Load Testing Results

[thinwater]

I typed a long response, then the damn website logged me out and lost it. Arg. A shorter version.


Assume:

• A splice between 2 long bits of webbing (not an eye).
• 25% stretch at high load.
• The doubled zone will stretch only 50% as much as the single webbing.
• Bar tacks, no thread stretch.

If we do an engineering evaluation (what we call deformable bodies) we find the following:

• 1 bar tack. 100% load.
• 2 bar tacks. 50% load each.
• 3 bar tacks. 50%/0%/50%. The center tack carries no load. Niether will any other stitching in the center of the splice. This is counter intuative, but take 2 rubber bands or elastic, conect over lapped, mark, and stretch. You will find the marks do not move relative to each other. No differencial stretch = no load.

In the real world, the thread gives and the bar tacks sink deeper into the webbing, giving some equalization. If the join is 1.5 inches long then the stretch over the join will be 0.22 inches, but only 0.07 inches to the inside bar tack (if there are 4 tacks); webbing and bar tack distortion can handle some of that.

Spectra has proven easier to sew for ths reason; less stitiching is required to reach the same strength because the load is even. The logical extention (the opposing case) is two 2x4s lashed with rubber bands; the boards don't stretch, the bands do, and the load is 100% even.


What do we learn from this?

• Keep joins very compact. Do NOT spread them out.
• Low stetch webbing is easier to splice.
• Low stretch stitiching has little value in terms of ultimate strength.
• Compacting the webbing with loads of stitiching will help because it further reduces stretch. but this is not needed with 4-5 rows of bar tacking.
• More than 5 rows of bar tacking don't help, because the center rows don't carry load. Fortunatly, 5 rows will break the webbing.
• High strength thread helps, since more hold can be packed in to less space. Small stitiches. Totally different from sewing sailcloth. The things you can do sewing webbing and rope would destroy a sail. I never ran into this problem hand sewing webbing because I can pack the thread count into a small space and because the waxed twine equalizes better.

What about a sewn eye? Imagine the above example cut in half (50% of the load carries through). The first bar tack carries the brunt, but now the throat has a pealing force.

• Keep the eye long and narrow. 7 degrees max. If it needs to be wider, use a sling.

What about sewn eyes in rope? Rope is far more compressible which allows greater load sharing, and polyester and hi-mod don't stretch much. This does not really apply.

What of web-to-web friction? thereis some, but it is ONLY in the very imidieate area of the stitch and the pattern won't matter. The total friction will always be some consistant fraction of the total holding power, perhaps about 20%. There is no reason to consider the effect of thread tension, beyond the important matter of wear. The real friction is generated by thread angulation when the real load comes on.

These factors also become unimportant when splicing rope, because the friction principles are very different from spot stitching loads. But it does suggest one more reason why the taper on a rope splice is critical; it is more than just a stress riser.

----

Throw rocks. I'm just a chemical engineer playing at mechanical engineering. But the math follows climbing gear practice.

[estarzinger]

Back to amsteel splicing a minute . . .bury length . . .

First I am not sure what the official definitions of "bury" are - whether it includes the taper or not. I suspect 'officially' it does not but that most people do in practice (they measure out and make the bury and then do the taper). Also, the bury needs to be measured on the 'core' and not the 'cover' because If you measure it out on the cover before you bury, the cover then gets shorter when you do the bury (because it get stretched out/increases in diameter).

So . . . 7/64" amsteel. I made up three pieces all with 1" tapers, with 1", 2" and 3" 'solid' sections before the taper). If you count the taper as 'bury these are 18, 27 and 37 bury ratios. If you do not count the taper they are 9, 18 and 27 ratios.

The shortest one slipped at 680lbs. The middle one broke at the taper at 2100lbs. and the longest one broke at the taper at 2080 lbs. (those two are statistically the same)

So, for a slow pull, somewhere between 18 and 27 diameters (counting the taper) makes the bury 'secure'. It might well be completely different for a shock loading pull.

[Lee Jerry]

Originally Posted by estarzinger

Back to amsteel splicing a minute . . .bury length . . .

First I am not sure what the official definitions of "bury" are - whether it includes the taper or not. I suspect 'officially' it does not but that most people do in practice (they measure out and make the bury and then do the taper). Also, the bury needs to be measured on the 'core' and not the 'cover' because If you measure it out on the cover before you bury, the cover then gets shorter when you do the bury (because it get stretched out/increases in diameter).

So . . . 7/64" amsteel. I made up three pieces all with 1" tapers, with 1", 2" and 3" 'solid' sections before the taper). If you count the taper as 'bury these are 18, 27 and 37 bury ratios. If you do not count the taper they are 9, 18 and 27 ratios.

The shortest one slipped at 680lbs. The middle one broke at the taper at 2100lbs. and the longest one broke at the taper at 2080 lbs. (those two are statistically the same)

So, for a slow pull, somewhere between 18 and 27 diameters (counting the taper) makes the bury 'secure'. It might well be completely different for a shock loading pull.

Evans,

Great stuff. Thanks for sharing all of it.

As for these splices, these buries and tapers seem a little short. I don't have any references in front of me (not at home), but I seem to recall (perhaps incorrectly) that the bury should be around 18x diameter for "normal" stuff and 36x for slippery, high-tech stuff (like Amsteel). That would correspond to about 2" and 4" for your 7/64" line. For the taper, the minimum would 12:1 and much better is 24:1. Therefore, 1.3" and 2.6" for this line. (And if not obvious, the taper is separate from the bury.) When I taper, I go down to a single strand (or pair): cut a strand to start the taper, move a half inch (or whatever is appropriate to the line size and construction) and cut another strand, move a half inch and cut another strand...

Any chance you could test a 4" bury with a 2.5" taper?

(After typing this up, the numbers don't seem as 'off' as they first appeared. Perhaps it's just that I'm not used to line that small. Usually 5/16" and up, therefore longer buries.)

[estarzinger]

wet vs dry line . . . .

For dacron double braid . . .bowlines . . . 3 pull average breaking strength 1780lbs dry vs 1720lbs wet, with a 54lb stdev. So that's not a significant difference at the 95% level.

All the literature says nylon looses 10-15% when wet, but that polyester is pretty much the same. I will pull some nylon bowlines when I get a chance.

[estarzinger]

On wet nylon . . . . 3 pulls dry = 1190lbs and wet=1107lbs, with a stedev of 56 and a 95% CI of the mean of +- 64, so that is a statistically significant result (7% less when wet).

[thinwater]

Regarding hand stitching webbing, the results look a little different, but in many ways, perhaps more repeatable. below is a dead-average result. strength is 92% of webbing strength, with a total range of 4% over 8 trials. 9/16-inch Bluewater Climb Spec webbing, 2300-pound manufacturer break strength. It is impossible for me to know if this is a good number, since I have no splice alternative. I tested 2 samples with overhand loops (best knot in webbing) and they failed in the knot at 1600 pounds (70%).

Stitched with doubled 50# whipping twine, snugged at each stitched.

Sorry about the big images, but I though folks would want a close look at the thread, particularly near failure.

Before


At 2000 pounds (87% of BS, 94% of load held. This is as close to breaking as I managed to photograph. Viewed from the side, the ends are pealing up.


Notice that thread angulation increases towards the tail. This is the unavoidable result of stretch within the splice. The only way out is Spectra webbing, which is good for some things and not others. Notice the slight spreading at the throat. The throat angle of this sample was above optimum, though I doubt this was the failure point; however, it should be obvious that at wide angles it would be.


After (2150 pounds, 93% BS). The surviving splice is completely unchanged visually (the little thread trim-off tail hasn't even pulled in). The failed splice exploded, breaking thread and webbing in an instant. Notice that the twine fails slightly longer than machine thread--not every pass fails at the same instant, though most do. In the most extreme cases, one adjacent stitch pulls through. It does not zipper.



I have played with a number of patterns--Xs, outlines, bar tacking--but simple dense stitching seems to work best when hand stitching. Because of the size of the needle, bar tacking is not effective with whipping twine; breaking strength is about 5% lower and the webbing tears.

The thread load efficiency is only about 65% in this example. That seems to be about the best I can get in webbing. I have tried more stitching also; no improvement.

[Seaworthy Lass]

Originally Posted by thinwater

The thread load efficiency is only about 65% in this example. That seems to be about the best I can get in webbing. I have tried more stitching also; no improvement.

Have you tried doubling the thread when hand stitching?

[thinwater]

Originally Posted by Seaworthy Lass

Have you tried doubling the thread when hand stitching?

Always doubled, visible in the photos.

I get significantly higher efficiencies at lower load factors and sewing line, but not with webbing near the limit of the webbing. High thread efficiency isn't the point, high splice efficiency is.

If I double the amount of stitching I see about 3-5% more splice efficiency, but even that is not certain (measurement error).

The real point of doubling the stitch count, which I do for my own projects, is durability. Day-one doesn't matter. Of course, in most rigging there is a 5-10:1 safety factor too, so the cushion is large.

[Seaworthy Lass]

Originally Posted by thinwater

Always doubled, visible in the photos.

I can see the doubling now that I have enlarged the photo .

Originally Posted by thinwater

I have played with a number of patterns--Xs, outlines, bar tacking--but simple dense stitching seems to work best when hand stitching.

I use a "Backstitch" when hand sewing anything important. It makes a big difference in strength generally for a minimum number of holes in the fabric. If you are experimenting it may be worth giving it a go.

On the back the stitch size is doubled, but overlapped:

[thinwater]

SL, I'm on it!

[Seaworthy Lass]

Originally Posted by thinwater

SL, I'm on it!

If your needle diameter is large and has therefore left a largish hole, it is possible with a bit of care to insert the needle exactly into the previous hole, halving the number of holes needed (not as hard as it sounds).

[thinwater]

Originally Posted by Seaworthy Lass

If your needle diameter is large and has therefore left a largish hole, it is possible with a bit of care to insert the needle exactly into the previous hole, halving the number of holes needed (not as hard as it sounds).

Yes, I understand.

Failure. Too much stress in too small an area. It simply pulled through the webbing at 15% lower load. A good idea that doesn't seem to apply to tubular webbing. 2 trials. I'm a little bummed. I've done this in sailcloth many times, but this different. Idon't think holes matter so much in webbing, or bar tacking would not work. Puzzeling.

[estarzinger]

Sewing polyester webbing and V138 (very heavy) thread.

This was a test of 'small pattern' sewing samples to identify the component strength of bigger stitching patterns.

I tested three variables: stitch type (straight vs zigzag), stitch count (20 vs 40, the thread ends were tied off to ensure there was no unraveling from the ends), and stitch orientation (across the webbing vs down).

Raw data
Straight/20/across 660lbs (33lbs/stitch)
Straight/20/down 710lbs (35.5/stitch)
Straight/40/across 1340lbs (33.5/stitch)
Straight/40/down 1250lbs (31.3/stitch)

ZZ/20/across 770lbs (38.5/stitch)
ZZ/20/down 800lbs (40.0/stitch)
ZZ/40/across 1290lbs (32.3/stitch)
ZZ/40/down 1260lbs (31.5/stitch)

Averaging across the tested features:
Across= 29.0/stitch vs down = 28.7/stitch - no difference
Straight=28.3/stitch vs ZZ=29.4/stitch - a small difference toward zigzag
20 stitches=36.8/stitch vs 40 stitches=32.1/stitch - a big difference toward fewer stitches

Conclusions, (a) like in nylon webbing, there is a diminishing marginal return to stitch count, and (b) the two highest efficiency samples were the two 20 count stitch zigzags (which I made as close to bar tacks as my machine can do).

[thinwater]

Originally Posted by estarzinger

Basically a series of very close together but relatively long individual zigzag stitches (eg a lot of zig but not much zag ). The individual stitches are 'down' the webbing' but the bar is 'across'. This is ideal for the sort of stress riser loading I have been seeing.

Attachment 73937
Attachment 73938
Attachment 73936

By the way, to really see clearer 'stitch pattern' test results it may be that I need to get some spectra webbing, so that webbing stretch will not compensate/hide the pattern effect.

I don't think you WANT to separate the stitch pattern testihg from the material, just as knot eff is material dependant. The value of what you are doing is in realism.

[estarzinger]

I just did one 'composite sewing pattern' a 1" square with a single diagonal across it. It broke at 2080lbs. It is a little hard to get an exact stitch count because the stitch length is different on the across vs down legs (because the down legs are only on one walking foot and thus move more slowly than the across which are on both walking feet) . . .but I am guesstimating 59 stitches = 35lbs/stitch.