Author Topic: The Dave RICHARDS 2004/5 Knots Test Report  (Read 228 times)

Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 3764
The Dave RICHARDS 2004/5 Knots Test Report
« on: June 28, 2019, 09:55:37 PM »
Today I happened to make another examination of this
moderately detailed test report, and am pleased to see
that it is again accessible via the U.S. National Speleological
Society's caves.org site, here:
http://caves.org/section/vertical/nh/50/knotrope-hold.html

Although I submitted comments about errors in this report
to the NSS CaveChat forum, I don't think that any corrections
were made to the original document.  In short, the main error
is that the bar graphs for the 12.5mm ("static") & 10.5mm
("dynamic") ropes are swapped/mis-labeled each for the other.

That the tables of raw data & associated Notes are correct
comes from the given values themselves --they are only
applicable, in the one case, to the strongest rope; and the
others must come by elimination.  Therefore, it can be
seen that the bar graphs are wrong.

Additionally --and surprisingly(!)-- the final Comments
in which a few results are highlighted are wrong in the
cases about the butterfly & fig.8 eye knots.  (Esp. re the
former, which has one stand-out value that is (a) only
1 of 2 >80% and (b) 9%-points higher for the given
knot & rope than in the other two ropes, its surprising
that the rope used would be mistaken --one would think
that it would've been remarked at and remembered!

I will re-print my post to cavechat.org, below, which
repeats w/maybe more detail the above notes.

In short, this seems to be a good test & results to have
for reference!

--dl*
====

 - - - - - - - - - -
Quote
My take on this all is this.  Richards has done one more set of tests
similar to ones done by others, and has data worth tossing into the
knowledge pot for consideration.  We need to determine what is the
correct labeling for the data tables & graphs --my surmise is that the
former are correct, and thus three graphs (re ropes S & D) need to
be corrected, swapping labels, and much of his "Comments" need
removing (they don't add anything).  For the interim to a corrected
paper, this can be effected by a note to the correction.

Sadly missing from this, and from every other test report  that I've
ever seen, is a clear indication of what exact knot geometry met
the test --such as photographs of the knots when tied &  set, and
then when, say, 20% & 50% tensile strength force is achieved.

With the Fig.8 eye knot, we don't even know which of its tails
is loaded; with the butterfly eye knot, which is asymmetric, in
addition to having some various dressings, we don't know which
tail is loaded, either.  We might even wonder if Richards used
same-side vs. opposite-side sheet bends,** although that is the
more commonly presented version, or concordant vs. discordant
(same- or opposite-handedness) fisherman's knots.  But, as noted,
this failing is common to test reports; the aforementioned Lyon
report at least gives voice to some of these differences,
although it is less than perspicuous about exact tying,
with some confusing expressions.
[** 2019 insert : His Notes about sheet-bends slippage
  suggests that it might have been the opposite-sides
  (ends/SParts) version, which is known to slip, as per
  CLDay's 1935 Sailors' Knots comments,
  for manila rope.]

It is interesting to read Richards's observation that the butterfly
knot was "almost impossible to untie after a significant load of
1,000# was applied," as one rationale for using the knot is its
supposed ease of untying.  It makes one wonder which version
was used, and which tail was loaded --it is an *a*symmetric knot.


.:.  The Richards report could be re-done to correct the mis-match
between data tables and the graph(s); the several graphs could be
reduced to a single bar graph presenting the average efficiencies
for all of the ropes when knotted (also, do we need bars out to
100% for "break strength" ?), dispensing with the like graphs
per rope, and the curves graph.  Except for the notes about rope
slippage in certain knots, of results for testing the "cowboy bowline",
and the difficulty of untying some knots, there isn't much
more to say in the concluding, "Comments" section.
[/size]
« Last Edit: July 11, 2019, 01:11:10 AM by Dan_Lehman »

agent_smith

  • Sr. Member
  • *****
  • Posts: 993
Re: The Dave RICHARDS 2004/5 Knots Test Report
« Reply #1 on: June 28, 2019, 10:41:20 PM »
Thanks for sharing this report.

Although every time I see yet another "slow-pull-it-till-it-breaks" type test I roll my eyes and yawn.
The yawning is getting deafening these days.

And here is yet another mindless "slow-pull-it-till-it-fails" type of test.
Link: https://www.youtube.com/watch?v=YeDhft54mJA&t=364s

And here is another: https://vimeo.com/40767916
And here is another: https://user.xmission.com/~tmoyer/testing/EDK.html
And here is another: http://sarrr.weebly.com/sar3-original-research/a-review-of-knot-strength-testing
And another: http://sarrr.weebly.com/knots.html
And another: http://www.paci.com.au/downloads_public/knots/06_Joining-knots_DDrohan.pdf
And this one: https://www.youtube.com/watch?v=P5qDr3qYZ9o  (Richard Delaney)

The problem is that the vast majority of 'knot testers' are programmed to think in one dimension only - and that is "slow-pull-it-till-it-fails" mindset.
They pull the knot until it yields and then pop the champagne cork and celebrate.

It is all pervasive and dominates the thought processes.

Richard Delaney has been making an effort to alter the pull-it-till-it-breaks paradigm lately (which is good).
He has been doing some slack shaking and cyclic loading tests... and he also examined the #1085 (Double F8 loop) when one the eyes was cut (although he didn't show the cutting of the eye while it was under load).

The Dave Richards report doesn't even reveal exactly how the knots were tied. And he list 'Fig 8 end' and 'Fig 8 bight'... which is amusing to me (unless its a typo and meant to write #1411 F8 bend).
At least with Richard Mumford we can see the knot specimen and the geometry.

But it begs the question... "What is the ultimate purpose?" And; "What new revelation does it bring to the table?"

Richards also lists 'Bowline'. Great - another revelation.
Presumably he means #1010 Simple Bowline?
Or does he mean #1012, #1013, #1074, or some other variant?
If he did in fact test #1010 Simple Bowline, why?
Presumably he had a concept of [a] 'Bowline' in his mind? And he defaulted to that for the test...

With 'Butterfly' - what does Richards mean exactly. What was the loading profile?
#1053 Butterfly can be biaxially loaded (through loaded) from SPart to SPart or; it can be eye loaded.
This is not stated.
Also, there are different dressings of #1053 Butterfly - we have no photo/image so no way of knowing the precise geometry.

One (and just one) example of why raw MBS yield is almost meaningless is with regard to #1410 Offset overhand bend.
Is the MBS yield point of the knot of any real importance?
In my mind, what matters most is resistance to slack shaking, cyclic loading, flogging (now becoming one of my favourites) and circumferential (ir ring) loading (for eye knots).

« Last Edit: June 29, 2019, 12:40:51 PM by agent_smith »

Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 3764
Re: The Dave RICHARDS 2004/5 Knots Test Report
« Reply #2 on: June 29, 2019, 06:10:00 PM »
Although every time I see yet another "slow-pull-it-till-it-breaks" type test I roll my eyes and yawn.
The yawning is getting deafening these days.
...
The problem is that the vast majority of 'knot testers' are programmed to think in one dimension only - and that is "slow-pull-it-till-it-fails" mindset.
They pull the knot until it yields and then pop the champagne cork and celebrate.
Worse, it is everywhere done poorly or worse,
seldom with a hint of acknowledgement of other
testings --or, if so, of how different those might
be from what is being presented.

The seemingly comprehensive --at least, extensive--
set of testing/etc. listed by SARRR.WEEBLY should be
adequate for me/us to begin building a table of the
*knots* and the many test results,
to show how sometimes stark a contrast there is!
Like it or not --the break values--, one should at
least be concerned about the variance seen in these
results, and the continued void of pertinent information
(e.g. knot geometry) to fill out this information.

I was struck by the comparatively LOW break values
seen in that recent (2015) Yachting Monthly tests of
"Marlowbraid" rope !!  --often <50% (!) (though there
wasn't independent testing of the tensile strength,
and maybe Marlow tends to exaggerate that!?).
And an eye splice for that double braid being but 80%?!
(Some standards organization I thought had stipulated
that eye-splice strength BE what is given as tensile,
I suppose reasoning that users want to know useful
not mythical values.)

Quote
And he list 'Fig 8 end' and 'Fig 8 bight'... which is amusing to me
(unless its a typo and meant to write #1411 F8 bend).
No, it really IS the case that Richards, and also the
CMC Rope Rescue Manual, 3rd ed., give values for
both tyings --"rethreaded" & TIB-- of a fig.8 eye knot !
--WITHOUT COMMENT as to why.  (And in both cases,
the results of testing 5 specimens for each *knot*
is different by just 1%point, IIRC.)
NOW, **I** can think of reasons for doing this:
1) one tying method (thinking TIB) might induce
some torsion, which can affect strength;
2) or the resulting exact geometry might differ,
of there might be different ends loaded --and, yeah,
the testers don't think to observe & comment on this(!).

An empirical study of what users do and what results
from that doing would be informative.  BUT one should
make such points known.  Otherwise, we should expect
that once tied by end-reeving or TIB and "dressed & set"
that the same "fig.8" knot will result,
with no significance on behavior.

NB : As also shown for I think later testing, for their
3rd Edition, CMC found that the fig.8 eye knot loaded
*through*/end-2-end (thus, offset) (eye UNloaded)
was stronger than the inline fig.8 so loaded!
--and the point of the latter knot IS supposedly better
through loading (though for which it is essentially a
variation on the squaREef knot, the infamously
"never use qua bend" entity.

Also eyebrow-lifting : CMC's test of a wet fig.8 EK =91% !!!
Why so high?  Perhaps --conjecture ahead-- the wet line
slipped through the SPart-turn's nip more and thus gave
complementary gripping against the SPart before that
U-turn point, and these combined/broadened areas
of friction enabled the high value?!  --plus, as SS369 might
note, maybe some reduction of heating,
though one can read of heat issues in marine nylon
for cyclical loading!?

Quote
The Dave Richards report doesn't even reveal exactly how the knots were tied.
But none do, for the most part --we never know,
and often --forgo exact geometry-- not even merely
which-end-loaded differences or sheet-bend version
or butterfly end-loading (as it's asymmetric, but I
think many don't much realize this).

Quote
Richards also lists 'Bowline'. Great - another revelation.
Presumably he means #1010 Simple Bowline?
Of course he meant #1010 --what other "bowline"
is known & talked about in the general public?
HE DOES cite "Dutch bowline" as having been
cursorily tested and found roughly equal (though
we can remark that it differs in geometry more
than those diff-tied fig.8s should!).

As for butterfly presume eye-loaded; but here, too,
it would've been a better two-tester to do also
*through*/end-2-end loading.


But, yes, in sum, a century of mediocre, un-clever,
and incomplete testing, REPEATEDLY DONE, does
NOT advance our understanding (except for some
hints of things implied by such variances as noted
here and elsewhere).

Quote
In my mind, what matters most is resistance to slack shaking, cyclic loading, flogging (now becoming one of my favourites) and circumferential (ir ring) loading (for eye knots).
Ring-loading IMO is of parochial importance
--not something to be concerned with in many
uses, but maybe in some others.  (Same as
having an eye knot be TIB : pointless for a
tie-in knot, e.g..)
The security/stability behavior stands in need
of some fairly uniform ways of assessing.  But
I think that (increasingly?) many cordage users
realize that "staying tied" is a Really Good Thing,
& nevermind how excessively vs. adequately strong
the knot might be.
(But that Yachting Monthly article sure did imply
that "best" = "strongest".  Meanwhile, in doing some
small tasks of late (seizing basketball netting bights,
and other stuff), I value much EASE OF TYING (where
sometimes "ease" seems more like "ability to" than
anything resembling "easy"!   :-[  )


--dl*
====



Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 3764
Re: The Dave RICHARDS 2004/5 Knots Test Report
« Reply #3 on: June 30, 2019, 08:12:48 PM »
Here I'll dump some further remarks re extant knots
testing & reporting.  (Oddly, only this thread came up
for a Search(Yachting Monthly) but I know that we have
other references to it?!)
--sorry, to little time to be more circumspect-attentive
on this & other posts that should be referenced/related here.
 :-\

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

One interesting test of argued semi-common PP laid
rope w/fisherman's knot (single) vs carrick bend
saw former at 70% & latter in 50s (!!) --nevermind Ashley's
sometimes parrot'd praise of latter (and wrongheaded
dismissal of former as "bulky").  CLDay's _Sailors' Knots_
has a similar difference in ORDER, but not this magnitude,
for manila rope.
  But I will note a diff. of dressing for fish.knot :: that used
  by this tester seemed to pull ends down **around** opposed
  SPart, rather than have them early on pulled by own SPart
  up into knot.  I think that this *around* dressing can
  (a) be sustained in much cordage, and
  (b) will see each SPart crunching hard, thus, into TAIL and
   so give some protection to the tight-binding behavior of
   SPart-on-SPart, and > strength?   --have been noting that
  many photos of broken knots show break where there is
  such tight constriction.  (E.g., in sheet bend, CLDay's old
  book's image shows U-/bight-part being what broke,
  as loop part SPart turned around it (which suggests that
  as one has an ever thicker U-part when tying diff.-dia.
  ropes together, the break strength in % of smaller rope
  will RISE!).

And eye splice in Marlowbraid checks in at about 80% only?
--maker's own site has "85-95%".
And a mere 47% with commonly used bowline!?
--20 %pt.s less than in kernmantles in Richards's test!!

Though now making a quick check of values, I see from
Marlow's site's spec. for 12mm = 4,090/4,450 (low/avg) and from
YM report "broke at 2.1 tonNES" which by on-line
value equator equals 4,629#,
which is MORE THAN TENSILE strength !?
so one must wonder what sort of sloppiness or misleading
values or ... is going on. [edit : or that "quick check" was too quick :: KILOGRAMS!  oops  :-[  :-[ ]
If 47% etc. is right, than roughly 2.1 x 2 = 4.2 tensile ... 18mm
given tensile strength, approximately.


AND PROBABLY NO ONE WROTE TO Y.M. ABOUT THEIR TEST
REPORT ARTICLE (I've not looked for 2015ff issues to see).

ACTUALLY, KM#143 (current/June'19)
HAS REVIEW OF THIS Y.M. TESTING OF 2015 BY A LONG-TIME
IGKTer, AND IT HAS NO FAULTING OF REPORT;
per that, neither did I catch this value contradiction 'til now.

<sigh>
BUT, that blunder of mine righted,
the article is quite flawed.  One odd miss : though
it mentions "Carrick bend" twice on introduction
--as a tested knot--, there is no mention in the
report of actual testing!?


. . . and so on . . . .

.:.  Perhaps if I et al. get better focused on drawing up
critiques of extant testing so as to capture all aspects
into a form for future use by testers so as to HAVE
those aspects covered,
such a form can become *public* knowledge and the
testing bar will get raised.

--dl*
====
« Last Edit: July 02, 2019, 12:26:51 AM by Dan_Lehman »

Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 3764
Re: The Dave RICHARDS 2004/5 Knots Test Report
« Reply #4 on: June 30, 2019, 08:21:16 PM »
Further to Yachting Monthly talk,
I see --so far, only the FACT of, not actual article--
that they'd some similar knot-testing article in 2006.
Here is as much as I've found re that,
from a Letter to Y.M. and ed.'s reply.

reference : https://www.yachtingmonthly.com/archive/readers-letters-extra-4-24598#TrElfRbEeq4hqwR5.99

Quote
Rope and knots-June 06
It does not feel right to take you to task as I have learnt so much
from articles in your magazine and take them with me when trying
to put things into practice but?..I was taught and, as a Scout Leader
for over 40 years, have been teaching that the two hitches in a round
turn & 2 half hitches should go round the same way. In your photo
you show a round turn and a clove hitch. Clove hitches are renowned
for slowly pulling through and if used to tie a dinghy to the pushpit
overnight will result in a missing dinghy in the morning. It is a bit
more difficult to see if the hitches in the Fisherman?s bend have been
similarly tied but is someone refers to a knotting book they will find
a different knot with just one hitch through the turns. The photo seems
to show three hitches and the last two look like a clove hitch.
A Double Fisherman?s bend has two hitches with the first going through
the turns and the second going round the standing part. The bitter end
can be made more secure when joinng rope to anchor or chain by either
using a stopper knot or seizing to the standing part or finishing with
a bowline. Your photo of the Fisherman?s Knot shows a Grapevine or
Double Fisherman?s Knot. A Fisherman?s Knot is formed from two
identical overhand knots pushed together but it is only suitable for
small diameter ropes or twine. The Double Fisherman?s Knot is more
like a pair of slightly complicated figure of eight knots pushe together
and as your tests indicate, is one of the strongest knots for joining
ropes or forming slings. Climbers tape the ends to stop them catching
on the rock face. If you do any more testing with Marlow in Hailsham,
can I help please? I fell in love for the first time and had my first kiss
with a lovely Hailsham lass called Avril one long hot summer long ago.
Oh! Happy days.
Richard Biggs

Toby Hodges, technical editor replies: The close up pictures of the actual knots
were ?mocked up? by our photographer afterwards in a studio to try and show
them more clearly as opposed to trying to do them while we were testing.
And I am glad you have raised the subject ? as I did exactly the same
-?when I first saw the pictures before we went to press ? namely the picture
of the round turn and two half hitches. I am used to seeing this knot, as are
you I imagine, in one way ? the way I tie it! But the picture makes it look
different unfortunately ? but it is still the same knot. Try tying it then looking
at it from the other side ? or simply rotate the free end by 90? ? and you?ll
see that you have the same knot as in the picture. A round turn and two
half hitches does effectively form a clove hitch on the standing part, which
is why it looks like this.
The Fisherman?s bend is again correct ? but Graham (our snapper) tied an
extra hitch in the end. He did one with and one without, but we used the
one with the extra hitch, so apologies about that.
And you are quite correct, the Fisherman?s Knot is indeed a double Fisherman?s
knot and should have been labelled so. While testing, we knew how to tie one,
and what to use one for, but couldnt agree on the correct name, so I think
?Fisherman?s Knot? stuck without being checked!


--dl*
====