Author Topic: Jam testing several bends  (Read 5673 times)

dmacdd

  • Exp. Member
  • ****
  • Posts: 120
    • My Knot Pages
Jam testing several bends
« on: May 17, 2010, 04:34:35 AM »
I have done an experiment that roo suggested to me in another forum. I did the experiment and have published the results on the web

I tested the tendency of several bends to jam as described at

http://davidmdelaney.com/jam-testing/jam-testing-several-bends.html

I was startled by how bimodal the results were.  The Carrick, Alpine  butterfly, and Zeppelin bends never jammed. The Hunter's and Ashley's bends always jammed. It seems that using a thin stretchy cord really amplifies any tendency to jam. Not surprising, really, when you think of the knot as locking up energy elastically, compressing itself after the external load has been removed.

I am not aware of any other comparative data of this sort being published on the web. Is there any?  Will these results be useful to anyone other than myself? Should they be used by anyone else?


Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 4278
Re: Jam testing several bends
« Reply #1 on: May 17, 2010, 06:46:17 AM »
I did the experiment and have published the results on the web ...
http://davidmdelaney.com/jam-testing/jam-testing-several-bends.html

Thanks for this.
 
Quote
The Carrick, Alpine  butterfly, and Zeppelin bends never jammed.
The Hunter's and Ashley's bends always jammed.

Noting that "Ashley's" = "#1452" (as opposed to other similar bends
--and some not so...-- that he also introduced, notably #1408, 1425),
I'm surprised at this jamming.  I suspect that you didn't tie it appropriately
to be jam-resistant.  Looking at your photos, I see that you've managed
at least two versions, judging by the ends' dispositions:  that image in
your test report shows ends pointing to the side of their color; that in
the tying instructions for the knot have them pointing opposite (which
would more likely be the non-jamming versions).  (Btw, the "Bosun's
method" seems a highly multi-stepped failure at tying expediency!
I always tie it by first making an Overhand in one end, and then
reeving in the opposite end according to what knot (version) I
want -- be it 1408, 1452, Shakehands, 1425a (SmitHunter's), or
1425a-improved.  I find that like method for Rosendahl's Zeppelin
Bend to be equally too-clever-by-half.  (And someone came up
with a 1425a adaptation, too.  bah, humbug!) )

#1452 (&1408) when loaded will twist the nipped ends in
a direction, as --unlike for the Butterfly-- the SParts rotate
the same way; if the ends are oriented on one way, this
twisting will only bind them tighter; but if they're neutral
(as your for-testing image shows), or oriented the other
way, the powerful twist will move them around to where
they can walk out a bit and the collars of the knot will
then expand and wrap around the knot body, locking
-- which might be just what one wants, or maybe not.

You wrote "I now think that sailors of the old days of sail wouldn't
have had much use for Ashley's bend, because it jams so easily"
,
but this just isn't so:  esp. in the rather static, non-shrinking (under
load) natural-fibre ropes of yore there'd be no jamming (or none
that some whacks w/appropriate tools wouldn't loosen).  This
bend is rather similar to the Carrick (#1408 can be seen as an
easy derivative-of/improvement-to the Carrick).

Quote
It seems that using a thin stretchy cord really amplifies any tendency to jam. Not surprising, really, when you think of the knot as locking up energy elastically, compressing itself after the external load has been removed.

Yes, and the small size made it tedious for you to try to get some
purchase on it to untie it.  OTOH, rope can be tough because of
the tremendous force it can see!  I just tried #1452 in 3/8" laid
(medium) nylon, loading it with perhaps 400#; it didn't jam.
Proportionate loading to your test, however, would've taken
maybe a half ton (I could try bouncing on the pulley ...).

(FYI, your URLink to the Butterfly bend doesn't work.)
(And thanks for posting (you?) to YouTube to correct those
comments calling Tiatt's knot a Zeppelin --surprising one did
so even with the correct citation into ABOK for #1452!
Unfortunately, your reply comment only showed (now, to me)
after clicking View All Comments, and not immediately on the
first page w/ the video.)

--dl*
====

dmacdd

  • Exp. Member
  • ****
  • Posts: 120
    • My Knot Pages
Re: Jam testing several bends
« Reply #2 on: May 17, 2010, 02:20:00 PM »
Dan

Thank you for your detailed and thoughtful response!

Quote
I'm surprised at this jamming.  I suspect that you didn't tie it appropriately
to be jam-resistant.  Looking at your photos, I see that you've managed
at least two versions, judging by the ends' dispositions:  that image in
your test report shows ends pointing to the side of their color; that in
the tying instructions for the knot have them pointing opposite (which
would more likely be the non-jamming versions).

I have just spent some time verifying that the two tying instructions I gave for 1452 produce a knot topologically identical, and I believe identical in all respects, to the method indicated in Ashley's drawing. Taking the "lubber's" and "bosun's" methods of my web site, and Ashley's drawing in #1452 as three distinct methods, I tied the bend this morning half a dozen times by each method, playing with the dressing of the knot.  I found that for all three methods the bend winds up with each working end pointing toward its own standing part most of the time (5:1?) when no attempt is made to control which way they point.  Occasionally each working end winds up accidentally pointing to the other rope's standing part.  They can always be adjusted to point either way by adjusting the way they nestle in the pocket in which they are nipped.

I was not attending to the way the standing parts were pointing when I made the photos, and by bad luck, both methods wound up showing the low probability final orientation. I believe I should revise the photos to show the high probability final orientation of the working ends.

It seems likely that I tied it with both resulting orientations in the tests. If that is so, then it jammed both ways.  Even if it were jamming only one way, the knot should not be absolved of the charge of jamming if such subtle and apparently random differences in the outcomes produced by its tying methods (all of the tying methods under discussion, including Ashley's) produce jamming configurations.  Besides, I just ran two tests, each with a different orientation of the working ends: they jammed equally.

I believe I should change the photos to show the higher probability outcome: each working ends pointing to its own standing part.

Quote
(Btw, the "Bosun's
method" seems a highly multi-stepped failure at tying expediency!
I always tie it by first making an Overhand in one end, and then
reeving in the opposite end according to what knot (version) I
want -- be it 1408, 1452, Shakehands, 1425a (SmitHunter's), or
1425a-improved.

Surely there is some value to a tying method that eliminates
the flexibility of outcome you admire!

Quote
You wrote "I now think that sailors of the old days of sail wouldn't
have had much use for Ashley's bend, because it jams so easily",
but this just isn't so:  esp. in the rather static, non-shrinking (under
load) natural-fibre ropes of yore there'd be no jamming (or none
that some whacks w/appropriate tools wouldn't loosen). 

This is entirely plausible. I'll revise my comments along these lines.

Quote
Yes, and the small size made it tedious for you to try to get some
purchase on it to untie it.  OTOH, rope can be tough because of
the tremendous force it can see!  I just tried #1452 in 3/8" laid
(medium) nylon, loading it with perhaps 400#; it didn't jam.
Proportionate loading to your test, however, would've taken
maybe a half ton (I could try bouncing on the pulley ...).

Fair enough. But surely the fact has some significance that Carrick, Alpine, and Zeppelin
are extremely easy to untie with bare fingers under these loading
conditions while Ashley's is impossible....  I may have made too much
of the significance of the difference for ropes that are not as stretchy, and I'll fix that,
but some attribution of jamminess should stick to Ashley's because of that difference.

Quote
(FYI, your URLink to the Butterfly bend doesn't work.)

Thank you. Fixed.

David








roo

  • Sr. Member
  • *****
  • Posts: 1926
    • The Notable Knot Index
Re: Jam testing several bends
« Reply #3 on: May 17, 2010, 04:26:47 PM »
I have done an experiment that roo suggested to me in another forum. I did the experiment and have published the results on the web

I tested the tendency of several bends to jam as described at

http://davidmdelaney.com/jam-testing/jam-testing-several-bends.html

I was startled by how bimodal the results were.  The Carrick, Alpine  butterfly, and Zeppelin bends never jammed. The Hunter's and Ashley's bends always jammed.


I'm not surprised at these results.  They match my experience.

Boy, that is some small line in that picture!
« Last Edit: May 17, 2010, 04:32:55 PM by roo »
If you wish to add a troll to your ignore list, click "Profile" then "Buddies/Ignore List".

Notable Knot Index

roo

  • Sr. Member
  • *****
  • Posts: 1926
    • The Notable Knot Index
Re: Jam testing several bends
« Reply #4 on: May 17, 2010, 04:47:13 PM »
I just tried #1452 in 3/8" laid
(medium) nylon, loading it with perhaps 400#; it didn't jam.

400 lbs may very well be a mere 11% of the breaking strength.  I'm not impressed.
If you wish to add a troll to your ignore list, click "Profile" then "Buddies/Ignore List".

Notable Knot Index

dmacdd

  • Exp. Member
  • ****
  • Posts: 120
    • My Knot Pages
Re: Jam testing several bends
« Reply #5 on: May 17, 2010, 05:16:19 PM »
Quote
Boy, that is some small line in that picture!

Yeah. I got 1200 feet of it a couple of years ago at the local ship chandler. It's very nice stuff to have around the house. It's 1/16 inch / 1.6 mm.    8(?) lay braid. (They're hard to count.)

I can't break it knotted in that test setup. although I can stretch it.

roo

  • Sr. Member
  • *****
  • Posts: 1926
    • The Notable Knot Index
Re: Jam testing several bends
« Reply #6 on: May 17, 2010, 07:47:05 PM »
I have done an experiment that roo suggested to me in another forum. I did the experiment and have published the results on the web

I tested the tendency of several bends to jam as described at

http://davidmdelaney.com/jam-testing/jam-testing-several-bends.html


As I read through your description of the bends, you seem not to bring up security much.  I find that the Zeppelin Bend is more secure than the Butterfly Bend, while the Double Carrick Bend finishes a distant third (among the three), with mediocre security.
« Last Edit: May 17, 2010, 07:48:13 PM by roo »
If you wish to add a troll to your ignore list, click "Profile" then "Buddies/Ignore List".

Notable Knot Index

dmacdd

  • Exp. Member
  • ****
  • Posts: 120
    • My Knot Pages
Re: Jam testing several bends
« Reply #7 on: May 17, 2010, 09:25:23 PM »
Quote
I find that the Zeppelin Bend is more secure than the Butterfly Bend, while the Double Carrick Bend finishes a distant third (among the three), with mediocre security.

How would we test that objectively but simply?  Redo Ashley's jerk testing? (#68) When he did his jerk testing, he placed the Carrick below #1452, the Barrel Knot # 1413, and the ring knot #14 12, but above above the Water Knot #1414, which looks like a pretty secure knot, and the double sheet bend, both of which finished above the overhand bend #1557, which the climbers use. He didn't know about either the Alpine butterfly bend or the Zeppelin, of course.  (Ashley's results described in #1543)



BTW, I'm gonna get a better handle for my jam testing test rig and go up to 1/8 / 3 mm nylon braid.

roo

  • Sr. Member
  • *****
  • Posts: 1926
    • The Notable Knot Index
Re: Jam testing several bends
« Reply #8 on: May 17, 2010, 09:42:22 PM »
Quote
I find that the Zeppelin Bend is more secure than the Butterfly Bend, while the Double Carrick Bend finishes a distant third (among the three), with mediocre security.

How would we test that objectively but simply?  

After setting the bend, performing some slack, repetitive motion is the usual route.  Different rope types are good, but slick and stiffer rope like polypropylene tends to amplify security issues.

Strict repeatability is hard without a mechanism.  By hand, it's easy to get tired or introduce different motions (compression, twists, waves, impacts) with differing degrees of amplitude.  

On one hand, you could try to avoid favoritism with machines or with blindfolds.  Or you could pit the untyer against the knot in a anything-goes timed competition with the only rule being that the person can't touch any part of the rope closer than, for example, within 20 diameters of a properly-set knot.  Setting the knot itself might demand a pre-defined tension on the free ends and standing parts in order to be objective.

It depends on how in-depth you want to go.
« Last Edit: May 17, 2010, 09:44:04 PM by roo »
If you wish to add a troll to your ignore list, click "Profile" then "Buddies/Ignore List".

Notable Knot Index

Dan_Lehman

  • Sr. Member
  • *****
  • Posts: 4278
Re: Jam testing several bends
« Reply #9 on: May 17, 2010, 10:17:41 PM »
I have just spent some time verifying that the two tying instructions I gave for 1452 produce a knot topologically identical, and I believe identical in all respects, to the method indicated in Ashley's drawing.

I'm not sure you're saying what you intend:  topological equality
is irrelevant to practical effects -- the Anchor Bend has that equality
to the Strangle hitch, but those are markedly different geometries
(it is a question lingering about what a Uni-knot is supposed to be
-- transformed into strangleness or left in a sort of reverse anchorB'dness?!).

Quote
I found that for all three methods the bend winds up with each working end pointing toward its own standing part most of the time  ... . They can always be adjusted to point either way by adjusting the way they nestle in the pocket in which they are nipped.

That points to unwanted ambiguity in the tying method, to leave
this important, behavior-(e/a)ffecting geometry unspecified.

Quote
(all of the tying methods under discussion, including Ashley's) produce jamming configurations.  Besides, I just ran two tests, each with a different orientation of the working ends: they jammed equally.

Ashley's image etc. are ambiguous.
See here for a definite image of the knot:  http://igkt.net/sm/index.php?topic=1446.0
Now, in this case, I left the collars rather large, expressly so that
they would come up around the knot body; the ends could've pulled
them shorter and then they'd effect more of a choker-hitch geometry
at knot entry, and should not jam.  So, it is not only the orientation
but the amount of material in the collar (and perhaps to that point
the elasticity of the material).
I also show a version of SmitHunter's Bend that resists jamming pretty well.
One can achieve this version directly from the original, by just
loosening one collar and then tucking the opposite (adjacent) end
out through it, making a sort of twist of the ends' passage through
the knot's nip -- dress to even this twist and have some bit of each
end within each collar circle.

Quote
Surely there is some value to a tying method that eliminates
the flexibility of outcome you admire!

Rather, that "flexibility" is something to be specified per need
--I enumerated several of the interlocked-Overhands bends
that can result-- ; it's not something left to chance.  (E.g., were
one tying together some of the common slippery-springy PP rope
for some makeshift dog leash or whatever, one might wish to use
a secure-when-slack knot such as the Jammed#1452, well aware
that untying would be easy in said material & loading; so, then
the ends would be positioned ... and the not set accordingly.
But for closing a rope sling used in impromptu towing of some
vehicle out of being stuck, another version would be deliberately
employed.  The tying guidance needs to note the differences;
the image(s) should show the results clearly (unlike nearly ALL
anglers' knots images!).)

Quote
I now think that sailors of the old days of sail wouldn't
have had much use for Ashley's bend, because it jams so easily

Given some of the egregious misrepresentations of knotting history,
one can wonder at the goings-on in general.  As Inkanyezi once said
here, in choosing a basic set of knots, bending ropes together isn't
such a regular need.  In modern day, the popular Discovery Channel
t.v. series The Deadliest Catch (about Alaskan crab fishermen) showed
the use of the Carrick bend to bend on a pot-line extension when
needed for deeper water; in some cases, I think that they might put
on a seizing of electrical tape, but not always.  (I've been unable to
connect so far w/some fishers up there, but did e-chat w/one fellow
who'd gotten wet in that capacity (or had been in that neck o' woods).)


Quote
Quote
I just tried #1452 in 3/8" laid (medium) nylon, loading it with'perhaps 400#; it didn't jam.
Proportionate loading to your test, however, would've taken maybe a half ton (I could try bouncing on the pulley ...).

[Roo] I'm unimpressed.

Roo doesn't impress easily.
I guess he likes exceeding safe working loads?

In any case, I cannot say for sure what force I applied,
but when taking things even to 300# in material running
smack next to your vitals it is impressive enough for me.
That said, to satisfy the skeptics, I did much the same (but
with greater caution, LESS bouncing!) with the small size
of hollow-braid (16strand) nylon binding cord which is in
common use by commercial fishers; it probably has around
an 800# tensile strength?  This material takes on a flat
profile (like thick tape).   #1452 in one version was able
to be pried open easily enough.

And I just tried lifting 62.5# (barbell weights) with thin,
nylon (polyester?) venetian-blind cord; lift-off was prevented
by a SNAP!!!   :o  Maybe it got to 50# force, but ... --100% plus.
The break was at #1452, not the Dragon eyeknot or the double
Pile Hitch around a pen; the knot remained intact, with the
break point right at the entry/collar area yet holding well
enough for me to see that the opposite collar could be easily
enough pried back.  (Maybe the real problem is that you're
inept at untying?   ;D  )
Round#3:  Just tried this with a jamming version and used
just 50#, but had the same result (and it both cases the
break was on the lower SPart) (!); the opposite collar
however was still able to be pried off, pretty easily (I don't
think this would be the case for serious material --rope!).

As for more exact size/info re this fine cord, hmmm:  it is
just less than 1/16th, >1/32, and a guesstimate of 1.8mm
is, well, in the ballpark such as one can be sizing tiny cord
with a (crudely?) marked rule.  It has a single core strand,
and (I think --but it's hard to see) over-2/under-2 braid.

But you might want to try using some barbell weights
to get a better feel for force:  if this stuff breaks at >50#,
you might be also down in this force region, though I'm
not going to equate strengths as clearly this stuff needn't
be all so strong for its task.
(I like to *harvest* such cords from discarded items.)
((Btw, seems to me that MY cord didn't have all so much
stretch, and I'll guess is weaker than yours, too (but one
can be surprised by what barbell weights show!).))

There's another version of #1452 which you should be able
to see from a critique of the Carrick, where you think "What
if the ends were run out up through the central nipping circle
of the SParts?" -- likely what led Ashley (and me) to it.  Here,
jamming doesn't have a chance (but I'll guess that strength
suffers a little, and it's a bit bulkier).  But THIS version really
does involve some subtleties of construction/dressing!

Quote
But surely the fact has some significance that Carrick, Alpine, and Zeppelin
are extremely easy to untie with bare fingers under these loading
conditions while Ashley's is impossible....  I may have made too much
of the significance of the difference for ropes that are not as stretchy, and I'll fix that,
but some attribution of jamminess should stick to Ashley's because of that difference.

The "Alpine":  somewhere along the line "alpine" got worked into
someone's notion of name for this.  There is some thread on origins
on this site (Search; and include "Thrun" in keywords).  In short,
C.L.Day pointed to "linesmen's loop" having publication antecedent
to Wright & Magowan's Alpine Journal presentation of the eyeknot
in 1928.  A caver-auther, I think, might've decided to qualify W&M's
"Butterfly" with "Alpine" in contradistinction to some other such knot
that can be tied similarly (Slip-Knot & Half-hitch'd end).  And it does
nicely fit the medium for W&M, but ... ; I prefer to leave it off, and
not gift this other knot with wings.  (I don't mind the other knot as
much as some.)

Rosendahl's bend is probably the least jammy; I think it would serve
those crabbers quite well in their stiff, hard-laid lines!

For security testing, one could tie a series of ropes/cords to
some stick and thereby deliver like force/jitters to the set of knots.
Then one could try dragging relatively unweighted knotted material
across some kind of rough surface (carpet, lawn), leaving ends not
so long (to invite their being pushed into knot).  I think that this
testing is really going to be largely pass/fail.  (Another test would
be to see how the knots behaved in HMPE rope, but I suspect that
they'll hold (but recalling that a Dbl.Bowline did NOT!).)

--dl*
====
« Last Edit: May 18, 2010, 06:42:21 PM by Dan_Lehman »

dmacdd

  • Exp. Member
  • ****
  • Posts: 120
    • My Knot Pages
Re: Jam testing several bends
« Reply #10 on: May 18, 2010, 11:48:51 PM »
I have substantially re-written the text of Jam Testing Several Bends, http://davidmdelaney.com/jam-testing/jam-testing-several-bends.html.

I have also reworded the text of Ashley's Bend,
http://davidmdelaney.com/ashley-bend/ashley-bend.html

The changes were substantially motivated by Dan Lehman's feedback here,
although he may not entirely agree with the result.

Thank you Dan.


 

anything