TEST REPORT: Offset overhand bend (#1410)

Practical Knots Knot Test Reports and Reviews
1

TEST REPORT

Test date: 07 Dec 2018

Knot specimen: #1410 Offset overhand bend (aka ‘EDK’)
Knot class: End-to-end joining knot

Test objective:
To determine the effect of rotation.

Rope material:
EN standard: EN892 dynamic rope
Diameter: 9.0mm (‘half’ rope) - only tested equal diameter ropes to rule out any variables - unequal ropes will follow later…)
Manufacturer: Edelrid (Germany)
Condition: 10+ years old (retired from lead climbing)

Tester: Mark Gommers

Test configuration and parameters:
As per attached images.
Equal rope diameters used to form the end-to-end rope join (9mm to 9mm).

Assumptions:
It has been theorized that a simple rotation can improve stability (per Dan Lehman).
#1410 is typically tested in its mid-rotation state. This mid-rotation state is the default mindset. I am of this opinion because firstly, it is never reported and secondly, you have to make a deliberate effort to perform the rotation - its not a random occurrence - and only one particular rotation state appears to be effective. Furthermore, its never been photographed in detail by other testers.
As far as this tester is aware, nobody has tested and published the effect of different rotation states - with specific attention given to identifying the rotation state which is most effective at improving stability.

Conclusions:
The rotation state where the choking rope segment is lying furthest from the axis of tension is most effective.
At a peak load of 6.0kN, no instability was observed.
The rotation state where the choking rope segment lies closest to the axis-of-tension induces a torque - which causes the knot to revert to its mid-rotation state as load is increased.

Due to limitations of the test rig and equipment, peak loads were well below the MBS yield point of the knot.
Follow up testing is required to determine the threshold of instability with this rotation state compared to a ‘control’.
The mid-rotation state would be the ‘control’.
At this stage, I would posit that this rotation state will boost the threshold of instability compared to the ‘default’ mid-rotation state.

I encourage others to try to replicate my results using EN892 ‘half’ ropes (or a similar facsimile rope if EN892 rope cannot be sourced).
NOTE: In my test, I used ropes of equal diameter. This was to done to rule out any potential variables.
Unequal diameter ropes would be tested at a later date…

2

These statements might –leaned on hard-- be wanting
for factual support. By which I mean that although we
might see the knot presented in this orientation,
and even if we find by empirical observation that it is
frequently tied so,
likely we cannot be sure of the tested knot.
(IMO, you’re probably correct; but we just don’t know
for sure, as this aspect hasn’t been noted and reported
on by anyone --testers or otherwise, for that matter
(save me/us)!)

–as you remark:

As far as this tester is aware, nobody has tested and reported on the effect of rotation.

As for

It has been theorized that a simple rotation can improve stability (per Dan Lehman)
, more to the point is that while I have opined that the mid-rotation state might be worst, we simply need to examine behavior while paying attention to such states. So, the pure point is that **orientation** might make a difference.
RANT ON
furthest
Why do folks forget "far/farther/farthest" --measures of distance, of *linear* position & sense (even if said metaphorically) ?!? These are (were) common, long-used words. "Further" is a note of continuance or difference in broad measure, un*linear-like*. (Similar and sharper rebuke comes for the now seemingly vogue esp. newscaster abuse of "reticent" for the perfectly known word "reluctant" --egadz, folks are going lame-brained. RANT OFF

I, too, have found one of the extreme-of-rotation states
not so stable, wanting to turn into the mid-span state;
but I do believe that this extreme case CAN be SET
and hold, to rupture, if not perfectly extreme in the
formation of “arc” & “loop” of its SParts, at least
more biased thus than is the mid-rotation state.
(Given the at least sometimes need for special
setting & dressing, though, we might conclude that
it’s a state unlikely to occur in common usage.
But it would be worth making the effort, at least,
to get a clear test of it. Maybe also much of a
“YMMV” per material, likely stiffer cordage being
more prone to shift orientation than soft stuff
which in dressing & setting is more obedient.)

I encourage others to try to replicate my results using EN892 'half' ropes (or a similar facsimile rope if EN892 rope cannot be sourced).

ALSO DISsimilar ropes, as doing so with e.g. something
pretty elastic vs. static might shed light on the mechanics
of the knot. AND, the general in-practice state is of ropes [<- edit “if”=>“is”]
dissimilar in both elasticity & size : i.e., of a 6-7mm “haul
line” joining the 9-10.5mm dynamic rope. (Whereas the
doubly long “twin” 8mm ropes are used sans knot on their
full twin length (does anyone ever join two such ropes
and thereby skip alternate belay/rap stations?!!)

Of course, in the case of dissimilar ropes, there is the added
orientation of which one is the “choking” rope (with the ideal
being the thinner & more flexible one).

ALSO, I suppose one can do some at least quick’n’dirty
testing of the offset fig.8 e2e bend, which maybe allows
even greater variance in orientation. Maybe here, though
not seeking to recommend this knot no matter … ,
we’ll find that there is an especially dangerous orientation!
(Recall that some French folks seem to PREFER this knot,
and to have test data in which it performs better --i.e.,
rolls at only higher forces-- than the offset water knot!?
–which surprised me, but it is a presentation that came
up to my awareness in the past two years.)

Ropes available to me for my crummy 5:1 pulley stressing
–lousy pulley, but bouncing on it surely imparts surges
of force up to what ought to come by good sheaves alone–
are : ancient "Goldmantle’ --the name says it all (~40yrs)–,
some old 11mm Mammut dynamic, some other old bit
(discarded top-roping anchor lines), newer 11mm gym
rope, some newer & pretty unused 8mm ropes,
Sta-set 11ish mm yacht rope, 3/8 BW II, and the
incredibly intractable PMI pit rope & some Canadian
stuff that makes the former seem, er, not-so-bad
(!! fankly, these might be unlikely to form an offset knot
that any sane person would actually use!), and other
7mm kernmantle low-elongation ropes, & 6mm.

–dl*

3

Hi Mark.

I was thinking about this test, and btw, thank you for the work.

What I am thinking about is this: How about making a sling using two ropes and tying a comparable knot in each pair of ends. They could be oriented oppositely for that kind of comparison, even using different diameter, etc., ropes. Sort of as a tug of war, side by side.
It might be useful to have the ropes go around some pulleys to aid separation and limit any influences at the anchor points.

I wish you resided closer because I would help with labor, materials and friendship.

Again, thank you for your contributions.

SS

4

per Dan Lehman:

These statements might --leaned on hard-- be wanting for factual support.

By which I mean that although we
might see the knot presented in this orientation,
and even if we find by empirical observation that it is
frequently tied so,
likely we cannot be sure of the tested knot.
(IMO, you’re probably correct; but we just don’t know
for sure

My position is this:
Nobody has specifically reported it in their tests.
As in, not one single report on planet Earth specifically deals with rotation.
I am willing to go out-on-a-limb and state that the default orientation in all tests is the mid-rotation state.
Another important piece of evidence (for me) is that you have to make a definite conscious effort to align the knot in the proper manner to achieve a choke - its not a random task - its a deliberate action. You have to know what you’re doing. As such, surely a knot tester would have reported it? But no one has… (until now with yours truly - me!).

There are two ways to find proof of this assumption:

  1. Track down every person who published test reports on #1410 - and ask them directly.
  2. The other way is to do a ‘reversal of the onus of proof’ - that is, find evidence to the contrary.

Analogy with space flight: Prove that it is safe to fly or; prove that it will fail (ie prove that it is unsafe to fly).

but I do believe that this extreme case CAN be SET and hold, to rupture, if not perfectly extreme in the formation of "arc" & "loop" of its SParts, at least more biased thus than is the mid-rotation state.
I only examined 3 rotation states for #1410. I considered the mid-rotation state to be the default mindset - and the one most vulnerable to instability. Of the 2 other rotation states - only 1 was found to be stable - and this was the orientation with the rope segment lying furthest from the axis-of-tension. Of surprise to me, the orientation state with the rope segment lying closest to the axis-of-tension was transient - it reverts under load to the mid-rotation state. And this is partly why i believe that the mid-rotation state is default mindset... any inadvertent rotation with the rope closest to the axis of tension would simply transition back to the mid-rotation state.
Recall that some French folks seem to PREFER this knot (offset F8), and to have test data in which it performs better --i.e., rolls at only higher forces-- than the offset water knot!? --which surprised me, but it is a presentation that came up to my awareness in the past two years.
I would suggest that the testers are likely blissfully unaware of the effect of rotation and various dressing states for both #1410 and offset F8. Dressing state of offset F8 also has what appears to be a default mindset orientation - and no one has reported anything different. Its a topic that is devoid of any data. The reason is that climbers/canyoners/rescue technicians etc aren't knotting experts! For example, [i]most [/i]climbers wouldn't have a clue about rotation of #1410 and/or particular dressing states. They are great at the gymnastic aspects of climbing and finding ways to push a route but not so great at the technical detail about knots. They just follow routines they have learned and trusted over the years - and largely dont question them. Also, technical details about knots is considered to be the realm of academia and perhaps a little nerdy - and perceived to be 'nice to know' but not 'need to know' information. Knot book authors are partly to blame - by publishing information that is either inaccurate or simply parroted. I would suggest that it should be possible to achieve a dressing state with offset F8 that is resistant to instability (up to a point) - but for the average climber its probably not going to be practicable (ie not worth the effort or risk) when you can more easily fiddle with #1410 or simply add another binding turn (as per my variant).
ALSO DISsimilar ropes, as doing so with e.g. something pretty elastic vs. static might shed light on the mechanics of the knot. AND, the general in-practice state if of ropes dissimilar in both elasticity & size : i.e., of a 6-7mm "haul line" joining the 9-10.5mm dynamic rope.

I chose to rule out any uncontrollable variables in my test.
Joining unequal diameter ropes introduces additional variables.
It also makes it harder for other testers to try to reproduce my results (because they would have to source the same rope diameters for their tests).
Be that as it may - further down the track, I would hope that other testers do follow up work to investigate this area.
But for me, at the moment, I am staying with equal diameter ropes to reduce the number of possible variables.

I also initially focused on EN892 dynamic ropes.
I think it is also important to test EN1891 type A (low stretch) ropes in the 9.0mm class
For example: https://www.edelrid.de/en/sports/static-ropes/safety-super-ii-9-5-mm.html
In other words, we need data sets for both dynamic and low stretch (‘static’) ropes.

5

per Scott:

What I am thinking about is this: How about making a sling using two ropes and tying a comparable knot in each pair of ends. They could be oriented oppositely for that kind of comparison, even using different diameter, etc., ropes. Sort of as a tug of war, side by side. It might be useful to have the ropes go around some pulleys to aid separation and limit any influences at the anchor points.

A good idea but my issue is the amount of force required to achieve the same result compared to a linear setup is doubled.
With a doubling of force - comes increased risk - and more effort on my behalf.
I have been trying to use ropes in the 9.0mm class because that is the most commonly used (although it could in reality be anywhere from 8.5mm to 9.1mm). Larger than 9.1mm is less common for climbers using a double rope system.

Alan Lee has a more powerful test rig…he could easily managed this type of ‘round sling’ configuration. Might have to see if he is willing to give it a shot…

I wish you resided closer because I would help with labor, materials and friendship. Again, thank you for your contributions.
Thanks, much appreciated. It would be nice to have another knot god living somewhere close!
6

Hi Mark.

My bad, I thought that was your test set up initially because of the photo of Test configuration on Test page 2.
I definitely understand about the increased risk and efforts and only offered the idea, no critique at all.

Yes, I do hope that Alan or someone with a suitable rig might try and contribute. I’m not able to invest in a test cell at this time, so my contribution(s) may be considered anecdotal.
I have been considering a drop rig and doing cyclic tests to some knots, but that’s not for here in your thread.

As for “knot god”… just knotting enthusiast please.

SS

7

Mark, thank you for bringing us a most interesting problem.

I believe it is an example of rotational cogging that seems to be manifest when the primary nipping loop (choke ?) is adjacent to the entry side of the knot, yet is more stable when it is located adjacent to the WE side of the knot.

The question of course is Why?

A lovely challenge.

Derek

8

I think I am starting to get a handle on understanding this effect. Unless I am mistaken (quite possible), this effect is much like the Whatknot, where SP orientation one way locks the knot, whilst the opposite orientation allows positive rotational cogging to proceed.

With this knot, we have two parallel OH knots with the SP legs splayed 180 degrees from each other. The consequence of this is that one OH retains its open shape, while the other one is deformed into a nipping helix. The susceptibility to the rolling effect, then depends on which of the two OH knots is deformed into the nipping helix - the one sited nearest to the SP legs or the one sited furthest from the SP legs.

If the nipping helix is formed furthest from the SP legs, then under load, it bears on the OH between it and the loaded SP legs. This tends to act as jamming packing and so resists the rotational moment induced into the helix by the load.

If on the other hand the helix is immediately adjacent to the two loaded SP legs, then there is no buffer and the helix rotational torsion bears only upon the two tail ends and starts to rotate them into the belly of the knot.

Using slick cordage, the effect is quite noticeable, caused simply by rotating the knot 180 degrees relative to the line of the loaded SPs.

Hats off to Marks for spotting this and highlighting it for investigation.

Derek

9

No, you’re looking at something else, here.

The point that I raised --some years ago, but to few
paying much attention, unlike Mark-- is that the knot
can be set so that the orientation of the body to the
axis of tension can be at any of approximately 180degrees
of “rotation” --like saying when looking down upon such
a knot pulled across a table top such that the loaded
SParts run from 9:00 to 3:00 in a clockface orienter,
the tails could point to any point in this clockwise range
from about 7:00, 8-9-.. 1:00, say. (or mirror this).

And the two SParts will take respective general shapes
/curves up into the knot atop this table of:
extreme-1) left-side “arc” & right-side “loop”;
mid-range) both SParts semi-arc & turn;
extreme-2) left-side loop & right-side arc (opp. ext-1)

The “arc” structure more pries open the knot,
whereas the “loop” better surrounds & grips/nips
its enclosed parts (the tails).

(I’ve gotten pretty good at being able to simply
change torsion on the SParts of a tied knot in hand
such that I can show one extreme to one person,
then turn and --torquing appropriately-- show the
opposite extreme to another :: I envision having
two people believing that they each see the same
knot (–a philosophical issue!) argue that the other
is loony in saying it’s opposite to what they saw!)

Of the twinned parts forming this knot,
the opposed pulling of the SParts will be into
ONE until that one opens/yields and then the
twin part will also feel the prying force, which
works to flype/“roll” the knot into another version
of itself. (With the fig.8 structure, this flyping
comes in a more dramatic way, consuming more
tail material; whereas the overhands are said
to tend to just rrrroolllll around.)

And a way to preclude such capsizing is to tie off
the “choking” (initially pried into twin part) strand
with a stopper overhand --preventing the draw
of its tail to feed the widening of its choke–, or as
Mark’s further-tucked version goes, with that
further tucking; or with my two-turns (making the
overhand into a fig.9) version making
the choking turn a full round turn, which is less
accommodating of the prying-opening.

–dl*

10

Dan,

I think (again, unless I am wrong) that we are essentially in agreement. Where you have used the term "flype/“roll” ", I have used the term “positive rotational cogging”. I see your aspect of “one opens/yields” and need to look more closely at the part this is playing in allowing/ blocking the "flype/“roll” mechanism.

Such a simple knot, yet a complexity of responses to loading.

Derek

11

I don’t see “rotational” or “cogging” as apt, here.
The rotations at issue are relative to tying/dressing
(and needing the setting to sustain), not mechanics
of handling force.

Or not so much. The choking strand gets tension,
and it can thus draw out some of its tail and then
open the choke … and lead to capsizing. (Whereas
with the whatnot there is more of a flow of
material. And a flow, as asserted by Xarax, that
can lead to a shift of orientation into locking
–or not, depending upon cordage (firm & round
I think is likely to keep flowing; compressible &
flexible can shift & lock more nearly into #1408).

Yes, #1410 /Thumb, Openhand, … / Offset Water knot
is much more than meets the eye!
But, so too, is #1452, which allows of various dressings
for various effects --one can induce locking/jamming
(this might be for light enough loading so as to be
loosenable, but just secure-when-slack until …)-- or
other geometries. And likely a good many others
have subtleties of dressing & setting.

And re these sorts of things, I have retreated from my
once dismissive position vs. the “EDK-backed EDK”
(tying two of the knot adjacent, the tail-closer one
serving qua stopper knot to the inner one); I now
advocate on its behalf as a foolproof & quite easy
(to do; to remember how to do) solution to the
rap-ropes-joint problem. While I do know of many
others, such as the offset 9-Oh, offset water knot
w/stopper, & <Marc’s favored solution>.

–dl*

12

Just wondering if I could see Marks report as I get

404 - Attachment Not Found

13

You can find his fuller report/paper on offset joints
at his PACI site (please note these are copyrighted
–as is the usual password, “COPYRIGHT”!)

Look for this article:

Analysis_Offset-Joining-Analysis_Offset-Joining-Knots
which has several other interesting reports or URLinks to such.

Happy reading,
–dl*

postscript

Re
“I explained all of this to Dan but obviously
he either chooses to ignore my request or he is forgetful.”

Mostly the latter, as I struggled to recall exactly what
the issue was, too tired to search for it,
and thinking “But it’s all out there in public?” w/o
the above concerns.
This post has been edited to remove the offence!

14

Ahh yes I have that somewhere but I don’t think this is included and in particular the ring pull results

Screenshot 2024-10-27 at 9.27.52AM.png

15

Hello Aoraki,
Perhaps you are a Kiwi (NZ)?
I’ve climbed Mt Cook six times (by various routes), and Mt Aspiring twice (via Colin Todd hut and also SW ridge).
Beautiful country.

…

The conclusion reached from my limited testing is that a rotation can improve stability (and therefore delay any instability).

Dan Lehman was likely the first to advance the claim that a simple rotation of Offset overhand bend (#1410) could improve stability.
He persisted in berating and generally harassing me until I undertook load testing to prove his proposition.

As is the usual outcome of Dan’s claims or propositions, he was correct.

A simple rotation does indeed boost stability.
That is, the load threshold at which we might see instability in this offset knot is raised.

My findings were that only one type of rotation/orientation was effective.
The best link is here: https://www.paci.com.au/knots.php

I had previously asked Dan not to give the direct link to the pdf document because it bypasses the source page.
Also, it is password protected - and you need the source page to find the password.
I explained all of this to Dan but obviously he either chooses to ignore my request or he is forgetful.

per your comment:

but I don't think this is included and in particular the ring pull results
Not sure what you mean by "ring pull"? Maybe you meant "ring loading"? But, it makes no sense because #1410 is an 'end-to-end joining knot'. It isn't an 'eye knot' (aka 'loop knot').

Off-topic:
The term ‘ring loading’ is ambiguous in my view… a better description being:
transverse loading; and/or
circumferential loading (hoop stress).
Although Dan will bemoan my preference for these more accurate loading descriptors
citing his preference for ‘ring loading’.

In any case, transverse and/or circumferential loading is not what happens with #1410.

1 Like