jam up against the anchorage (however in doing so, you lose the benefit of an offset knot profile as stated in #1 above).
The context in which this type of system would be employed is ‘sport climbing’. This is a type of climbing where the route is fully equipped with existing permanent bolts to ‘protect’ the route plus a well defined existing ‘rap station’ (or belay station) where a descent can be initiated or the climbing route continues up to the next ‘rap/belay station’ (ie a multi-pitch sport route). This also assumes that that route is vertical - so the presence of ledges and protuberances may not be a factor - and so the abseil ropes will retrieve easily. ‘Rap/belay stations’ can often be strategically sited so as to permit a clear and unobstructed path for the rope retrieval process.

EDIT: I have finally found the accident link related to this type of jammed knot loading profile:
Link: http://www.rockandice.com/climbing-accidents/rappel-knot-fails-climber-falls-300-feet-to-death
Victim: Brian Ellis
Route: Linkup of Serenity Crack (5.10d) to Sons of Yesterday (5.10a) Yosemite, California USA
Outcome: Fell to his death

Commentary: http://www.traditionalmountaineering.org/FAQ_ReepschnurRappels.htm
Reepschnur translation = accessory cord

The art of double rope climbing is in slow decline…with the majority of modern climbers opting to use a single rope system. I did say ‘majority’…there are still some very active climbers (like me) who continue to use ‘double rope’ lead climbing techniques. The big trad routes - and in particular, new routes (ie first ascents) still see double rope techniques employed. I personally enjoy the added safety benefits provided by double ropes (including self-rescue options). You need a competent belay partner who is proficient in belaying with double ropes…

Mark G

EDIT NOTE: Clarification on loss of benefits afforded by offset knot profile… If you use a thin pull cord to retrieve your abseil rope, you have to assess the risk of the knot pulling through the anchorage link. It may turn out that you need a larger knot mass and therefore you will have a larger footprint - and it most likely may not be an offset geometry.

agent_smith · 10 June 2016 01:57

Here is another accident link I have found… this one is closer to home.

Link: http://www.theage.com.au/victoria/melbourne-university-climbing-club-culture-branded-dangerous-after-twin-accidents-20150708-gi7eu3.html

It appears that the club responsible in both abseiling accidents has a ‘culture’ of silence.
Because of this culture - we may never know the facts which caused the accidents…
NOTE: There were no fatalities, however both were serious injuries requiring hospital treatment.

Another link to the Mt Buffalo accident: http://www.abc.net.au/news/2015-04-26/man-falls-15m-while-abseiling-at-mt-buffalo-after-rope-snaps/6423018
(news reports are usual unreliable or manifestly wrong…so its hard to determine fact from fiction).
And another link: http://www.bordermail.com.au/story/3040981/mountain-miracle-abseiler-survives-fall-after-landing-in-worst-possible-spot/
There is a clue given that it may be the ‘inadvertent’ release of the abseil ropes from above by an unknown person or persons in the party.

Mark G

Dan_Lehman · 10 June 2016 20:50

I don’t see this --one still has an offset end-2-end
knot (if that was indeed what was tied). Although
there might need to be some test-inspection of how
various offset knots when loaded qua stopper in
the various (largely thick+thin) materials and expected
forces (NOT at Mobious forces!) react / distort.

–dl*

Dan_Lehman · 11 June 2016 14:40

Wow, I’ll quote a bit of that article:

It followed another accident on April 26 in which a male member fell about 20 metres while abseiling on the north wall of Buffalo Gorge, in the state's north-east. The 24-year-old was lucky enough to fall into a pool of water, but suffered back injuries and was trapped at the bottom of the cliff in sub-zero temperatures overnight. Search and rescue police, paramedics, and State Emergency Service crews hiked in to the gorge and winched the club member out the next morning and he was taken to the Royal Melbourne Hospital, the ambulance spokeswoman said.
The club did not report the severity of the Buffalo Gorge accident
and told the university the injured club member was in a “good”
condition, therefore not triggering a “critical incident response”
by Melbourne University Sport, a university spokesman said.

Ahhhh, yes, right : all’s well that ends well (the frightening
in-betweens be darned) !!

Where falling XX meters (okay, even 2) and being trapped
and isolated overnight (and at freezing temps, no less)
is not a critical incident in need of accounting,
I’m … not joining THAT group!

–dl*

roo · 9 September 2020 21:29

from: https://rockandice.com/climbing-accidents/rappel-knot-fails-climber-falls-to-death-on-the-goat-wall/

ANALYSIS

Two Okanogan County search-and-rescue members, Ottis Buzzard and C.B. Thomas, recovered Kautz’s body about an hour after Jackson found it. Kautz was still attached to the rope and his rappel device was correctly threaded. Thomas carefully checked both ropes and found no damage. There were no knots in the ends of either rope.

“Ryan had told me he had never done a multipitch before,” Johnston wrote in an e-mail to Rock and Ice.

“That group was very inexperienced. It was obvious from their gear, the way they talked and the way they were climbing and belaying.”

Johnston went on to describe the most likely scenario for the accident.

“They had been using a flat figure 8 backed up with overhands all the way down the route. [Kautz] tied the same knot, no doubt, but didn’t back it up, or it would not have come untied. A flat overhand is the preferred knot for tying two ropes together for rappelling, but the figure 8 looks more substantial so it gets used, too. But a flat figure 8 can walk and unwrap itself when loaded. If the tails are too short, it can walk right off the ends of the rope.”

Dan_Lehman · 11 September 2020 21:41

Four-year anniversary, nearly (of this tragedy)?!

Of course, I don’t fancy “flat” as the adjective used
vice “offset”, but that’s a trivial matter (“flat overhand”).

What is missing from these knot analyses is that knots
simply aren’t well presented, to the point of specificity
of formation --though in some cases, such as the Fig.8
eyeknot, anything goes pretty well.
But not for the offset end-2-end Fig.8. There is some
French site that actually recommends this knot OVER
the EDK (offset water knot), and claims higher values
for tension on which it rolled (or broke?). And IMO
there IS a way to SET the knot so that it shouldn’t
be an issue, in abseiling; but such specification just
isn’t in the community literature --at most one sees
“pull on all ends” which is partly WRONG; and here the
order of pulling, if not direction, matters.

–dl*

agent_smith · 3 January 2021 00:22

Found some more info about the Kautz accident: (although it appears to be a ‘copy and paste’ from the original Rock and Ice article.
http://publications.americanalpineclub.org/articles/13201214416

And here are some more accidents that involved knot failure:

and so on…

…

With regard to using an offset F8 bend (ie #1411 tied in its offset form) - there really hasn’t been any peer reviewed testing to determine precisely which geometries provide stability and security. We have ample empirical evidence that shows the offset F8 to be unreliable.
However, it may indeed be possible to arrive at a particular geometry which is stable and secure up to a reasonable load threshold.
To prove this theory out, well designed and documented testing needs to be performed (and then repeated by other testers to confirm or refute the results).

I would be happy to publish such test results with acknowledgement in an update to my existing technical paper on offset rope joining knots…

Dan_Lehman · 7 January 2021 00:36

Don’t have their URL on hand, but there is some FRENCH
cavers(?) site who recommend the offset Fig.8 ends joint
over the infamous EDK / offset water knot, and show test
results that speak to its (slight) favor --both knots showing
proper behavior (which is staying tied, not holding trucks)
through reasonable loadings.

IMO, the situation points to the lack of specificity in
knotting :: knot books give half-assed presentations,
and users are all over the place; “dress & set” might
be uttered, but almost never with exactly what that
means!
(And this indeterminateness will make others’ test
reports of little use --you won’t know what exactly
was tested! The specification needs to precede … .

E.g., in the case at hand –offset fig.8 & water knot–,
the setting (tightening) of the knot should be done
by pulling hard on the SParts TOGETHER with the
knot held qua stopper in resistance; and done
iteratively (perhaps one iteration, though) with
reversing the orientation to set the tails tight,
which means tightening their choke around
the SParts --which have NOT BEEN PULLED
APART YET :: key point!!!

THEN, the knot takes a tightened state that
resists being pried apart.

IN CONTRAST to setting by “pulling on all ends”
aimlessly, which sees tails opposing SParts and
the choking turns sorta snug BUT the tails not
hauled down as loading will do, and so the
initial loading tends to open the choke more
than were the tails hauled in by SParts’ pull
as described above --a pull that doesn’t open
the choke.

Possibly, as you’ve done in some instances via photography
& text (and the calibrated-forces loading to be shown),
you could shown a pulled-by-tails-vs.SParts setting
and pulled by SParts vs. held-knot setting (perhaps
even sans a following pull on tails IN THAT STATE**),
AND THEN
the resulting knots upon receiving, say, 100# force!?
.:. My belief is that there will be a noticeable difference
in the choking parts’ size/radius --and vulnerabilty
to flyping.

[**When loading SParts-adjacent vs. nub, one will
fold over the tails, and probably incur some bit of
slack in them that some iterative setting of the
tails --careful to keep their folded-over state–
will remove.
]