Load Testing Results

[Cotemar]

Originally Posted by estarzinger

The beauties of the soft shackle is that it is 'just a piece of rope' that anyone can make up in about a minute. And for most people, the diamond is plenty strong enough. I mostly use soft shackles made with 1/4" line on Hawk, but I realize after this testing that 1/8" line would be strong enough for almost all my applications.

Some people really get carried away with these Soft Shackles, not realizing the small ones may be over kill for their application, so they size it up.

Tests like yours may help cruisers understand what soft shackle sizes are really needed to get the job done.

Keep up the great work.

[SVNeko]

Originally Posted by estarzinger

There is some indication that a shorter bury is better than a longer one (because it is hard/impossible to get equal tension on the bury and the cover), and the 'no bury' 'Kohlhoff' may be best, but it is hard to sort that out in the data from the constructional variation noise.

Although the Kohlhoff style doesn't suffer from the cover/core tension differential, wouldn't that be offset by one leg of the Kohlhoff bearing more of the load unless the two legs were EXACTLY the same length?

[estarzinger]

I found someone else who had machine sewn and broken a lot of webbing samples, but had not done much to analyze his data. I cleaned up the data a bit and did the most basic analysis:

The overall summary number is that his samples show an average loop breaking strength = 125% of the thread tensile. However there is very significant variability, with a range from 66% to 309% and a standard deviation of 49%. There was no correlation at all with the sewing pattern/shape. And no other way to explain the big variability. I am surprised by his (and my very limited) data because I thought one of the big advantages of machine sewing was its consistency/repeat-ability.



So, I am still puzzled by our machine sewing results.

[Paul Elliott]

I wonder if you're seeing a chaotic chain-reaction failure here. With low-stretch line, and low-stretch thread there's going to be less load-sharing among the stitches. although the "more stitches = more strength" trend line obviously shows some load-sharing. Still, there may be more variation in the load seen by an individual stitch, and when that stitch breaks the thread slips and the whole thing unzips.

When the thread breaks do you have lost of broken stitches, or just a few and the rest pull out? If it's not obvious how the stitches are failing you might try a controlled experiment where you load up a stitched sample just under the breaking point then deliberately nick a stitch with a knife.

[estarzinger]

Paul,

That graph is all dacron thread in nylon webbing, so it is not really 'low stretch'.

usually when my samples break, it looks like many of the stitches break at once - eg there are lots of little pieces of thread left in the webbing rather than one or two long pieces . . . but I have not examined it closely.

I am just surprised that we are getting less 'efficiency' in machine stitching than hand stitching . . . I am wondering if there is something fundamentally weak in the geometry of the machine lock stitch.

[Cotemar]

Your line is pulling in a linear direction and your stitches are in shear.
Shear on a thinner thread will always go before a line in a linear direction.
It’s sort of like a scissor action on the thread.

The reason that stitches hold on strapping is because they do a box and then an x inside the box as wide as the strapping.

A line does not have the surface area to do this kind of x box stitching on.

[Paul Elliott]

Originally Posted by estarzinger

That graph is all dacron thread in nylon webbing, so it is not really 'low stretch'.

One of the great tragedies of life is the murder of a beautiful theory by a gang of brutal facts.
– Benjamin Franklin

[estarzinger]

Originally Posted by Cotemar

Your line is pulling in a linear direction and your stitches are in shear.
Shear on a thinner thread will always go before a line in a linear direction.
It’s sort of like a scissor action on the thread.

The reason that stitches hold on strapping is because they do a box and then an x inside the box as wide as the strapping.

A line does not have the surface area to do this kind of x box stitching on.

This IS a test of stitching on webbing, not on line!

And unfortunately that's not what this data shows. That's what my comment about "no correlation between stitch pattern and breaking load" meant. The sample were stitched with a wide range of thread patterns, from all "across" to all "Down" and various box shapes in between. . . . And there was no correlation to stitch strength.

I don't know why but that's what this data set shows.

Originally Posted by Paul Elliott

One of the great tragedies of life is the murder of a beautiful theory by a gang of brutal facts.
– Benjamin Franklin

But, your comment raised the thought that the nylon webbing is stretching more than the polyester thread. It might be better if they had matched stretch or if the thread was stretchier.

[Paul Elliott]

Originally Posted by estarzinger

But, your comment raised the thought that the nylon webbing is stretching more than the polyester thread. It might be better if they had matched stretch or if the thread was stretchier.

Interesting! I can't think of a reason why stretchy thread wouldn't be best. It wouldn't have to match the stretch of the webbing (or line), the stretchier the better for the thread. You would just have to use a lot of stitches for strength.

[Cotemar]

Or try using Tenara thread

[estarzinger]

Originally Posted by Cotemar

Or try using Tenara thread

Tenara's big selling point is that it is UV resistant. I use it when doing sail covers and such. The sail makers I have worked with don't like it because it is slippery, and thus hard for them to get constant stitch tension on their machines.

Do you know what it's stretch characteristics are vs Dacron and dyneema? I have never seen those specs. I know that it is chemically Teflon (polytetrafluoroethylene) but I don't know what it's stretch is like?

[Cotemar]

estarzinger,

Yes, Tenara thread is a Teflon thread. The slippery attribute makes it great for hand stitching, but not so good on the sewing machine. I always use it for hand sewing when on the boat.

[estarzinger]

google . . .

UHMWPE
Young's modulus 113,500
Elongation 3.6%

PTFE
Young's modulus 580
Elongation 450%

Source: MatBase

[estarzinger]

I am just looking at the sewing data a little more closely.

I notice three things:

1. The very high stitch count samples all have negative residuals (eg are under the average strength). I do not have data on the length/area of stitching, but this suggests you weaken the sample by jamming too many stitches in (which sounds logical).

2. The 95% confidence interval for the breaking lbs/stitch is 24.7lbs to 30.5lbs, or (rounding a little) 110% to 140% of the thread tensile.

3. The residuals are a close fit to the normal distribution (excluding the very very low and high samples). So, the linear model is probably pretty good. We just need to find additional variables that explain more of the variation.

[thinwater]

a. I'm not sure we've explained why the thread modulus needs to match the rope modulus. Could be true, but I've not seen a provable explanation. I can lace a tramp to a rigid boat or tie myself to unyielding rock with nylon. In fact, the nylon will have a greater ability to self-equalize, like a lacing. By the way, I've always found high mod line on tramp lacing to be pretty silly--the wrong material for the application.

b. I've broken ~ 50 samples over the last few months--line and webbing--and I have NEVER seen variation beyond 15%, high to low within a given material combination. Most of that variation was due to changes in stitch count (I presume some samples with fewer stitches had less uniformity). These were all hand sewn, and though I may be experienced, I'm not an obsessive craftsman.

c. They don't zipper. In fact, I've cut threads at full load intentionally to reduce strength. At most the efficiency drops 10-20% if the count is sufficient.

d. The scissors analogy makes no sense, the bends are not that sharp in proportion to the thread; if you see the thread pattern under load, it is more like a lacing.

e. There is plenty of area on line; I've broken lines up to 5000 pounds with only 2 inches of stitching.

f. I've seen no evidence that box stitching is magic. I think there are certain practical matters; it gives a peal resisting row at all edges and it is simple. But climbing gear is all bar tacked. I'd be happy to look at the numbers, but my testing indicates that a number of cross rows is more efficient.

g. Once zippering starts, there is no pattern that will resist it; it's fast. Better, figure out where it is starting and stop it.

h. Machine stitching depends on even tension because it is inherently unable to adjust to the load (interlocking loops). On short runs I can see that this is a challenge. Bar tack machines do not suffer from this because of the high thread count (that last is a guess).

i. Whipping twine will generally be more UV resistant than fine thread, but in any event, stitching will require protection. I can think of many proven methods.

j. Tenera thread, to my knowledge, is limited to 19#. Even after 5-8 years, plain 50# nylon twine will hold 20 #, and portions out of the sun far more. Great for awnings, but not very interesting for high load applications, IMHO. And there are heavier twines (#8 is ~ 100#) for larger lines.


Opinions, generally backed by testing. But I'd love to see more data. I'm sure I can learn more.