Two by Four knot (or what knot is this)

[Dan_Lehman]

two-loop-bends.

Pardon my laziness in not checking back in this thread
(or my lame memory ['laMEory' in fusion-writing  ])
but what is this : the eyeknots simply pulled with eyes
through each other, simply --no sort of knotting--;
or (another option) the clever structure in which each
knot receives the other's eye leg for completion (which
CLDay called "twin" in the case of bowline knots) ?
I THINK that we've only dealt with the former, in testing,
here, but have discussed the other.

Now, my observation is this :
the overhand eyeknot was stronger in its single form
than in the "two-loops" configuration!?  This would seem
to imply that the break occurred at the eyes' union, in
the latter.  (THIS degree of Where did it break? observation
is available any time post rupture!)

Now, I guess I should infer my answer from your test of
joined-bights : that they sustained 140% or so of the line's
strength, so clearly the failure must come at the "nub".

In that case, we can look at the admittedly small difference
between the joined-eyes & isolated overhands and maybe
just shrug it off as a small difference that might swing in
either direction?!

<sigh> and I should double-check the questions/answers
about exact geometric orientation (and beyond that one
can wonder if indeed things were as stated, always, esp.
in such small, fiddly material!).

Any of the specimens that you have suitable to photograph would suffice.
Closeups of the broken area, perhaps naming the type of knot please.

Of course, video is out of the question --even hard-to-obtain
super-high-speed video won't show a break very well.

Photographing BOTH of the two-loops knots should give
one complete and near breaking, and that might be the
most useful.  But this IS tiny material, after all!


--dl*
====


ps:  In these latest results, it appears that the zeppeliln
is an *outlier* and should be removed, by some thinking ... 

[X1]

The order of the tests and the knots tested were as follows:

Is  this "order" the temporal order you have performed the tests - which should not be expected to make any difference -, or the spatial order of the successive segments / portions of the rope, on the reel of line you have used ?

In these latest results, it appears that the zeppeliln is an *outlier* and should be removed, by some thinking

  True, but this could/should have been expected - in a way. I believe that the Lazy dog bend should be compared to some variation of a double Zeppelin bend, not to the single, common Zeppelin bend. Or to the Strange bend, or the Cow hitch bend, i.e. with bends where we have a double line turning around a double line. See also Re#109 (1).

1. http://igkt.net/sm/index.php?topic=4150.msg26032#msg26032

[James Petersen]

The order of the tests and the knots tested were as follows:

Is  this "order" the temporal order you have performed the tests - which should not be expected to make any difference -, or the spatial order of the successive segments / portions of the rope, on the reel of line you have used ?

Good question. The order was both spatial and temporal. It later occurred to me that it may have been better to test all iterations of each knot successively in that all tests of one kind of knot would be from one continuous section of  line. But I rotated in the order given.

In these latest results, it appears that the zeppeliln is an *outlier* and should be removed, by some thinking

  True, but this could/should have been expected - in a way. I believe that the Lazy dog bend should be compared to some variation of a double Zeppelin bend, not to the single, common Zeppelin bend. Or to the Strange bend, or the Cow hitch bend, i.e. with bends where we have a double line turning around a double line. See also Re#109 (1).

1. http://igkt.net/sm/index.php?topic=4150.msg26032#msg26032

Actually, I simply used some knots that I feel are known quantities to test alongside the EOL loop bends. I don't suggest (nor do I feel it would be prudent to suggest) that they-are better than any of the knots I tested them alongside.

I didn't include the 2x2 Lazy Dog variations in this round of tests because I don't like how hard it is to untie. The 1x2ssl version, in particular, of the Lazy Dog Loop has tested stronger than the 2x2 variations and it seems especially easy to untie, even after fairly heavy loading. I hope to do some tests of this by loading some of the loops to near failure and then see with what ease/difficulty they untie.

One aspect of the results of this round of testing was the knots which had the smallest standard deviation -- two overhand loops, the zeppelin bend, and the 1x2ssl Two Lazy Dogs. I was impressed with the  consistency of the former two and pleasantly surprised by the latter.

[X1]

I didn't include the 2x2 Lazy Dog variations in this round of tests because I don't like how hard it is to untie.

   Here you test how strong is a bend, not if it jams or not - which is an altogether different issue. The 2 x 2 Lazy dog may well be a very strong bend, that can be used as a permanent end-to-end joint. You should test it. Ropes are cheap nowadays - it does not cost too much if you have to cut off a small segment, in the rare case you will have to release a permanent bend and you find difficult to do this.
   I can not explain or understand why the 1 x 2 Lazy Dog eyeknot-to-eyeknot compound bend is stronger than the 2 x 2 version ! I would have bet for the opposite - and I would have lost, again !  However, the prudent thing is to imagine, to always have a (provisional) theory, to test this theory, and to falsify it. It is not to dare to suggest something only when you will be sure, 100%, in advance, that you will be proven to be right !  The strength of complex structures made of non-homogenous materials is a very difficult thing to predict. Nevertheless, we should always predict, and then test our predictions, just like you did. We can not understand anything if we do not have a theory - it is the theory that determines what is observable.
   I had not predicted those numbers - they would have seen way too hgh to me ! The only explanaton I can think of is the "flattening effect", where the lengths of the paths of the individual yarns and strands are forced to be equalized, to a high degree. So, perhaps the future tests on thicker ropes would show lower strengths.
   For the work of a colleague of yours, see :
   http://hilarynelson.com/Hobbies/Knots   
   (The data from those tests is available as a .zip file)

[Dan_Lehman]

The order of the tests and the knots tested were as follows:

Is  this "order" the temporal order you have performed the tests -
--which should not be expected to make any difference--,
or the spatial order of the successive segments / portions of the rope,
on the reel of line you have used ?

Good question. The order was both spatial and temporal.
It later occurred to me that it may have been better to test all iterations
of each knot successively in that all tests of one kind of knot would be from
one continuous section of  line. But I rotated in the order given.

Here, "spatial" is meant that line flowed into whichever
indicated tested item as indicated --and not that all of
the items were in a single test specimen!  (One could
have four, though : anchoring eyeknots at the ends,
and expected-to-be-broken (one of...) end-2-end knots
between these (which would yield a survivor loaded to
near rupture, for examination).)

I disagree that order "should not be expected to make
any difference" : rather, while we would hope that it
did not, we would do this staggered/mixed testing on
the chance that exercising of the device DID have some
effect --not expected, perhaps, but not inconceivable.
(And here is where I have issue with statisticians,
who would disclaim a careful ordering to ensure this
diversity, and demand instead randomness --which
could, as statistics books like to point out(!), result
in quite non-random-looking, non-MIXED order!
The device doesn't care how the order came about,
but if it does tend towards some bias upon exercise,
will show that on whatever specimens it encounters
at those times!)

I can not explain or understand why the 1 x 2 Lazy Dog eyeknot-to-eyeknot compound bend
is stronger than the 2 x 2 version ! I would have bet for the opposite

Perhaps if you saw what they looked like at high load,
you would make a different guess --that you might
retain your theory of strength, and see that in the
state of high load, it pointed rightly to something
one couldn't easily recognize in its set geometry.
Or maybe other and more substantial materials
would show a different result.


--dl*
====

[X1]

One could have four, though

   I do not trust experiments where two or more knots we want to test independently the one from the other, are tied on the same tensioned segment. Vibrations and sudden dislocations can well be "transmitted" through the line, and influence things in ways we can not foresee ( they can play the role of a triggering effect, for example, that will affect a nearby knot more than the knot where they had appeared as first signs of an impeding rupture ). Also, the tensile forces delivered through the anchor points can be absorbed from one knotted structure, at a certain instant, and from another, at another instance, futher complicating the loading pattern. To my mind, independence means separation - each knot should be tested in isolation of any other knotted structure tied on the same tensioned line, that could possibly influence its behaviour.

I disagree that order "should not be expected to make any difference" : rather, while we would hope that it did not, we would do this staggered/mixed testing on the chance that exercising of the device DID have some effect --not expected, perhaps, but not inconceivable.

Right ! ! ! We should not suppose that the apparatus itself is left unchanged after each test. 

you might retain your theory of strength

   Frankly, I start to think that "my" theory is too simple... I have a conseptual "image" of some static structure, made from articulated elastic rods, something like a metal building or bridge, which is confronting tension, compression and torsion. Perhaps that is only a very limited view of the story. There might be other things that determine the outcomes of those experiments...I can imagine the material being melt at some points, but not at some others, due to some obscure reasons. Perhaps more fibres, the one next to the other, make the flow/transmission and dissipation of the raising, by friction, temperature, more easy within a thin line, and more difficult within a thicker one. I admit that I could not have predicted those high values - so I guess I have to wait patiently for the next series of tests, on thicker ropes. 
   J.P., go on ! 

[Dan_Lehman]

One could have four [knots in a test specimen], though
[ anchor-knot ....... tested_end-2-end#1......tested_end-2-end#2......anchor_knot]

   I do not trust experiments ...
 Vibrations and sudden dislocations can well be "transmitted"
through the line, and influence things in ways we can not foresee
...
...
To my mind, independence means separation - each knot
should be tested in isolation of any other knotted structure tied
on the same tensioned line, that could possibly influence its behaviour.

But such hyptothesized problems remain just that,
and at cost of getting some real information.  Knot
use, after all, hardly comes in such pristine conditions
as some folks seem to want for testing --which conditions
beg the question of relevance, if so much must be taken
away lest it tarnish results!  What good is my knot if it can
only survive to this test_result in *germ-free* conditions?
I cannot use it out of doors!


--dl*
====

[X1]

if so much must be taken away

  " It turns out to be very difficult to devise a theory to describe a knotted structure all in one go. Instead, we should break the problem up into bits and invent a number of partial answers. Each of those partial answers would describe and predict a certain limited class of observations, neglecting the effects of other quamtities, or representing them by simple sets of mumbers. It may be that this approach is completely wrong. If everything in this knotted structure depends on everything else in a fundamental way, it might be impossible to get close to a full solution by investigating parts of the problem in isolation .Nevertheless, it is certainly the way that we have made progress in the past. "
( with just a little bit of editing, by an arm-chair knot tyer)