Break testing 3 knots

[046]

totally agree ... nice job!

have been saying for years... how a knot is dressed out can make a big difference.

besides the strength issues, dressing out a knot out "pretty" makes for easy and positive ID that knot is indeed tied correctly. for a life line termination knot, that could have dire consequences.

Fantastic job!

i think the point should be maid that how a knot is dressed out can make a lot of differance; so someone else might not get the same results with different final "seating" of the mechanics. Also, this Tenex is flat on mount, so would have less deformed dimension/ no height at the bend on that axis(for less strength/efficiency loss there) perhaps affecting results. Global staemeants are hard to make, due to variances of dressing, materials, braid etc.

A Butterfly test should perhaps include a test of just 1 leg to eye; due to some of the ways it is used, that don't conform to leg to leg with no or less load to eye and pulled perpendicualr to line.

Thank-you very much.

 

[moray]

Good points, Spydie and 046. I am expecting this knot stuff to make splices seem like child's play.

Oh yeah, gorgeous triple fisherman's, 046. If you were going to use one all the time, but it was going to be down in a well or something where you would never be able to see it again, that's what I would want it to look like...

 

[moray]

I have put together a few photos from today's pulling tests. The rope in all cases is 3/8 inch Samson Stable Braid, with a nominal break strength of about 5600 lbs. Since the rope is not new, and had experienced one large load of unknown magnitude a couple of years ago, I don't trust any test results with this rope. It does help me work some bugs out of my equipment, and some of the pictures came out pretty good...

First picture shows a knotless pull test with the two bollards. The awl on the right (a sharpened $.93 screwdriver) locks the final anchoring overhand knot. With one full turn around the bollard, and half a turn around the clevis pin, only about 20% of the full tension reaches the anchoring knot. Far less still reaches the tail of the overhand knot, so the locking awl may not be necessary at all.

The next pic shows the 5:1 tackle I use to pull the cylinder back out after a test. You can do it without the pulleys, but they make it easy. The butternut log on the right supports a chain anchor at each end. The ends are 21 feet apart.

 

[moray]

This experiment shows a test on two double fisherman's loops. The small clevises have a working load of 4.5 tons. The first and second photos show the test underway. Tension is 420 lbs.

The last photo shows the result. The rope broke where it entered the knot, or just inside the knot. The scorpion tail saw a lot of heat, and as it pulled through the fibers melted and fused together. The knot was squeezing the rope so hard that once the rope was out of there, the knot clamped shut. It is very hard to see where the rope had passed through the knot.

 

[moray]

In this test an alpine butterfly is compared to a bowline. In the first photo we see the two knots under 480 lbs. tension.

The rope broke where it enters the butterfly, as shown in the next photo.

The last photo is a closeup of the break area. Knots are very irregular creatures: some parts are under far more stress than others.

 

[pdqdl]

I am a bit confused by the pictures. The alpine butterfly is tied as a midline knot, right? Did you attach the clevis to the loop formed, leaving the tail unattached, or is something else at work there?

Can we presume that the Bowline outperformed the loop portion of the alpine butterfly? Was there any damage to the bowline? I would expect to see fused rope if it was close to breaking strength. Can the knot still be untied?

 

[moray]

...The alpine butterfly is tied as a midline knot, right? Did you attach the clevis to the loop formed, leaving the tail unattached, or is something else at work there?...

Yes, the alpine butterfly is a "midline" knot, but I think that just means you can tie it in the middle of the line without access to either end. The standard bowline cannot be tied without access to an end. So I tied it in the middle, the middle being 4 inches from the end. Yes, the tail was left unattached.
I searched my break bucket for the bowline; here it is:

The arrow shows how the cover of the loaded rope accordioned after the break. I never thought to try untying it. The double fisherman's were clearly way too tight to untie--I ended up cutting them off. To my surprise, the bowline untied easily! It had 2392 lbs. tension on it when the butterfly broke, and even though it had been visibly stressed, I didn't see anything that looked like actual damage.

 

[pdqdl]

Cool!

So we can safely conclude (on a single test) that bowline beats alpine butterfly for strength and ease of untying?

Another question: you did not report the "load at break point" for the double fisherman's knot. It would be interesting to see how it compared to the alpine butterfly.

 

[blewgrass]

Funny story between the butterfly and the bowline. I used to do a bunch of boat mooring work, which in maine entails attaching chain to huge pieces of granite. To move these moorings, we would hook onto them with a boat and drag them off a shoreline until they hung from the bow of the boat. We had a set way of lowering them to the bottom in a controlled fashion which included the use of a bowline on a bight. Thinking I was going to improve on the technique and impress all who cared, I tied a butterfly instead of a bowline. After that ton and a half of maine granite hung on the butterfly it was impossible to untie. I remember my boss handing me the rope and saying "here you go, get that untied." The three strand nylon was fused and probably pretty close to it's breaking point. Learned my lesson.

I've definitely seen many knots/ropes fail in my years with marine work, but it's never been in such a controlled setting as moray. Great to be able to see a number put to a picture and have hard data from an expected rope failure.

 

[moray]

...Another question: you did not report the "load at break point" for the double fisherman's knot. It would be interesting to see how it compared to the alpine butterfly.

The DFL broke at about 3300#.

...but it's never been in such a controlled setting as moray. Great to be able to see a number put to a picture and have hard data from an expected rope failure.

I wouldn't take any of this too seriously at this point; as I mentioned earlier the stable braid is in an unknown condition. However, the general picture is starting to emerge.

The knotless pull with two bollards gave a reading of 4374 lbs., well above any of the knots. Even though we may dress and set the knot until it is perfect enough to be a museum piece, the knot immediately starts changing shape under load for the simple reason that some parts are loaded more heavily than others. The actual configuration when the knot breaks may be rather different from the unloaded configuration. This clearly varies from knot to knot; the DFL seemed little changed, but the butterfly changed a great deal. Finally, melting seems to be present in every break. Even for the non-melting fibers, we can safely assume great heat would be present at every break. Perhaps a knotted rope would be stronger if the tension increased very slowly, allowing the heat to dissipate? Or, to say it the other way, perhaps a knotted rope would be much weaker under a fast (shock) load than what a slow-load test would indicate.

 

[pdqdl]

I have often thought about the melting or fusing of rope under load, and I have come to a conclusion, based mostly on my limited understanding of thermodynamics: it requires a great deal of energy to break a rope, and that energy must be physically expended somewhere on the rope. Since breaking the rope requires deformation, it makes perfect sense that the energy released while deforming the rope might generate enough heat to cause melting. Heat from deformation predisposes adjacent fibers to fail, while simultaneously reducing the available fibers to hold the load. PRESTO! You have an accelerating localized rupture of rope fibers; very shortly, a broken rope.

My initial thoughts upon seeing fused rope in knots had me thinking that it was the massive friction caused by tightening the knot, but I think otherwise now.


************************************************

Has anyone ever seen wire rope break? It does it a bit differently, in my experience. The twisted fibers begin to unravel, and they all pop individually in a big frayed-out mess of sharp steel wires. The energy released from the individual wires seems to go into each strand, bending the broken ends away from the core of the rope, leaving a splayed out end.

They make a metal singing sort of "rrrip" sound, and then you had better duck behind something big.

 

[treemandan]

I have to say I am one for tying the rope to the truck and all the ropes I broke did not break at the knot.

 

[Philbert]

Has anyone ever seen wire rope break? It does it a bit differently, in my experience. The twisted fibers begin to unravel, and they all pop individually in a big frayed-out mess of sharp steel wires. The energy released from the individual wires seems to go into each strand, bending the broken ends away from the core of the rope, leaving a splayed out end.

They make a metal singing sort of "rrrip" sound, and then you had better duck behind something big.

When I was 16 I had the privilege of watching a tow truck operator try to pull a stuck fork lift out of a rut with brute force. As you note, the cable splayed out into a giant fan/funnel shape in a split-second. Very impressive.

Philbert

 

[treemandan]

See what I mean about the rope breaking in the middle rather than the knot? Granted these are old bull ropes and such I am using but the " pop" usually starts about there for me.

 

[moray]

See what I mean about the rope breaking in the middle rather than the knot?

Ah, the picture tells the whole story!

In various tests I have done to measure the friction effect of a rope wrapped around a wooden post or limb, as a rough rule of thumb I came up with this: for every half wrap the tension in the rope diminishes by at least 50%. A second half wrap would knock it down another 50%, so a full wrap reduces tension to 25% of the original.

Applying this rough rule to your photo, and noting that you have about 3/4 wrap around the trunk, the knot at the upper arrow would see something like 35% of the tension on the main line. Since your knot can easily handle 35% of the rope breaking strength, it isn't the weakest part of the system.

But why isn't it breaking right where the main line enters the loop? That would seem to be the weakest spot to me.

 

[pdqdl]

I have an answer for that:

Knots weaken the rope by putting a bend in the rope that has a smaller radius than the rope can handle. Then the outside of the curve carries most of the load, and it tears the rope fibers most at the outside curvature.

Since the bend of the loaded leg of the rope as it passes through the bight is not even a full 90° turn, the rope is not as significantly weakened there as it would be if it were in a full turn as part of a knot. The loop end of the rope only holds enough force to balance the line deflection of the loaded leg and whatever line pull is present after it makes the 3/4 wrap around the log.

So...less likely to break at the hitch than a casual observation reveals.

 

[TheTreeSpyder]

i think the answer(s) are more multi-dimensional. We have the inner and outer effects as stated, the outer arc will stretch further and carry most of the load.

But also, the rope only resists distance on the inline axis; and then also only in the tension direction. Because the volume of force is power X distance; only by resisting distance can we raise force from the volume of available force. Thus, any compressed fibers on the inside of a curve are not supporting at all; and other fibers then do this work/carry the inner fiber's share of the load. Now, flat line has less bent dimension on this axis, and less resistance to bend (typically); with much less / if any 'strenght'/efficiency loss on bend. But, also a bent leg of support holding 100# isn't inline all the way. So only the sine of it's angle is supporting the load. So, line tension at that point must be higher until the line tension X the sine (percentage of tension on inline axis to support load)=100# needed to be supported. Or 1/sine (cosecant) x load to be supported. With 2 angled lines supporting, that would be load /2 X cosecant to show line tension to support the given load at angle.

Then, also; the rope deforms around the mount, as tension on the outer curve pulls around as the friction on the inner part of the curve is pushed backwierds by the passive/ responding force of the spar. As tension decreases (because of the friction reducing force) so, therefore the rope is thicker at this stage. All of these distortions of form and force happen in a small area; so they 'impact' with their cumulative changes. Note, if we form a bend by a Dbl.Noose/ Scaffold Hitch. the bend, size etc. distortion is spread out over a longer distance,so is less impacting. This effect is also seen in Buntline and Lobster Buoy etc. But, is essentially lost, if a Round Turn rather than a simple Turn around the mount; for the line tension is reduced after the Turns, so the finishing choke around Standing is less forceful, and doesn't grip line as hard to buffer the forces of distortions etc. Taking away most of the 'strength' increase of Dbl.Noose, Buntline and Lobster Buoy over singleNoose, Turn + 2 Halfs, or Turn + 2 halfs opposing (respectively). Also finding, that Hitches opposing as a finish (Cow, Turn + 2 Halfs, Lobster etc. ) are typically easier to untie.

The last knot lacing should break where the knot forms, any of these things should get many permutations before we really drew hard data to forecast by.

As to the Bfly; i've never h-eared it said it was maid to be the strongest etc. To me it has a typical primary distortion point of a Half-Hitch; like a Bowline to each side; but then this primary distortion is longer like a dblBowline, suggesting more strength/by same change over longer distance, for less impacting change. But, the Bfly doesn't have the 2nd tier level to possibly help leverage inversion like a dbl.Bowline. Bfly though does have other distortions pushing to side that Bowline doesn't. And also distortions just past that , that could place 'back pressure'(?) of distortion, in the imagery of the fibers getting stiffened inside the lacing and this stiffening giving further distortion/ weakness to the primary(?) if close enough.

 

[brnchbrkr]

:jawdrop:



rep's on the way!

 

[moray]

...Since the bend of the loaded leg of the rope as it passes through the bight is not even a full 90° turn, the rope is not as significantly weakened there as it would be if it were in a full turn as part of a knot. The loop end of the rope only holds enough force to balance the line deflection of the loaded leg and whatever line pull is present after it makes the 3/4 wrap around the log.

Since I would NEVER nitpick, I agree with this as far as it goes. An engineer would look at the junction where the main rope passes through the loop and say something like this: "Since we don't have a system in motion, the vector sum of the tensions in the 3 legs at the loop must be zero, that is, they must be in balance (as you noted). Any set of 3 angles and 3 tensions that is balanced is a legitimate solution to the problem."

Take this one (very common) example. The 3 angles are each 120 degrees. That means the 3 tensions are exactly the same. Which means the main line, where it makes the 60-degree bend through the loop, is under full tension all the way. Surely then, it is weaker where it bends than where it is straight!

Now we don't expect this scenario because as load was applied to the rope and it was tightening and stretching around the trunk, the trunk friction was applying significant counterforce. The result would be much less tension at the knot and loop than in the main line. Then the angles cannot be equal (120degrees).

But what we care about is the main line before and after it passes through the loop, because that is where the big tension is and where we expect the break. Call the main line before the loop "B" and the line after the loop "A". In order for the tension in A to be different from B, the angle from the loop to A must be different from the angle from the loop to B, that is, not balanced. This also means, equivalently, that there is a frictional force where the loop holds the rope that is pushing either towards A or B. Let's say we have 200 lbs. pushing towards A. That means the tension in A is 200 lbs less than in B. If the loop were a steel ring, I would say sure, it can easily apply a sideways force to the rope. But it is a 3-strand rope, and I think it would try to roll to relieve any sideways force. Conclusion: with the setup shown in the picture, the tension in the rope would change very little where it passes through the loop, and therefore that bent passage should be the weakest spot in the entire system.

 

[moray]

...The last knot lacing should break where the knot forms, any of these things should get many permutations before we really drew hard data to forecast by...

Spydie, I have never tried to forecast knot performance beyond the extremely general idea that a really sharp bend under heavy load (overhand knot) is going to perform worse than a more complicated knot that presents less load to the first sharp bend.

The butterly illustrates the problem of forecasting. First off, there are two ways to load it, and it seems plausible that they could behave differently. Second, the one I tested changed shape significantly even under very modest loads. Obviously the knot breaks in its heavily loaded shape, not the configuration it had when you tied it. If you can't predict the final shape then any forecast of performance is going to be somewhere between weak and worthless.

Though I have not yet tested this, it seems likely that rope material in the guise of rope on rope friction is going to be a factor affecting the shape-shifting knots like the butterfly. It won't be able to change shape as much if there is a lot of friction, and so it will break in a different configuration than a slippery rope of the same construction. One might be able to test this just by squirting oil all over one knot while testing another one dry.