Comparing techniques used for eye splicing double braid polyester (=Class 1)

[Seaworthy Lass]

Originally Posted by s/v Jedi

About the goal of this thread; learning more about rope and splicing is #1 of-course. Apart from that, it seems that finding the strongest splice must be among the goals. We had a similar exercise with the Dyneema soft shackle which lead to great improvements.

But this is not a soft shackle. For an eye splice, the maximum achievable is 100% of line strength, while this is 400% for the soft shackle. The latest and greatest soft shackle is now at 230% so just over 50% of theoretical maximum, while the eye splice is reported to be at 90% or better, which is 90% of theoretical maximum.

A few comments:
Firstly the 90% value is just for constant pull. What about cyclical loads that are what many splices are subjected to?

Secondly, not all splices may have a system strength that is 90% of line strength. I can find no studies comparing splices.

I doubt very much all the techniques result in the same strength.
Which is best?
That is the burning question to start with.
Then we can perhaps see if it can be improved.

Originally Posted by s/v Jedi

This raises the question: why are we still below 100% then? I think the answer is that a splice, even though stronger than 100% itself, is causing a disturbance within the rope and this is the location of the failure. Logic (at least my logic) seems to dictate that balancing the core and cover leads to a stronger end result, as well as tapering.

Yes, I agree absolutely.
I made that comment here in post #90 and I thought it important enough that I put it in red .
This is why it freaks me out seeing how much core is being cut off in some methods and what promoted starting this thread!

Originally Posted by Seaworthy Lass

NOTE:
If there is excess cover below the core then the relationship between core and cover will then be different. The breaking strain of the line will probably then be reduced and the line will break somewhere along its length when it does break, not at the splice. It may even break some distance from the splice.

The problem with this is that unless you know exactly what load caused the line to break, you will say "Nothing wrong with my splice, it held. It was the line that broke".

SWL

Originally Posted by s/v Jedi

But most of the lines we use are put to work at a maximum of 20% of breaking strength. Will we change that to a higher percentage when we get a better splice?

Splice fail though. Is this occurring at only 20% of maximum line strength?
If so, it means cyclical loads or other factors (that are not tested for) may have a huge affect on splice strength.

Originally Posted by s/v Jedi

I therefor think our goal should be to make the splice as easy as possible without giving up too much. In the end it comes to testing and the only tests I've seen are the published papers from the UK and the guy with the two cars/vans. Maybe it's possible to create a number of splices and send them to someone who can test and likes to join the effort. We could do something like a Samson, a Selma and the Premium splice.... and the version we come up with if other than those.

Getting any decent testing done may be difficult. There is no money for manufacturers in this, so funding will be limited. This could be a masters/phd thesis for someone. Both the protocol and analysis need to be good to give statistically significant data. There are so many variables that it would be beyond the scope of any one study to cover them all, so results will be limited.

So I think all we can realistically do is look at current studies that have been done on both line strength and eye splice strength for double braid polyester, then look to see what methods minimise the risk of strength reduction (either in the splice itself or the line after changes have been induced by the splice).

This may enable us to suggest which of the methods is likely to be best.
It may also lead to suggesting possible ways of increasing strength.

SWL

[Seaworthy Lass]

Juho, I will read and respond to your posts #108-114 later (2:00 am here). I have not forgotten.

I think it is important that we each read all posts carefully. I do find feedback is very helpful and I am sure others must as well. I do have a tendency to get enthusiastic and rush ahead .
So I will do this before tacking the spreadsheets and videos and other reading material.

SWL

[Alan Mighty]

Originally Posted by Seaworthy Lass

Alan, this link you gave won’t open. Could you please check it:
http://www.cruisersforum.com/forums/www.paci.com.au/downloads_public/knots/15_Report_knot-terminations_2006.pdf"]www.paci.com.au/downloads_public/knots/15_Report_knot-terminations_2006.pdf[/URL][/FONT][/COLOR]

Also, was there a link to Milne & McLaren’s work somewhere? I can’t find it

The paci.com.au URL links to the ready-to-print version (the printer's version) of the Katherine Milne & A J McLaren paper. It should work if you point to: http://www.paci.com.au/downloads_pub...tions_2006.pdf.


If not, get the same paper (in standard pdf resolution, a smaller download) direct from:


http://personal.strath.ac.uk/andrew....%20McLaren.pdf


and if that doesn't work, go to Andrew McLaren's project page (Dr. Andrew McLaren's Project Page) and read down until you find K.A. Milne & A.J. McLaren "An Assessment of the Strength of Knots and Splices ..."


Strathclyde U has the equipment (a Tinius Olsen screw-driven fatigue tester and an Instron 8802) to test both static strength and also cyclical strain - for the latter, they set a splice (or a knot) in a machine, set it to a frequency of cyclical test and let it run. They monitor the break in three ways: a digital camera taking photos to see where the rope or splice breaks (but the camera works on a certain frequency of photos, sometimes it misses the break), a two computer techniques (one involving a plotter) to chart the strain before and at the break.


Strathclyde uses the Marlow splicing technique, which to my eyes just slightly abbreviates what I call the canonical Samson Rope or Brion Toss techniques (SR and BT follow the same steps, just that BT increases two dimensions just a little and almost always specifies dimensions in rope diameters instead of fid lengths).


Strathclyde uses Liros cordage. And starts the tests at 75% of published breaking load and then lets the machines ramp up. All of their tests show surprising variance in the strain at which the rope breaks. Their tests (which are not the same as the tests rope makers use) usually show rope breaking at a lower BL than the ropemaker. And some 'identical' splices do better than others - so they calculate the statistics in an attempt to handle the variation.


Some splices seem to be more than 100% of rope strength. Others less. Just the way it goes. Reality is chaotic. Life experience is unsatisfactory/full of misery (but in misery there is unbounded grace, according to one authority).


I'm off to buy the morning newspaper and get breakfast at my local cafe. Be back in a few hours.

[Alan Mighty]

Originally Posted by Alan Mighty

Strathclyde U has the equipment (a Tinius Olsen screw-driven fatigue tester and an Instron 8802) to test both static strength and also cyclical strain - for the latter, they set a splice (or a knot) in a machine, set it to a frequency of cyclical test and let it run. They monitor the break in three ways: a digital camera taking photos to see where the rope or splice breaks (but the camera works on a certain frequency of photos, sometimes it misses the break), a two computer techniques (one involving a plotter) to chart the strain before and at the break.

As usual (when writing in haste), I'm wrong. The Tinium Olsen does the static strength test. The Instron 8802 does the cyclical fatigue tests.


The Mechnical Engineering department at Strathclyde U, with Andrew J McLaren and his students doing their projects, is the only place I know that has published cyclical strain tests of knots and splices. Others may have done cyclical fatigue tests. I've not seen them publish the figures.


Everything we do as cruisers is about cyclical fatigue: every wave, every variation in wind from change of apparent wind angle to puffs and gusts of wind speed, every cranking of a winch: all are cyclical strain. Not static strain.


Splices can break, but usually the rope breaks near the splice (and we don't know whether it is because of an imbalance between cover and core: remember, a Class 1 double braid is 'balanced double braid', meaning that cover and core share the load.


Contrast that to Class 2 double braid, with a high modulus core. The core (Dyneema/Spectra, PBO etc) carries the load. The cover is there to protect the core from abrasion or UV, and to give you a nice hand feel.


Splices can slip. When tested statically, the people at Strathclyde found that you have to reduce the bury, the length of tucked cover, to 1/3 or even to 1/5 of the standard length to get a reliable slip. I don't think they have tested that in a cyclical way (Heather McLaren flagged it as research that ought be done).


Bob Norson with his 2 vehicles doing a snatch force test never found a slipped splice - his rope broke near the splice. The breaks recorded by Strathclyde Uni are in similar locations.


The increased density of polyester threads at the splice may mean that the splice almost never breaks.


Question is why the rope almost always breaks near the splice: if the test was done many times, the break could occur anywhere along the rope.


If


(1) imbalance of cover and core is caused by the splice; or


(2) abrasion of the core fibres is caused by pulling it out of the cover and then milking it back in,


then you might expect the break to be near the splice.


So the ideal splice is probably:


(a) easy to understand;


(b) buries a goodly amount of cover and core; and


(c) does not abrade or break many fibres and threads (which may mean that the section of core through which the cover is tucked is a particular length, not too long because that causes bunching up and bending and abrasion of core fibres) and perhaps not too short).

[Seaworthy Lass]

Originally Posted by Juho

One possible trick would be to connect a thin cord to both C and C' before the crossover is milked in. Those cords would have parallel marks further down. If CC' grows during the milking and stretching process and loading, we should be able to measure the amount of change from the two marks. The cords now hang out from the hole at R, and difference in the position on the two marks can be measured.

Good idea, but:
It would have to be done in a way that minimised the risk of it inadvertently affect CC’.
It would also have to be done enough times to determine if what we are finding is statistcally significant. It may vary between rope brands and for different rope diameters. That is a lot of line to buy and a lot of splices to make .
Then we don't know the significance of it shifting unless load testing is done on each sample.

Originally Posted by Juho

In the New England Ropes video that s/v Jedi just posted above they milk the rope after extracting the core, after which the core in clearly longer than the cover. I have had similar experiences with my ropes. Especially if the cover and core have been melted together at the end you may find surprises when you let them loose. Some sort of milking may thus be needed to rebalance a rope whose cover and core have become misbalanced for some reason. The milking process should emulate rope under load.

I don’t think the cover and core have “become misbalanced” and as far as I am aware, we have no reason to think that pulling the core out then milking it by hand emulates what the rope does under load.. Hopefully with good quality rope the manufacturer has explored what relationship is best between length of core & cover and the rope is made that way. It is possible cheap rope could have any combination.

I think the reason for milking is purely because if we are physically incapable of returning the cover:core relationship to how it was originally along the length that is secured off with a knot, then it will be much harder (not impossible, just harder) to milk the splice properly.

If it is important not to tamper with relationship between cover & core for the line below the eye (and I suspect this is the case), then the core should be undisturbed before the mark at the extraction point (that we are calling P’) is placed. I think this initial milking that some methods suggest (and it is not all that do this) is a bad thing.

If we can't return the cover:core relationship to how it should be, we should really be focussing on disturbing it as little as possible. Just enough to enable the minimum amount of core to be pulled out for the splice.

That is maybe why a couple of methods I have looked at (Samson & New England from memory) are very specific about trying the knot only 5 fid lengths from the second mark made on the cover. They may not just be being pedantic. I can see that there may be a good reason for this - a bare minimum amount of core is disturbed.

Originally Posted by Juho

The Samson video was interesting because they cut quite radically half of the strands of the embedded core away at point R'. I think that's about as far as one can go. The load is shared by two covers and two cores there, so it is possible to cut half of the strands away without destroying the overall strength of the splice.

They are the only ones who do this. I agree you could theoretically lose half the core here (or, as Jedi has pointed out, even all the core if the cover provides half the strength in new rope ) without system strength suffering. It may even be beneficial, as it reduces bulk just below the eye. This makes milking easier and distorts the cover less. The taper itself will also distort the cover less.

Without load testing we have no idea if this is a good or bad thing to do.

Originally Posted by Juho

The strangest thing I found in this video was that they embedded the end of the core inside the cover at point R. I.e. no distance between R (the loop) and the insertion point. Samson used point T (C in out latest terminology).

The same thing is done in the Marlow video. It does mean a full tail length of cover is buried in the core. This may be very beneficial.


Post #110 is the last one I need to reply to, but it is lengthy and I need a clear head for that, so I will do it first thing in the morning.

Alan, I will look at your spreadsheet after that.
Thanks for the links to the Milne & McLaren paper.

SWL

[Seaworthy Lass]

Originally Posted by Alan Mighty

Life experience is unsatisfactory/full of misery (but in misery there is unbounded grace, according to one authority).

In suffering or in misery? There is a big difference.

I've found misery only makes most people mean and cranky, not full of grace.

[Alan Mighty]

Originally Posted by Seaworthy Lass

I suffering or in misery?

I've found misery only makes most people mean and cranky, not full of grace.

Part 1 of that "Life experience is unsatisfactory/full of misery" is one of the core teachings of Buddhism. The Pali word 'dukkha' (Sanskrit equivalent duhkha) has been variously translated as "unsatisfactory", "suffering", and "full of misery". (parenthetical note: some translators take the stance that the only words that have a one to one correspondence between two languages of different language families are the number words. Almost every other word has a connotation that means translators need to know two different cultures, each of which is changing all the time anyway)


Part 2 "in misery there is unbounded grace" is an English translation of a teaching from an Roman Catholic Pope. I forgot which one; I think it's the one who used to wear a hair undershirt all the time and was stabbed through the hair shirt when he visited Manila in the Philippines (thus for the first time revealing that he always wore a hair undershirt).


More (or Less, depending on your viewpoint) importantly:


Yes, Marlow has no cover length between the Reference Point R and the Taper/Crossover (T in Samson, C in International Standard Name). There are many published splice techniques based on the Marlow splice. Rather than put that dimension as 0 in my spreadsheet (because it gives me a headache to do so) I arbitrarily decided to focus on the 'canonical' splices that have a non-zero RC dimension and the others (Marlow, two splices from Premium Ropes).


* note that Premium Ropes even has a video of the Selma Fid splice, so Premium seems not wedded to any one splice technique and has made videos for several different splice techniques. Whether that confirms the suspicion that they've not tested the static or cyclical strength of the different splice techniques is a good question.


I might reconsider my arbitrary decision about that RC dimension. Comparability is always a good thing, even if it involves a headache.

[Alan Mighty]

Originally Posted by s/v Jedi

But this is not a soft shackle. For an eye splice, the maximum achievable is 100% of line strength, while this is 400% for the soft shackle. The latest and greatest soft shackle is now at 230% so just over 50% of theoretical maximum, while the eye splice is reported to be at 90% or better, which is 90% of theoretical maximum.

This means we have very little room for improvement, while the soft shackle used to be at 180% which is well under 50% of theoretical maximum. So where are we at? I believe that our splice is already at well over 100%. Like one of the documents explains: when using just the cover, we can already achieve a 100% splice because the load splits over two legs. This raises the question: why are we still below 100% then? I think the answer is that a splice, even though stronger than 100% itself, is causing a disturbance within the rope and this is the location of the failure. Logic (at least my logic) seems to dictate that balancing the core and cover leads to a stronger end result, as well as tapering. Even when tapered to the ultimate, the first little strand of extra core or cover that appears may still be the cause of heat generation under load, or cause some fibers to be bruised or even snap, leading to the avalange of fibers failing and finally the rope breaking. Our effort may be futile. Or they may bring 5% better strength.

But most of the lines we use are put to work at a maximum of 20% of breaking strength. Will we change that to a higher percentage when we get a better splice? how about rope stretch and creep if we do? Personally, I go to Dyneema quickly when strength, stretch and creep come into play. I even had Vectran but shy away from it after having experienced it's vulnerability to UV exposure. Samson Amsteel Blue is pretty price efficient.

I therefor think our goal should be to make the splice as easy as possible without giving up too much. In the end it comes to testing and the only tests I've seen are the published papers from the UK and the guy with the two cars/vans. Maybe it's possible to create a number of splices and send them to someone who can test and likes to join the effort. We could do something like a Samson, a Selma and the Premium splice.... and the version we come up with if other than those.

Thanks s/v Jedi for more of your excellent points. Sorry that it takes me until my weekend to respond to them.


1. On the strength of splices compared to the strength of the rope: knot and splice specialists agree that a splice is around 90% efficient. Some of the tests done at Strathclyde Uni have revealed individual splices over 100% efficient (not many, and the appearance of those over 100% may be just statistical variation).


It would be optimum if all my splices were the lucky 108 percenters, no? And I'd be happy with 100% or the high 90s%.


Yes I agree with, we have little room for improvement.


Bob Norson (the Aussie guy with the abbreviated splice techniques who did a snatch test with a van and a lightweight 4WD) showed how easy it is to make a splice that does not slip.


I've talked to one ropemaker who makes the highest quality polyester double braid I can find and afford. He (a company representative) is complacent and also thinks there is little point in doing more research (but then he won't have to climb up my mast to recover a broken halyard in what likely will be lousy conditions).


Brion Toss, I think it was, said something on the lines that when we exert all effort to do the best (make the best splice, use the best appropriate materials), we build in a reserve that allows us to neglect or stop worrying about things.


So part of this thread is about building in that reserve, in this case in splicing.


2. With polyester double braid, we hope that we are only subjecting our lines to 20% of breaking load. Once over 20% (such as if I reef late), then I likely exceed the limit of elastic elongation (and permanently stretch the line).


3. A Class 2 rope does have a massively higher Breaking Load and hence a much higher working load. But is it appropriate if my sails are Dacron (i.e. polyester fabric with various fillers etc applied after some sort of heat setting treatment)?


Let's say that I'm in a squall and am late reefing my mainsail. If I have a Class 1 polyester double braid halyard, the line will stretch perhaps over its 20% of BL and might be damaged in the sense that I've exceeded its limit of elastic elongation. So the line is permanently stretched. What happens to the sail? Perhaps the stretch in the polyester double braid halyard will be enough not to push the sail polyester over its 20% elastic elongation limit (within limits, stretched sails can be repaired, but I'm sure you know that a stretched sail is not much fun).


Now let's say I have a Class 2 cordage in the same halyard in the same conditions. A Class 2 halyard, with a Dyneema/Spectra etc core, will likely not stretch at all in those conditions. My Dacron sail will.


Which is less expensive: a new mainsail or a new halyard?


For a racer who budgets for breaking gear to maximise her wins, she already has an exotic laminate sail matched to Class 2 halyard and (possibly) sheet. And those are a good fit for her, with her deep purse. But not for me. I sail to deliver higher quality life for myself and family. And my funds are limited.

[Seaworthy Lass]

Originally Posted by Alan Mighty

Yes, Marlow has no cover length between the Reference Point R and the Taper/Crossover (T in Samson, C in International Standard Name). There are many published splice techniques based on the Marlow splice. Rather than put that dimension as 0 in my spreadsheet (because it gives me a headache to do so) I arbitrarily decided to focus on the 'canonical' splices that have a non-zero RC dimension and the others (Marlow, two splices from Premium Ropes).
……
I might reconsider my arbitrary decision about that RC dimension. Comparability is always a good thing, even if it involves a headache.

I think it would be good to include the Marlow type methods (the New England video is similar). Having an RC of zero is just part of the spectrum of RC possible. This is not a different technique. Not specifying core length to bury the cover is not important, as long as their instructions of allowing enough to bury the core are followed. The Samson type methods just specify an excess amount to make sure you do allow enough.

Looking at more videos, I have the impression 2 types of methods were developed originally by or for:
Samson (who made the first double braid polyester rope) - they specify the beginning of the cover tail in the core.
Selma (who made the first “Swedish type” fids) - they specify the end of the cover tail in the core.

All the instructions I have seen are simply variations of the above two methods.

In addition, they ALL use the same basic principle: they just pull out the core, create an end to end splice with the crossing somewhere at or below the base of the eye, then they milk all this in, BUT lengths of bury are quite different for each method, as is tapering, if any.

Originally Posted by Alan Mighty

* note that Premium Ropes even has a video of the Selma Fid splice, so Premium seems not wedded to any one splice technique and has made videos for several different splice techniques. Whether that confirms the suspicion that they've not tested the static or cyclical strength of the different splice techniques is a good question.

I can’t find a Premium method that doesn’t use the Selma technique. Could you post a link please?

SWL

[Seaworthy Lass]

Originally Posted by Alan Mighty

Thanks s/v Jedi for more of your excellent points.
......
Yes I agree with, we have little room for improvement.

I disagree with Jedi there .

System strength of 90-100% of line strength is only found with new line and constant pull. Cyclical testing has been limited.
Splices (or the line near a splice) do break, and it is just a theory, but I very much doubt it is because the line was subjected to a load equal to its usual line strength had the splice not been there.

Furthermore, do all methods give 90-100% of line strength? It would be good to gain some insight into which is likely to be best.

Originally Posted by Alan Mighty

Brion Toss, I think it was, said something on the lines that when we exert all effort to do the best (make the best splice, use the best appropriate materials), we build in a reserve that allows us to neglect or stop worrying about things.

So part of this thread is about building in that reserve, in this case in splicing

I do, however, fully agree with this .

SWL

PS Alan and Juho, I have now caught up with the discussion here and will go look at the emails with spreadsheets.

[Alan Mighty]

Originally Posted by Seaworthy Lass

I can’t find a Premium method that doesn’t use the Selma technique. Could you post a link please?

See the Youtube video I referenced in post #117 in this thread. It carries the Premium Ropes logo, was uploaded by Premium Ropes in 2012, and the comments in the discussion below it are made by Premium Ropes.


Discriminating between the Marlow technique and the Selma technique from a video is the hard part!

https://www.youtube.com/watch?v=Ym1-rI0SdaA

For what little it is worth, Premium Ropes even has a video of making a halyard knot! Published in 2013. In the discussion, Premium Ropes argues that a splice is superior.

https://www.youtube.com/watch?v=mqzJj6lYtj8

[Seaworthy Lass]

Originally Posted by Alan Mighty

See the Youtube video I referenced in post #117 in this thread. It carries the Premium Ropes logo, was uploaded by Premium Ropes in 2012, and the comments in the discussion below it are made by Premium Ropes.

Discriminating between the Marlow technique and the Selma technique from a video is the hard part!

The Marlow technique is based on the Samson technique.

The Premium technique is based on the Selma technique.
The minor variations are:
End point of bury of cover tail in the core is 1+3/4 fid lengths for Premium and 1+2/3 for Selma.
Premium cut the cover tail obliquely, Selma have a straight cut.
Premium bury the core tail 10 cm below the eye. Selma bury it just 1 inch.

Alan, the brilliant thing about doing this as a joint venture online, is that different skills are brought to the table, advancing work in leaps and bounds .

SWL

[Seaworthy Lass]

Things may be crystallizing for me regarding the length of core cut off, so before we go tabulating all the methods, let me share my thoughts.

The diagram below is universal for all methods I have looked at so far (except Premium’s new video, and I think they simply made a mistake there using a smaller fid for their measurements):










As far as I can see, the only region where the cover and core are not affected a similar way during the bury is RC (where core tail is not buried in the cover) and what we are calling E’L (the region where the core tail is buried below the eye):






The change in length of the cover compared to the core should be roughly just dependent on what the two do in these two regions.
The amount of core cut off should equal the difference between lengths of core and cover tail below the eye, plus the change in length above that actually occurs.

I think a lot of methods add a huge allowance for what occurs below the eye, imagining the cover is altering dramatically in length compared to the core over the entire fid length of cover tail.
In addition, they disrupt the core excessively below the eye before making the splice, then start measurements from there.

So, that is why a frightening amount of core is cut off in many techniques.

--------

In conclusion, I think the point the cover starts to bury in the core should simply just equal the length of the cover tail plus eye size, (ie the spot where the core was extracted from the cover without any initial pulling out of the core) plus RC chosen (RC may be zero). The further allowance needed is simply unknown. It will probably depend on rope manufacturer and maybe rope diameter.

The reason I think we don't need to worry much about the allowance is that core tail is always buried a shorter distance below the base of the eye than the cover in all the methods I have seen. Unless the core tail is left very long below the eye, there is plenty of room to allow for the variation in lengths of core and cover in the bury. The amount of core cut off will therefore simply end up a bit less than the difference in length of the two tails below the eye.

I think all the methods that are cutting off much more than the difference in length between the tail lengths of cover and core below the eye are disrupting the core:cover relationship in the line below the splice and in the process weakening the line in this region.

Any feedback from anyone regarding this?
I could be way off base, so it would be good to have holes poked in what I am saying .

SWL

[Seaworthy Lass]

I am going to cycle down to the closest chandlery in a moment and buy some good quality double braid polyester with a diameter around 10-12 mm and do some experimenting.

I will start with what I have suggested in the previous post (I will not pull and milk the core after extraction from the cover and I will make start point in the core for cover a length of RC from the extraction point).

Then:
- I will tie the knot the shortest possible distance I can get away with (maybe a couple of fid lengths at first).

- I will start with making RC 4 strands from R (= B in Selma and Premium methods).

- The cover and core tails will be tapered.

- I think I will bury the core tail below the cover about half a fid length.

It may be completely impossible to milk the splice doing this. But I will try.
The above is likely to need modifying a bit.

If this works, I will put my money where my mouth is and use this unstitched on our new headsail sheets when the new boat is completed (a soft shackle will connect these to the clew). My gut feeling is that this should be a good eye splice.

Stay tuned .

SWL

[Seaworthy Lass]

If I was inclined to swear, that is precisely what I would be doing at the moment.

This was just sold to me as Marlow "double braid" .
I had quizzed the chandler carefully to make sure it was double braid before I bought it: