Knot testing - Life support knots - procedures and results

Practical Knots
1

Thought I’d start a thread to set the scene for some testing that I plan to undertake soon.

I have the following equipment/materials:

5 ton Dynafor digital load cell (can be set to output kN or kg and can remember highest load achieved - 50ms sampling rate - will check this) - check
concrete driveway with screw-in eye bolts (love that word ‘eye’) - check
5.0mm diameter kernmantel construction accessory cord - check
Safety glasses (check)
Hearing protection (check)
Winch (yet to purchase - currently looking at a Tirfor or a chain/lever hoist - 2 ton capacity ought to be sufficient)

I have opted for 5.0mm diameter accessory cord (kernmantel construction) because 5.0mm cord conforms to an EN standard (cords less than 5.0mm in diameter do not conform to any EN standard). Also, by using 5.0mm cord, the force required to trigger failure is not excessive…I estimate most breaks will occur in the vicinity of 4kN give or take a bit - which gives me a margin of safety to operate in. This diameter also enables me to tie a knot with some recognisable shape and form and it will be easier to distinguish features in a photograph.

The specs on the cord is as follows:

Manufacturer: Sterling rope company USA
Construction: Kernmantel / accessory cord
Material: Nylon
Diameter: 5.0mm
MBS: 5.2kN (1169 LBS)
Certification: EN 564
Batch coding: A050AS0100 Lot #R6-092507KT
Condition: New
Purchase date: 02 Jan 2009

Test conditions:

  1. I’ll test the ultimate breaking strength of the 5.0mm kernmantel cord first - maybe 3 breaks to obtain a mean value - MBL (will be interesting to compare it to manufacturers claims)
  2. Care must be taken in tying and dressing the knot specimens
  3. Need to take care to isolate the forces with respect to the knot - will need to use large D shackles to obtain reasonable diameters at end termination anchor points
  4. Results obtained will be compared to unknotted MBL of accessory cord
  5. Will need to perform 3 tests on each knot to obtain at least some form of statistically valid sample (bearing in mind cost factors involved for each length of cord I break which eats into budget).
  6. Will pause test at various loads to photograph the progressive knot geometry - eg pause at 1kN, 2kN, 3kN, 4kN and maybe at a point momentarily before failure (without risking personal injury)
  7. Will weave small lengths of white cotton thread into sheath to act as ‘markers’.
  8. Will use a ‘tensionless hitch’ wrapped around a large diameter pin of a D shackle which will ensure stress and strain is at the knot specimen as not the end termination anchorage.

EDITS:

  1. Will tie two (2) identical knots on each specimen - one each end - so both knots are simultaneously subjected to the force (as per Dan Lehman’s recommendation).
  2. Will use restraints to ‘catch’ the explosive recoil at instant of failure.
  3. Load cell will be protected with bubble wrap, foam underlay and the recoil restraints.
  4. Photography of knots under load (and moment before failure) - It will be difficult to photograph the knots from all all sides - the knot will lie close ground level, so I will typically only be able to observe one full side. I would like to hear comment on which side is best to observe and photograph (ie front or rear aspects as per my Bowlines paper).
  5. As per Roo’s comment, yes I will only be testing a select range of knots. This is dictated by budget as I can’t break and endless range of knots. I will come back and put a list up for review. One thing I can tell you with absolute certainty is that the following knots will be on the test list:

Right hand Bowline - ABoK #1010
Yosemite Bowline - (ABoK #1010 variant with tail wrapped around eye leg and then through collar)
EBDB - (ABoK 1013 variant with 3 rope diameters encircled by the nipping loops)
EBSB (agent smith varient to DL’s EBDB)
Rosendahl bend (Zeppelin bend)
Figure 8 eye knot (ABoK #1047)
Butterfly knot - ABoK #1053 (with specific attention to loading profiles)

It is beyond my means and budget to test larger cord diameters (eg 10-11mm diameter range). EDIT: Larger diameters will require larger forces and hence higher risk. Also, I can’t afford to purchase EN 891 dynamic rope for destructive testing (there is no opportunity for me to ask around for older samples either… and in any case, I want to test new cordage with known batch coding).

I am of the view that thinner cords will still yield a useful relationship between knotted to unknotted % ratio. Provided the same accessory cord is used with the same end terminations and same ambient temp/humidity conditions, the results gathered should provide a good baseline of data to compare knots.

At this stage, the only thing holding me up is the mechanical winch - which I’ll purchase tomorrow or the next day…

agent smith

2

Safety overlay or framework for catching ‘exploding’ rope recoil forces might be in order, along with specific protection for mea$uring device.

3

If this is related to your earlier questions, you can cut your costs dramatically if you only test the knots that pass your security requirements and ease-of-tying or ease-of-remembering requirements.

In other words, it may be best to start with the low-cost tests. :wink:

4

And if this goes well, maybe the Guild and members would like to connive some time for well tested knot testing and explorations.

SS

5

It’s interesting to see in Richards’s tests for rope strength that in each of the thicker ropes
there was greater variance/Std.Dev. than in some of the knots’ data!? The two really low
figures are for the 12.5mm low-elongation & Grapevine and the 7mm unknotted.

2. Care must be taken in tying and dressing the knot specimens
Maybe not so much, in that you should have photos of each specimen; given that, some variance in structure might be seen to correlate w/strength--though, yes, that begs the question of reliability of the data given the implicit 1-case-only sampling. OTOH, if all cases are identical (magic wand set, say), then one can question the significance of the reliable data for other-tied/-dressed like knots. (You can't win!)
5. Will need to perform 3 tests on each knot to obtain at least some form of statistically valid sample (bearing in mind cost factors involved for each length of cord I break which eats into budget). [b]Dan Lehman suggested 2 tests should be adequate

??? I said no such thing, unless you do some implicit logic. What I DID say was
that you should have TWO specimen knots per material sample–one at each end,
in the case if eyeknots–; in this way, the break at say 1,000# force will show that
BOTH knots held 999# force (and the survivor might be regarded to some small
measure >1000# strong (this I dubbed “Speculative Statistics”)).
Conceivably, with adequate material length, you might be able to run say your
3 rope-piece tests, break 3 of 3x2 knots, and then re-test with the 3 surviving knots
(to see how strength was affected by such severe loading–pretty much academic,
I think, since such material should be retired).

6. Will pause test at various loads to photograph the progressive knot geometry - eg pause at 1kN, 2kN, 3kN, 4kN and maybe at a point momentarily before failure

While this stands in contrast to standard continuous slow-pull loading, it should
be okay for the purposes of testing variance between knots (and slow-pull isn’t
all so assuredly like what practice might throw at a knot).

7. Will weave small lengths of white cotton thread into sheath to act as 'markers'.
Having some different color in the middle of set could help you quickly count/pinpoint the break ("ah, 2nd mark away from red thread going away from Spart"). The camera will show us where the marks lay in the loaded knot. This is REALLY intriguing!!
It is [b]beyond [/b]my means and budget to test larger cord diameters (eg 10-11mm diameter range).

Not so: it is … if trying to do the full, multiple-repetitions set, agreed; but you might care
to break some bona fide DYNAMIC rope (get some donated old ends chopped from a long
length, which some climbers do as a way to extend the usable life of a rope (since it is at
the ends that the most degradation occurs), or a complete old rope–you should have some
contacts who can pitch in!). The idea would be to show, via camera, what likeness, what
differences obtain in the actual target material–such as a tendency to deform more in some
way. We might be able to project some results to be checked by sampling. (“Hmmm, since
it seems that dynamic rope will do , … , then THIS variation of the Bowline should
improve strength by <…>.”)
So, you might not give up the idea of doing SOME breaking of actual climbing rope,
which should occur with range of your set-up, yes (around 3,500# judging by Richards’s
results for 10.5mm rope; less for used & thinner rope).

–dl*

6

The comparative test rig I set up, used a hydraulic jack as the force source. The advantage (apart from low cost) was that there was virtually no elasticity in the jack, so when the sample broke, there was virtually no jack reaction.

My test cord was 3mm polyester - 1kN MBL, yet even at this low level of breaking strain, the energy stored in the sample was huge and the stronger knots would literally explode off the rig with a puff of smoke. I ended up working with a long camera lens and a long jack handle to keep me at a safe distance. I was also at right angles to the trajectory of the fragments.

If you plan to use a chain hoist, how do you plan to stay out of the path of reaction?

Advice gleaned from my experience…

Keep the sample as short as possible, this will reduce the energy stored in the sample.
Make the loop fixing point as realistic a diameter as possible commensurate with intended use.
Make the SP fixing at least 5 turns onto a large diameter non rotating bar and clamp the free end to the bar, not to the specimen
Use two tells, 1cm apart and photograph the knot under pre break tension (you will hear the cord start to ping and crackle as it nears rupture), this will allow reasonably accurate measurement of failure point.
Photograph the fragments including tells to allow determination of rupture point.

Take great care to document the exact dressing of each knot tested, your photography is excellent, so take several perspective shots including end on.

I doubt it will be much use to you, but here is the protocol developed for comparative testing (unfortunately funding for the test cord fell through)
http://igktworkshop.pbwiki.com/Testing+Protocol

It is probable that you have never been up close to a modern rope failure - please be exceptionally careful, the energies released at rupture are frightening.
(you can easily store 1kJ in the cord, that’s enough energy to sling a 10g ‘magnum’ bullet at 450 m/s - faster than the speed of sound !!).

Derek

7

Having obtained a static loading failure value, you might consider exposing the same knot to a dynamic load of a similar magnitude (or say the 50% failure value)

500kg load dropped 1m onto the test sample would simulate 5kN loading dynamically.

Be prepared for the samples to fail at a far lower loading though.

Derek

8

Maybe I’m missing something, but without knowing the elasticity of the system, how do you know 500 kg dropped 1m yields 5000 N?

9

Yes, what you are missing is a brain in gear - mine, not in gear that is.

Of course, you are right roo. Dropping 500kg by 1m gives us an energy of 4,900 Joules (not Newtons force) and without knowing the dynamic elasticity, we cannot know the forces involved. For that we would have to integrate the area under the Force/elongation curve for the system (which of course would be available from the testing carried out by AS if he records extension along with force.

Mea culpa - what a way to start 2009 - 0/10, must try harder…

Derek

10

Strictly speaking, I think that there is some potential bias from this, in terms of
how much material there is to allow some fibres to elongate and distribute load,
but otherwise I imagine that circumstances of the device & cost will so dictate shortness.

Make the loop fixing point as realistic a diameter as possible commensurate with intended use.
Not sure about this, as the break will come at the knot, and is unlikely altered by other than unusually wide pins (or one so sharp as to cut). I'm going to guess that 'biners might serve, or the like. --but not harness tie-in loops.
Make the SP fixing at least 5 turns onto a large diameter
For the tensile check? For per-specimen-knots, they are to be knotted at both ends (which maybe saves some material, and otherwise yields behavior on TWO knot-tokens per exercise of device).
Use two tells, 1cm apart and photograph the knot under pre break tension (you will hear the cord start to ping and crackle as it nears rupture), this will allow reasonably accurate measurement of failure point. Photograph the fragments including tells to allow determination of rupture point.

Not sure of what’s asked, here. But coming to photograph material so near rupture
that it’s sounding–and I’ve read accounts of there being no such warning (a difference
betwee the Offset Ring Bend & (regular) Ring (“Water”) Bend in tape, I recall)–seems
like tempting fate. And I suspect that beyond a certain highly loaded (say, 45% tensile?)
point, the change to geometry is both minimal and well indicated by the recent change
to that point–so the report might remark “and it continued to … very slightly up to break”.

AND, with TWO knots per test, one won’t know which one is going to burst first
–so (1) won’t know where to train a camera, and (2) must protect against the
burst from both knots.

Take great care to document the exact dressing of each knot tested, your photography is excellent, so take several perspective shots including end on.

One efficient way to do this is to tie several specimens, label them, and have them all in
the pictures together–perhaps even with some bearing of a load/weight so as to have them
in tension. Saves on photography effort (your doing, & later reviewing, printing).
It might be worthwhile to deliberately vary the dressing on each 3rd specimen set,
leaving the collar a little looser (which can be seen to model tying in firmer or more
dynamic rope or more quickly/carelessly or … ).

And you might budget in saving some material for special re-testing if something
strikes you as odd/interesting and you want to have a further look, or to try to alter
some result by making a particular change, as suggested by … .

[b]It is probable that you have never been up close to a modern rope failure - [u]please be exceptionally careful, the energies released at rupture are frightening.[/u] (you can easily store 1kJ in the cord, that's enough energy to sling a 10g 'magnum' bullet at 450 m/s - faster than the speed of sound !!).[/b]

Worth simply reiterating in bold. (I have also not been … .)

Thanks,
–dl*

11

Since the ultimate breaking strength of the unknotted cord is the baseline for all the other tests, it should be determined to more accuracy with more breaks than the other tests.

I once read an Alcoa aluminum handbook. It said that engineers might not be concerned about the average strength of an alloy. Instead, they might be interested in the minimum strength of an alloy. The minimum strength of a knot should be reported.

12

Could I also suggest that you test the cord failure periodically throughout the knot testing phase. This will show any drift in the testing system and even potentially allow you to compensate if it is shown to be present.

As you will be comparing your results against manufacturers claims, could I ask how you intend to calibrate your system. If you do not calibrate, then you will not be able to make a comparison with the manufacturers data and your results will have to be declared as % of MBL(measured).

As this is your own project, rope, equipment and work, could I ask if you intend to share the results with the Forum?

The minimum strength issue is an important one and although I believe comparison should be made on mean performance, the presence of low individual values are important and deserve much greater attention than simply being reported, or as in some statistical approaches - deleted.

A low figure will mean either the knot was tied happen stance on a weaker portion of the rope, or more likely, the construction of the knot created an adverse set of forces, causing an earlier failure. It is these lows that have to potential to direct us to specifics in dressing which magnify the weakening effect of the knots structure. If you find ‘spurious’ lows, they deserve hyper attention in the hope we might find the cause of the early failure, especially attention to the point and structure of the failure (that is structure of the debris, not the knot).

If we can identify the causes of critical weakness, then we can advise instructors to specifically include training to check to ensure the absence of these weaknesses.

I would have hoped that the Guild might have helped by funding the test cord to be used, but if the response to such a request is as forthcoming as the request for help to fund the 3mm cord for the trials I proposed, you will be a long time waiting.

Derek

13

and

Could I also suggest that you test the cord failure periodically throughout the knot testing phase. This will show any drift in the testing system and even potentially allow you to compensate if it is shown to be present.

As you will be comparing your results against manufacturers claims, could I ask how you intend to calibrate your system.

imply a measure of confidence/significance to this home-brew testing that strikes me
as way beyond the pale in having any value.
For one, how do you know that the data from elsewhere that you presume to make
comparisons with has any of this rigor and exactitude!? The UIAA had an big issue
with differing results from different test labs in its certifying tests; I suspect that those
places had more resources in expertise & equipment, but still ended up with challenged
results, requiring official redress.

Take a look at Dave Richards’s results: I’d think that the least variation (he did 5 tests per item)
would occur for the material (tensile strength) --one might reason that by logic anything that
depends on the material could be no better(!)–, and yet his smallest standard deviations
are for the Grapevine in low-elongation 12.5mm rope, for tensile in the 7mm cord, and
nothing so small in the dynamic rope, but (single)Fisherman’s knot & and ONE tying
–but not, should be identical(!!!) --the other tying (TIB vs. “rethreaded”) Fig.8 eyeknot !?

.:. Asking for some great precision of the tensile begs the question in any case as to
how valuable that is if specimen tests are less demanding, and comparisons to other
date again to their rigor, and in any case, their different material.
Agent_Smith isn’t about to build a lawsuit case to claim fraud against Sterling, is he!?

Oddly, I was thinking just the opposite: that, given the hoped-for consistency of the cord,
and the (we hope) smooth, uniform, terminal securing of the line for this test,
that a SINGLE test of it should suffice.
Frankly, the real comparisons of interest are of Knot-A to Knot-B to … ,
and –irrespective of actual value/strength(!)– the information about how/where knots
break! Indeed, if this were ALL that he reported, along with the images for us to understand
what the knot was–in terms of exact geometry–that would go much further in advancing
the “science of knotting, its study and practice” than yet one more set of not-really-possible
-to-understand batch of strength figures.

Really, the comparisons of prime interest of knots to each other, and one could even
just assume some out-of-the-air value to use as a baseline for producing %-of-X
values to compare. Think about this comparison, Bob: you have the material tested
for three days and 100 individual tests, and a MBL … etc.; then for each knot you have
but ONE break; contrasted with the opposite scenario, rigor per knot, and one test of
the material. I think that the choice is obvoiusly for the latter; very easy to take the
given results and adjust them by reason or later data on the material. Especially when
dealing with a material that should be pretty uniform in strength.

If you’re going for the original THREE breaks, I like Derek’s suggestion to distribute
the testing–for evrything, actually. Perhaps best is to set up a trio of full tests, and
within each test set, vary the ordering of specimens tested, which includes the
break of pure cord (and, if drift is much a risk, and you want an arguable surer
value for it, make the pure-cord tests all towards the center of your sets.

–dl*

14

On A, you should have a specimen pair in which the end is dressed-set to cross UNDER
itself on exit re the Front perspective–a position the draw of the SPart will then bring,
under load, more into the position that is typically shown for the end to start (from which
an end is brought up & a little to the side).
B, you see, positions the end (to the opposite direction) and holds it there.

On E, you might have on specimen pair in which you deliberately leave the knot
a little looser than the what you might expect setting in good practice to do–this
can case can be seen to model stiff rope or careless/quick tying, and should give
a geometry tending towards that of the Bowline.
However, I find the inclusion of this dubious in light of any advocacy for use in
the kernmantle [sic] worlds, as for joining slings, the Grapevine will rule, and
for joining abseil ropes, an offset bend is wanted; what does this leave? This
symmetric knot will work with dissimilar ropes pretty well, but I just don’t see
much market for it in your presumed target audience. (Still, at least a carefully
reported break test & photos will advance understanding … !)

On F, there are two prime ways of loading the knot. To my mind, if one is going
to load the strand that runs nearer around the eye, and that will press INTO its
shadow/parallel/twin partner, you should set the knot firmly by loading the end,
and by loading the eye in opposition to the nub (i.e., grasp knot, pull eye)
–the idea is to put more of a curved path for the SPart to take en route to
its tight turn around the eye legs–;
if loading the other strand (which will pull away from its twin, naturally), the
best setting might be to tension SPart against nub, to try to ease the force
that the eye legs will deliver to the SPart, which in this position seems to
have a broader curve around the body.

On G, well, there are two potential SParts here, too–although they depart
not in parallel/tandem but in opposite directions. You should see that W&Magowan
give specific orientation instructions, having the eye legs crossing in the nub,
which is in contrast to most of the images of it that I’ve seen modern books
give (having not appreciated the variety of orientations possible). So, here
you have an orientation of knot body/eye-legs, and THEN the question of
which end to load–already there are 4-6 distinct cases (two crossings & none,
times two ends). For this reason, I do not like the choice of this. We have now
no idea how the knot was tested elsewhere, which makes any comparison
impossible. (Well, yeah, this is somewhat true for everything, alas.)

And of course I’m very disappointed to not see the Janus Cowboy bowline
in the set, as it’s the simplest and best-looking of the bowline variations so far.
And it’s a knot that I’ve examined carefully, and so advanced, especially on
account of how the curvature of the SPart looks GOOD/strong.

It is [b]beyond [/b]my means and budget to test larger cord diameters (eg 10-11mm diameter range). [b]EDIT:[/b] Larger diameters will require larger forces and hence higher risk.
Again, I will remark: [b]the "test" of such a rope [i]need not be [u]to destruction[/u],[/i][/b] but only to a known, significant force, at which photo documentation can be used to compare the target rope (dynamic) with like-loaded (%-of-tensile) 7mm rope, and with that information, maybe some projections can be made. --same bit of rope (11mm) can be used repeatedly, thus; and just one test per knot, if ... . Again, it's a matter of gaugeing how the greater elasticity enables greater distortion and so affects effective knot geometry. Loading climbing rope to 500-1000# say will produce meaningful results; in terms of knot security on "cyclic" loading by sport falls, this is the sort of loads you'll see, and you can see if anything changes re security. .:. so, no greater system loads; no cost of rope, other than having ONE to use (at forces you are likely putting on it or are willing for it to endure anyway, in use).
(there is no opportunity for me to ask around for older samples either... and in any case, I want to test [i]new [/i]cordage with [i]known [/i]batch coding).

A simple ping to a climbing forum could bring ONE rope fairly quickly, I think.
–or just a worn end of one, adequate.

–dl*

15

If I have understood your comments correctly Dan, would I be correct with paraphrasing them as – ‘Home Brew is not professional so don’t bother with attempting to achieve anything of value’, and, ‘If you publish anything of value, it is likely to end you up in court contesting the results’.

If that is what you meant, then I would have to disagree.

The level of professionalism present in many hobby organisations is generally well in advance of any commercial organisation because hobbyists do not have to turn a profit or report to a Board of Directors, consequently, profit does not get in the way of progress. Just because something is ‘Home Brew’, does not automatically mean it is shoddy or worthless. A job may be being performed on an extremely tight budget, but the use of intelligence and ingenuity can yield breakthroughs that commercial organisations would never dare set their sights at.

From a different perspective, many people have run strength testing studies, and how many times have members of the IGKT (or at least this Forum) criticised the reports for failing to cover, identify or document issues critical to a knots performance. Probably the greatest gathering of knotting expertise is contained within the IGKT and is accessible through the IGKT and this Forum. There is therefore a strong argument that better trials will be run where there is a large proportion of knotting expertise being put into the project rather than a high level of lab technician expertise. I would consider my Analytical skills at knowing HOW to measure something as being far less important than knowing WHAT I should be measuring.

Finally the issue of having results contested in court. The better structured a trial is and the more detail that goes into both the setup and the interpretation, then the less chance there is of someone misunderstanding or disagreeing with your findings, and even less chance of them winning. Equally, rule 101 - never sue and empty pocket, tends to make it a non starter to take out an action against amateurs scattered around the world.

If these tests are not approached with the attitude of making the results meaningful, then what is the point is doing them? I can’t see Agent Smith going to all this work and then not publishing them ??

Derek

16

That is a misreading. I pointed to exactly the “at least” considerable value,
if documentation (images & marker threads) enable us to narrow down the
region of breakage; and I also believe that comparative strength values
will be informative (though we should remind ourselves that such values
can vary per situation). I have no fear that there will be a legal issue.

My reaction was to those bits of urging for exactitude I think is out of proportion
in effort and even practicability for this home-brew situation. Especially as I
read the protocol of there being interruptions to some hand-cranked? winching
for the sake of photography, I think that there will be siginficant deviations from
the industry standard (but perhaps not so entirely valuable/worthwhile) test
protocol of steady pull.

How does the measure (dynomometer) get calibrated? Well, okay, let’s see
about that (although if it understates forces uniformly, won’t the results be
pretty much useful, in terms of percentages?). But going to some extra effort
on figuring tensile strength, but doing less per knot specimen, strikes me as
backwards–as I stated (and gave a choice of extreme cases, for illustration).

(E.g., suppose some tests were later done at Sterling itself, on some of the knots
tested by Agent_Smith, and the results were comparable; but Sterling challenged
the A_S-measured tensile, and could suggest why those values differed. We
might come to accept the different 100% figure, and then it would be a simple
matter to re-figure the % breaks per knot, trusting that at least they had been
rightly recorded. (E.g., if A_S used a too-narrow anchor point for tensile or
something, and so had a lower value.))

–dl*

17

Am back on the case…

I have purchased a suitable hand operated winch and installed a pair of eye bolts into the concrete floor of my garage.

I completed 4 shake-out tests today to confirm the following:

  1. That the test rig as a whole system actually works
  2. That I can avoid personal injury
  3. That one (1) meter will be a practical length of cord to work with
  4. That I can observe and measure events as they occur

Anyhow, during the shake-out tests, I was able to confirm the following observations (Knot tested was ABoK #1010 - Right hand Bowline):

Note: The specs on the cord is as follows:

Manufacturer: Sterling rope company USA
Construction: Kernmantel / accessory cord
Material: Nylon
Diameter: 5.0mm
MBS: 5.2kN (1169 LBS)
Certification: EN 564
Batch coding: A050AS0100 Lot #R6-092507KT
Condition: New
Purchase date: 02 Jan 2009

Procedure:
I tied two identical #1010 Bowline knots within the one (1) meter length of Stirling accessory cord. I gave due attention to shape/form/dressing of both knots. The knots were cinched tight by hand. 100mm tails were left on each knot. The eye of one bowline was attached directly to an end termination eye bolt (5mm radius - simulating a carabiner). The eye of the other Bowline was attached to the load cell via a D shackle with 5mm radius. Once all connection were confirmed, a suppression system was attached to prevent violent recoil at the instant of knot failure.

  1. The Stirling accessory cord internal core failed at approximately 3.90kN - leaving the sheath intact (there was a very distinct ‘popping’ sound when the core yielded).
  2. Final failure occurred when the external sheath yielded at approximately 4.12kN (for the initial core: 3.90/5.20 = 75% and for the final yield point: 4.12/5.20 = 79%)
  3. For reasons I do not fully understand, failure always occurred at the Bowline attached to the end termination eye bolt (and not the other Bowline attached to the load cell).
  4. The precise failure point was not determined - as I was not specifically examining this phenomena. To more accurately determine the local region of knot failure, it will be necessary to use small tracer threads woven through the sheath at strategic positions within the knot structure. This became obvious when observing the amount of stretch occurring with increasing load. Dan Lehman already advised that this would be necessary earlier and in many posts over the past few years. I can confirm that he is 100% correct in stipulating this requirement. It may even be necessary to video the process!
  5. The recoil effect at the instant of failure was not as violent as I anticipated - I believe that this was due to the fact that the internal core yielded first, followed shortly thereafter by the outer sheath.
  6. I had some trouble with the load cell - it uses 3 x AA batteries (1.5v) - and the recoil effect after the initial ‘pop’ of the internal core induces enough of a ‘jolt’ to dislodge the batteries from their cradle position! This happened on 2 occasions - so I’ll need to tape the battery lid in position using gray tape. As soon as the batteries were dislodged, it interrupted power supply…and hence loss of digital output.

At this point in time, I am somewhat concerned by the mode of failure in the 5.0mm diameter accessory cord. With the core yielding first, it will be very difficult to determine the precise region of failure. I may need to use thicker cord - eg 6.0mm diameter cord…

I am open to suggestions/comments/advice…

agent smith

EDIT: I just completed a tensile strength test to determine the ultimate breaking strength of the 5.0mm diameter cord. I reached a peak force of 4.98kN when the internal core failed (with a distinct ‘pop’) followed shortly thereafter by the external sheath - also at 4.98kN. The load cell sampling rate is 40hz (the LCD screen updates 40 times per second when the load cell is set to read the peak force achieved). The end termination pin diameter used was 15.0mm.

Observations: The 4.98kN result is unreliable.
Reason: The end termination anchor pin diameter needs to be at least 20.0mm (15.0mm is too small). Cord failure occurred at a region where the cord immediately exited from the anchor pin. A tensionless hitch was used to secure the cord to the pin - 5 wraps were used and then the cord was terminated away from the ‘tensionless wraps’ to a shackle using ABoK #1047. Since the cord failed at a region close to, or at the pin itself, I conclude that the 7.5mm radius of the pin (ie 15mm diameter) caused undue stress and strain on the cord resulting in premature failure.

However, 4.98kN is not too far off Stirlings reported 5.2kN MBS (a difference of only about 20kg) - so I think a move to a 20mm diameter anchor pin ought to be sufficient.

This means I will have to go out and purchase bigger D shackles with larger pins of 20mm diameter - bugger, more money!

18

:smiley: I can only imagine that your misspelling has the rope manufacturer shaken at being ‘Stir’'d!

I’ll also add that all “cores” are “internal” (redundancy alert).

Now, about why “failure always occurred at the Bowline attached to the end termination eye bolt
(and not the other Bowline attached to the load cell),” I’ll toss out a conjecture: that there is a
difference in friction of the metal terminal points, and so the exact load on the eye legs differs
between the two knots–note that I’m not positing that an imbalance (at the more frictive end)
is weaker: maybe it’s what adds some strength!? Again, this is a seeming possible reason!?

Btw, what IS the core/kern, structure-wise: a braid, or a collection of (mixed-lay) laid strands?

A tensionless hitch was used to secure the cord to the pin -[b] 5 wraps were used[/b] and then the cord was terminated away from the 'tensionless wraps' to a shackle using ABoK #1047

  1. I think that the T.Hitch should have FEWER wraps–to limit the amount of (potential) movement
    of material around the pin–it STILL should be quite effective at reducing force transmitted to the
    ultimate termination point such that rupture occurs where you want it. --even ONE turn!?

  2. In what condition did you fine the Fig.8 eyeknot used as ultimate termination?
    –could you untie it? (seeing how much load it got through 5 tightish wraps)

  3. Rather than involving the necessarily significant amount of material consumed by
    an eye-knot, can you use something brief, such as a noose-hitch w/#1821form of the
    Overhand (looking at #1821, connect the SPart to the RIGHT side of the ring, if you
    will–as though the ring is in fact the SPart coming to turn around the anchor ping
    from the left, clockwise around and flowing into 1821 around itself. (This appeals
    to me for the draw on the wrapped SPart into the knot of this noose hitch
    will pull the end ever tighter, and somewhat pad the noose-SPart from the
    tight squeeze of the tightening knot. (Some tests of the “barrel hitch”, where it’s
    a Strangle knot (Dbl.Oh.) noose, showed the noose-SPart breaking at entry into
    the knot (and not, e.g., at the 10mm turn around the 'biner), so I’m looking for a
    diminution of compression at this point.)

  4. “I am somewhat concerned by the mode of failure in the 5.0mm diameter accessory cord.”

You indeed might acquire some other cord. But I think that if you have rigorous
test results from 5mm, and then do some sampling with say 6mm (and, as I
urged, at least l00k at the effective geometry of actual dynamic rope at loads
we can map into the 5mm results/images),
you’ll have provided a good basis for understanding what’s going on.

Perhaps the failure mode is attributable to peculiarities with the construction
of the 5mm cord, different from 6 or 7mm even though its roughly of the same
sort (in terms of braid in the mantle, and make-up (which is … ?) of the kern),
and not noticeably different as braid is from twisted – something that might
result from percentage of mantle to kern, or particular tightness of weave!?
(speculations, these)

I’m having similar doubts about my thought to use beached conch-pot warp,
worrying that its well-weathered, highly frictive mantle will be weakened much
by friction which will skew results. (Although also noting that IF one is in fact
working with just such cord, it is THAT behavior that matters, and not that from
pristine cordage.) Indeed, I recently BROKE A BOWLINE in aged CoExOlefin
(“CEO” I’ll coin! (“coextruded”, btw)) 8/16-strand rope, with rather moderate
force from a (crummy) pulley: it broke at initial entry point (as per Chisholm),
which I think is HIGHLY improbable for decent cordage; one can at times see
how much friction the surface fibres are fighting at such bend points through
a bight of surrounding material; but new/slick rope, will pull through and by
enduring greater force change the geometry favorably at that point.

–dl*

19

Suggestion - don’t increase the diameter of the MBS test pin from 15mm, DECREASE it to the same value (i.e. 10mm) that you will be using to test the knots. If it breaks at the shackles, then so be it, it is after all the weakest point in the system. Then you will be comparing unknotted cord with knotted cord with everything else equal, so you will be comparing apples with apples.

POI, what makes you state “Observations: The 4.98kN result is unreliable.”?

Derek

20

Reporting in here with some early test results for the IGKT forum…

This post will be updated on a regular basis as test results become available.

Cord specifications:
Manufacturer: Sterling rope company USA
Website: http://www.sterlingrope.com/supportingdocs/2008_climbing_cat_technicalspecs.pdf (look under ‘Accessory & Prusik cord’)
Construction: Kernmantel / accessory cord
Material: Nylon
Diameter: 5.0mm
MBS: 5.2kN (1169 LBS)
Certification: EN 564
Batch coding: A050AS0100 Lot #R6-092507KT
Condition: New
Purchase date: 02 Jan 2009

Load cell:
Model: Dynafor LLX5
Capacity: 5 ton
Sampling rate: Forty (40) times per second when set to record and display peak force achieved. Four (4) times per second when set to read current force.
Serial No. E04073
Calibrated by: Tractel S.A.S. (Calibration certificate No. 0414607)
Last calibration date: 25 April 2004

Accessory cord ultimate strength calibration test series:

MBS TEST #1
Test date: 10 Jan 2009
Knot: None
Cord length: 1.5m
Method of fixing: Each end of the cord was secured with a 5 wrap tensionless hitch around a 20mm pin on a bow shackle. The tail ends were then fed through a hole in the pin of another adjoining shackle using ABoK #515 as a stopper knot.
Peak force recorded at failure: 5.1 kN (a discrepancy of only 0.1kN from Sterling’s reported MBS of 5.2kN)

Comment: The #515 stopper knots had cinched very tight after the test. I could not undo these knots by hand alone. They were easily untied with the aid of pointy nose pliers.

MBS TEST #2
Test date: 12 Jan 2009
Knot: None
Cord length: 1.5m
Method of fixing: Each end of the cord was secured with a 6 wrap tensionless hitch around a 20mm pin on a bow shackle. The tail ends were then fed through a hole in the pin of another adjoining shackle using ABoK #515 as a stopper knot.
Peak force recorded at failure: 5.0 kN (a discrepancy of 0.2kN from Sterling’s reported MBS of 5.2kN)

Comment: I managed to fit 6 wraps around the 20mm pin of the bow shackle.
Note: The #515 stopper knots were able to be untied by hand with a small amount of effort.

MBS TEST #3
Test date: 12 Jan 2009
Knot: None
Cord length: 1.5m
Method of fixing: Each end of the cord was secured with a 6 wrap tensionless hitch around a 20mm pin on a bow shackle. The tail ends were then fed through a hole in the pin of another adjoining shackle using ABoK #515 as a stopper knot.
Peak force recorded at failure: 4.98 kN (a discrepancy of 0.22kN from Sterling’s reported MBS of 5.2kN)

Comment: I managed to fit 6 wraps around the 20mm pin of the bow shackle.
Note: The #515 stopper knots were able to be untied by hand with a small amount of effort.

BOWLINE ABoK #1010 TEST SERIES:

TEST #1:
Test date: 10 Jan 2009
Knot: ABoK #1010 (Right hand Bowline)
Cord length: 1.0m
2 x identical #1010 Bowlines were tied - to form an eye at each end of the cord
Tails: 100mm on each knot
Dressing: Attention given to shape and form - pulled tight by hand only
Peak force recorded at failure: 3.98 kN
End termination anchor pin diameters: 20mm

Observations:
The knot was cinched tight using hand force only. A single ‘tracer’ thread of white cotton was weaved through the sheath to mark the exact mid point of the collar on each knot. Under load, it was observed that the collar position did not move. The collar remained in a fixed position relative to the other parts of the knot. Examination of the specimen after failure indicated that the break/rupture did not occur at the collar. The precise break position could not be determined - more tracer threads need to be woven into the thread to determine the region of failure/rupture.

TEST #2:
Test date: 10 Jan 2009
Knot: ABoK #1010 (Right hand Bowline)
Cord length: 1.0m
2 x identical #1010 Bowlines were tied - to form an eye at each end of the cord
Tails: 100mm on each knot
Dressing: Attention given to shape and form - pulled tight by hand initially, then loaded to 1kN before tracer threads were inserted.
Peak force recorded at failure: 3.78 kN
End termination anchor pin diameters: 20mm

Observations:

Only one of the #1010 knots failed. On this occasion, the knot that failed was closest to the end termination anchor point (and not the knot immediately adjacent to the load cell).
The test specimen was loaded to 1.0kN then removed from the test bed. Three tracer threads were woven through the sheath in strategic positions in order to more accurately determine the point of failure/rupture.
*White tracer thread was woven at the collar
*Red tracer thread was woven at the crossing point (G spot - DerekSmith)
*Blue tracer thread was woven at a position on the loop where it passes over the bight (when viewed from the ‘front’ side)

Examination of the specimen after failure indicated that the BLUE tracer thread was missing. I could locate the white and red tracer threads, but NOT the blue. This indicated that failure/rupture was occurring at a region on and/or around the blue tracer thread. A further test with a fourth tracer thread will be required to more accurately home in on the region of failure/rupture.

The fact that the other #1010 knot remained intact enabled location and comparison of the tracer thread positions.

TEST #3:
Test date: 10 Jan 2009
Knot: ABoK #1010 (Right hand Bowline)
Cord length: 1.0m
2 x identical #1010 Bowlines were tied - to form an eye at each end of the cord
Tails: 100mm on each knot
Dressing: Attention given to shape and form - pulled tight by hand initially, then loaded to 1kN before tracer threads were inserted.
Peak force recorded at failure: 3.76 kN
End termination anchor pin diameters: 20mm

Observations:
Only one of the #1010 knots failed. On this occasion, the knot that failed was closer to the load cell. The test specimen was loaded to 1.0kN then removed from the test bed. Four tracer threads were woven through the sheath in strategic positions in order to more accurately determine the point of failure/rupture.
*White tracer thread was woven at the collar
*Red tracer thread was woven at the crossing point (G spot - DerekSmith)
*Blue tracer thread was woven at a position on the loop where it passes over the bight (when viewed from the ‘front’ side)

*Gray thread was woven at a position half-way between the blue and red thread.

Examination of the specimen after failure indicated that the BLUE tracer thread was again missing. I could locate the white , red and gray tracer threads, but NOT the blue. This indicated that failure/rupture was occurring at a region at, or immediately around the blue tracer thread.

The fact that the other #1010 knot remained intact enabled location and comparison of the tracer thread positions.

Further tests will be carried out to gather more data.

Thanks to Dan Lehman in particular for insisting that tracer threads be used and that 2 identical knots should be tied in each test specimen. Also thanks to DerekSmith for his suggestions.

I will post photos soon.

agent smith

EDIT: I would like to add the comment that this whole testing procedure is very time consuming and labor intensive. In particular, weaving the tracer threads through the sheath in strategic positions is tedious and testing both my eye sight and patience… I hope its worth the effort Dan!