Pull Test--Ice Tail Splice

[moray]

I have been using Samson Ice Tail for my hitch cord for the last couple of years. Among its excellent features are high strength, heat resistance, ease of splicing, pliability, and good surface friction. It's also relatively cheap. I like a lot of room between my belt and my friction hitch, so my split tail measures about 50 inches between the eyes. This has proved to be a little too long for some of my friends I occasionally take climbing, so I decided to make another shorter one.

The first thing I discovered when I dusted off Samson's splicing instructions for Ice Tail is that my old one is substandard in terms of bury length. Not to worry, both splices were backed up with locking Brummels and plenty of stitching. Now I like to follow the manufacturers's instructions when I can, but I absolutely required that the friction knot be entirely outside the splice zone, meaning each splice's bury had to terminate well short of the friction knot. I was quite sure much of the excellent performance of this cord came from the fact it was (1) extremely pliable and (2) only 5/16 in in diameter.

If I was going to do another substandard bury, a test was called for.

My intention was to put a halfway decent eye with stitching at one end of a 7-ft. piece of Ice Tail, and put a grossly substandard eye at the other end with no stitching. I would measure the force needed to pull the bad eye apart, then replace it with a fresh eye with a bury 2 inches longer. Pull that one to failure, measure the force, then repeat again with a still longer bury. I should end up with a nice little chart of numbers that should provide useful guidance for making the real split tail for field use.

I was in for a surprise.

 

[moray]

Thanks to an excellent suggestion from pdqdl about a year ago, I have put together a hydraulic rig that can apply forces up to 10,000 lbs. and measure them with 2-lb. resolution over the whole range from 0 to 10,000. Accuracy over the full scale is on the order of +- 10 lbs.--far, far better than I need.

For the first test splice I decided to use a bury of 4 inches, 2 of which was tapered. This is only 20% of Samson's spec, so it should be easy to pull apart.

The first picture shows the load cell with light tension on the test rope.

The second photo shows the actual splice. The distance between the arrows is 4 inches, and the righthand arrow points to the throat of the eye where the bury begins (unfortunately the camera is in the plane of the eye so the eye isn't visible).

The last photo shows the anchor for the eye under test. The screw link is there to provide a strong subsitute for a carbiner.

 

[moray]

I actually did two pulls on this first setup. On the first pull I reached 2052 lbs. and then released the tension. As this was the first pull with my new rig, and I had expected an easy pull, I had not provided any safety features for recoil. Over the next hour or so I remedied the safety situation and resumed the test. I pumped steadily on the hand pump, adding about 1000 lbs. every 5 seconds. After the gauge reached about 9800 lbs., the next pump stroke took the read-out out of limits, and the stroke after that broke the rope. Quite a maiden voyage for my new rig!

The results are shown in the photo.

The arrow points to the substandard splice, which is still intact. The circle indicates the two legs of the test eye, which extend beyond the photo. As you can see, a couple of strands in the main rope actually survived the torture.

When I make that new tail, I'll probably bury 10 or 12 inches and stitch it well. I will have 100% confidence that it will do its job and still provide a huge safety margin. (Do I need to put in a disclaimer--I am not recommending anything to anyone else?)

I am left wondering, though, why Samson doesn't provide less restrictive instructions for the Ice Tail splice. They know what the rope is for (they do call it Ice Tail), and they must know that stuffing the whole thing with buried core makes it fat and stiff and much less useful. They also must know you don't need 20 inches of bury for the splice to hold. I am guessing Ice Tail is a very minor money maker for them, and they have little inclination or motive to improve the one-size-fits-all hyper-conservative splicing instructions they have published.

 

[Blakesmaster]

THIS is interesting stuff. How did you build your testing rig? Have you tested other types of rope/splices. How are you certain the amount of pressure being appied on each test? Man, this is beautiful ####, I want to know more.

 

[masiman]

.....I am left wondering, though, why Samson doesn't provide less restrictive instructions for the Ice Tail splice. They know what the rope is for (they do call it Ice Tail), and they must know that stuffing the whole thing with buried core makes it fat and stiff and much less useful. They also must know you don't need 20 inches of bury for the splice to hold. I am guessing Ice Tail is a very minor money maker for them, and they have little inclination or motive to improve the one-size-fits-all hyper-conservative splicing instructions they have published.

I wonder how the splice performs over time, i.e. cycle testing, element exposure, gradual abrasions, etc.

 

[canopyboy]

Great test, I was thinking to test my bee line splice with locking brummel and 6" bury the same way. What kind of load cell did you use? We calibrated dynamometry all the time at my old job, and I had come up with a similar set up for high load cals that I wanted to use for testing splices.

Sure makes you feel better about what you're climbing on, doesn't it?

 

[moray]

It's good to see this has generated a bit of interest!

...How did you build your testing rig? Have you tested other types of rope/splices. How are you certain the amount of pressure being appied on each test?...

This was my first pull. I have a long list of experiments I want to do now that the basic rig has proved out.

The load cell is connected to an electronic display that converts the load-cell voltage to a reading in pounds. Load cells are very linear devices, so if it is accurate over some part of its rated range, it is probably accurate over the whole range. I have tested mine with known weights from 2 lbs up to 158 lbs, and it has been dead on every time.
My rig is still very much a work in progress, but I can tell you one of the toughest challenges is finding a suitable frame to anchor to. I am using part of a butternut tree I took down a couple of weeks ago. It is not ideal, but it is free. Loops of chain at each end of the frame provide adjustable, non-stretchy anchors. The tree itself is too big to flex much even under a load of 10,000 lbs. The big problem facing me is how to prevent a recoiling rope from whacking something I care about, like the load cell. The operator (me) is well off to the side at the end of a 7-foot hydraulic hose.

I wonder how the splice performs over time, i.e. cycle testing, element exposure, gradual abrasions, etc.

What I have read, and this is consistent with my own informal observations, is that splices become very stable after a few load cycles, as if the core and cover become more or less physically locked together. If the throat is properly stitched, the splice cannot creep apart.

...How does Ice Tail hold a scaffold hitch?...

I tied one just now to see... Ice has a very soft hand--it is noticeably softer and more pliable than the polyester in Tenex. It is also not as slippery as Tenex. I would think all those features together would make it shine at holding knots, but I never knot it myself.

...What kind of load cell did you use?...Sure makes you feel better about what you're climbing on, doesn't it?

The load cell was made in (where else?) China. I love your line about feeling better! A lot of people seem to have an incurable distrust of splices, but here is the cure, and it doesn't require a load cell: make a simple eye in some Samson Tenex, then try to pull it apart with your truck. Nothing like a little hands-on experience to change your thinking and dissipate your fears.

 

[pdqdl]

If you want to control the risk of flying components, load the rope or other dynamic elements with non-dynamic weight. Regardless of the force involved with breaking the rope, at separation you may have a lot of velocity, but not much momentum.

Throw some fabric over the stretched parts, like an old horse blanket, gunny sacks, or some other material of low value and sufficient soft weight to do the job. If necessary, you could add weight by wetting the fabric. [great idea: test a wet splice, and see how it performs]

The momentum of parts flying off to either side of the pull is rapidly absorbed by trying to move the extra weight of the fabric. Quick and easy fix.

By the way, I love reading your test reports. Keep up the good work.

I have a suggestion for getting more use out of your load cell: Rig your hydraulic cylinder to a pulley system attached to the load cell, then load the splice with the full force. I would suggest using two doubled pulleys for a 5:1 power gain. Using mechanical advantage to reduce the pull on the load cell and keep the force higher against the splice should give you more power to break splices and less risk of breaking the stuff you want to protect. By attaching the load cell to a line at the end of a pulley, you might also remove the load cell from the danger zone of flying parts.

 

[pdqdl]

Needs more information!

I can't tell from the pictures where the rope broke.

At a splice, near a splice, in the middle? Some point of deflection or bending?

 

[canopyboy]

I have a suggestion for getting more use out of your load cell: Rig your hydraulic cylinder to a pulley system attached to the load cell, then load the splice with the full force. I would suggest using two doubled pulleys for a 5:1 power gain. Using mechanical advantage to reduce the pull on the load cell and keep the force higher against the splice should give you more power to break splices and less risk of breaking the stuff you want to protect. By attaching the load cell to a line at the end of a pulley, you might also remove the load cell from the danger zone of flying parts.

Unless you're maxing out your load system, you don't want to do this if you can help it. Doing precision force and moment testing for the Navy has given me some insight into this type of thing. Pulleys aren't perfect. Highly loaded pulleys are even less perfect. Your readings will start to suffer in accuracy with every pulley between your load cell and what you're measuring. In some of our systems we found up to 10% deviation (actually about 10% for the first pulley, then 7% for the second, etc....) for each pulley we put into the system, and they were high quality ball bearing pulleys. Not that extreme accuracy is needed for this, but if you don't need the mechanical advantage to protect the load cell from being over-stressed, you might as well leave it attached directly in line with the splice.

 

[046]

moray... nice job!!

have always put a disproportional amount of weight to what few pull tests that's available.

this raises my confidence level for splices. traditionally have never trusted splices. much rather depend on a knot I personally tied vs a splice done by no telling who.

 

[pdqdl]

Well of course there will be frictional losses and changes from ideal using any mechanical advantage.

But if your system is designed to deliver 10k, and you max it out on the preliminary test, you need to invent another way to measure.

Since the Icetail rope is not rated for anywhere near that great a load, I suspect that Moray's equipment needs to be calibrated a bit better. What really matters, though, is the relative results he will get from splice buries of different lengths.

If you don't like pulleys, Moray, you could also rig a lever mechanism that would do the same job. Unfortunately, there are not many levers you can hit with 10,000 lbs of hydraulic force that will still do the job. By the time you rig that up, you will spend a fortune on metal.

 

[moray]

Good discussion of the very issues I have been wrestling with.

The rope broke right at the end of the short bury.

Mechanical advantage: I'm with canopyboy on this one. Besides the accuracy issue, you have the problem of acquiring a 10-ton or higher SWL block, extra heavy duty anchors, an extra heavy duty frame, etc. But worst of all, as I am now beginning to see, you now have big chunks of metal (the block) that want to fly. The sweet spot, for me, was 10K lb. This allows me to test any climbing rope to failure. Of course I am not interested in breaking ropes, per se, but in teasing out some of the principles governing the behavior of knots and splices.

I did a couple more tests today which I will report in a new thread, but I can tell you the recoil problem is quite serious. I purposely designed my system so the parts would have as little stretch in them as possible. If something doesn't stretch, it can't store energy and can't recoil. The longer and stretchier the rope under test, the more energy available for recoil. The Icetail was relatively short, maybe 3.5 feet, and very non-stretchy. What's more, it snubbed itself off because two strands remained intact.

I really saw some recoil today when breaking a piece of 5/16 inch Tenex. It was only about 30 inches long, and broke near the middle. The end connected to the hydraulic cylinder and load cell sent them flying about 2 feet in spite of an inadequate snubbing system I had in place. I am leaning towards a positive snubbing system in which a rope, as strong as the one under test, is connected in parallel with the test rope. Before anything can fly more than a couple of inches, the snubber goes taut and then stretches to absorb the kinetic energy of the recoil. As I say, this is still a work in progress.

 

[pdqdl]

Try the blankets. It works, even for chunks of metal.

The more weight, the more mass you will need to absorb the flying chunks.

 

[moray]

Pdqdl, I didn't mean to ignore you; I have been thinking about using some old rug that I have. A rug (or blanket) folded over the experiment should confine the recoil, one would think. But I don't think it would absorb much of the momentum. The wadded-up recoiling rope would whack the nearest clevis like a 2-or-3 hundred MPH baseball without its cover. The clevis pushes the load cell. Whack! The load cell pushes the cylinder. Whack! It's those Whacks I want to minimize. I can help myself a lot by moving the load cell to the other end of the cylinder, and I think I am going to have to do that.

BTW, I am inclined to believe the 10,000 lbs from yesterday, even though I certainly have done no calibrating at that end of the scale. The clean 5/16 inch Tenex I broke today measured 4400 lbs, slightly below spec. Could a sample of Icetail actually be 17% or 18% stronger than the published average? Would that be merely unusual, or unbelievable? In any event, I will stay alert for the chance to weigh a truck or some other large known weight.

 

[pdqdl]

Try the blanket, let me know how it does.

What thickness Icetail were you breaking? I thought you were doing smaller stuff used for eye-to-eye prussics. That doesn't come in anywhere close to 10k.

 

[moray]

Is there more than one size? Mine is the 5/16 in, 8500 lb, stuff meant for prusiks. 80% technora, 20% polyester. I will try the blanket.

 

[moray]

wet blanket

I tried the blanket today--actually a peice of rug. The rug was about 5 ft by 3 ft, more than enough to entirely cover the experiment. When the rope broke at 4460 lbs, the recoil still caused the cylinder and attached load cell to jump back a foot or more. If the rug did anything useful, it wasn't apparent.

I then implemented a rope snubber made from 7/16 stable braid. Once the experiment was tensioned to 500 lbs or so, I adjusted the rope so there was about 4 in of slack. There was still slack when the test piece broke, but the snubber clearly did a good job of absorbing the recoil, and the big hardware didn't move more than 2 or 3 inches.

 

[pdqdl]

I am a bit confused. The rope breaking might slap into the cylinder, but it won't have enough momentum to move the hydraulic cylinder.

I suspect that where your cylinder is anchored has more elasticity than you are giving it credit for. If the cylinder is moving, what is it attached too?
(some pictures of the whole rig would be nice!)

My suggestion for the blanket was to control frayed rope and small flying metal attachments, not the entire hydraulic cylinder. If the cylinder is attached with a long chain or cable, see if you can shorten that attachment in some way, then extend the other end longer. Any long stretch of material will have some stretch, which will rebound. The longer the run, the more re-bound you will have.
(But of course, with your understanding of physics, you will know that. I'm just thinking maybe that didn't cross your mind)

An alternative method might be to pre-tension the chain holding the cylinder back in the direction that the splice is being loaded with a second chain. When the rope fails, the cylinder is still under tension and goes nowhere.

By the way: if the rug was actually some really stiff old carpet, that wouldn't have worked to well. You need a flexible fabric that conforms to the rope (or other moving objects) so that the momentum is absorbed and transferred into the fibers of the fabric.

It sounds like your rope snubber is largely doing what I suggested.

 

[moray]

My bad, I think all your criticisms apply to me.
(1) The rug was quite stiff.
(2) There are 21 feet between anchors, so a lot of chain is available to stretch.
(3) Under tension, the cylinder is about 3 inches above the little table it sits on, and its chain anchor is about 8 inches off the dirt. When the test rope breaks, even if there were no recoil in the test rope at all, the cylinder has to fall 3 inches and the sagging chain will drag it back toward the anchor.

So I can't actually tell how much of the movement is due to rope recoil.

I like your pretension idea. If I can somehow pretension with some rope attached to the same clevises as the experiment, that will make for an improved snubber. I should probably stop letting the cylinder fall on the table by putting something thick and soft under it. The whole arrangement is embarrassingly primitive at this point! When it finally looks pretty, I'll post a picture.