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

[Seaworthy Lass]

Carrying the theory of line weakening further, I am thinking about what the cover is likely to be doing if you have cut off too much core.

During the splice you have milked with all your strength to try and bury the splice. I think all the excess cover (resulting from too much core cut off) will probably be jammed up pretty close to the splice.

So, what happens if the excess is not redistributed further down the line? Will the weakest point of the line then be somewhere close to the splice? Failure modes often given us a clue as to what may be happening.

In the splice tests (including car tests, Jedi ):
- Was the breaking strain of the spliced line compared to that of identical unspliced line
- When the line broke, did it repeatedly break within a short distance from the splice? This to me would indicate the line was not breaking at its normal breaking strength, otherwise it should break randomly along its length.

The amount of core cut off could well be affecting line integrity. This is not just fanciful thinking.

Any comments anyone?
It helps advance things if ideas are bounced back and forth.

SWL

[s/v Jedi]

Here is something interesting: in the Gleistein (German rope manufacturer) splicing instructions, at step 10 of the double braid eye splice they write:

"
The splice should look like the one shown in the diagram. We recommend a simple whipping of the eye at the crossover point. Should a hollow spot appear at the throat because the core tail was cut off too close, this is merely aesthetically irritating, but does not reduce the strength.
The reason for this is that in the eye of the rope, the tension load is distributed over both sides. Each side therefore has to withstand 50 % of the total load, a feat that the cover alone is capable of handling
"

So we can just strip out the core and make a Flemish eye

Oh, here's the PDF: http://www.gleistein.com/assets/down...ookenglweb.pdf

[Juho]

I did some more drawing to capture the diameter that I forgot, and to add point K and a similar style version of your drawing in case you need it. I used Norwegian letter Ø as the diameter symbol since it matched the current style better than the mathematical symbol ∅.

[Seaworthy Lass]

Originally Posted by Juho

I did some more drawing to capture the diameter that I forgot, and to add point K and a similar style version of your drawing in case you need it. I used Norwegian letter Ø as the diameter symbol since it matched the current style better than the mathematical symbol ∅.

That looks excellent!
I don't need the diagram for anything, but Alan will probably attach it to his table, so it is good having something well drawn.

I am just seeing what columns would be useful in Alan's table, so I was looking at your post #66 to see if we concur.
I have a few comments about that post, but nothing that affects the diagrams. I will reply shortly.

I think the notations you have come up with in conjunction with the two diagrams make analysing the techniques a lot simpler. Good work!

SWL

[Seaworthy Lass]

Hi Juho

I have gone back to read this earlier post of yours to make sure we were on the same wavelength regarding what columns would be useful in Alan's table.

I have found a discrepancy with how we regard the marks on the core.
As I understand it, the marks on the core such as L' (and therefore R') do not necessarily (and I think are unlikely to) line up with their counterparts on the cover when the splice is completed. I have responded in blue below as it would have taken too long to split it properly.

SWL

Originally Posted by Juho

Few more comments coming. I'll concentrate on making the splices understandable.

The spreadsheet approach of Alan Mighty is good. It makes comparison of different methods and picking up your favourite method (or building your own) easy. Having tapering and stitching as separate topics improves the structure.

The Milne and McLaren cross-section drawing is good since it makes it easy to understand what the final product is intended to be like.

To me it fees natural to present also the names of different points in that drawing. During the process I renamed the points in order to understand better the logic of the splice (see drawing below). Those names are not necessarily the best ones. I just want to point out the symmetry of names and points in the cover and core, and demonstrate that descriptive names / abbreviations of the points may be useful.

L = start of the Loop
R = Return from the loop
C = Crossover of cover and core
E = End of rope
X = eXit point

In the drawing points L, L', R and R' are always at the same place in the final splice.
If L' is the mark on the cover before it is removed from the splice, then for most of the methods that I have viewed, it is very difficult to imagine how it ends up there in the completed splice. The amount of core being cut off does not usually roughly match the difference between the length of bury of the tails of the core and cover in the portion below the loop.
It won't be exactly the same due to the core and cover shrinking and expanding as they bury within each other, but it should not be the large amount it is in some of these videos. This places L' in an unknown position in the final splice.

R' is actually marked at the junction after being milked into position there, but it may shift further into the loop as the splice is finally milked.
So I don't think L' and R' will line up with L and R in the completed splice, even before additional load has been applied.


Points C and C' should be placed some small distance to the left, usually (but not necessarily) right next to each others.
In all the methods I have looked at, they aim to make C and C' as close to each other as possible, including the Premium method. The junction is secured (eg temporarily by tape or permanently with stitches). We don't know, other than by dissecting a finished splice, whether they stay close together, or seperate (or the impact of this occurring) although we can speculate according to how much extra core is cut off than "expected".


Points E and E' should be far enough to the left to generate sufficient friction (or base for stitching). Often E and E' are intentionally not at the same place. Points X and X' should be little bit to the left of E and E' respectively.

Based on this analysis the key differences in markings and structure of the final splice are length of the loop (L↔︎R), length of R↔︎E, length of R'↔︎E', length of R↔︎C, and length of C↔︎C'. I guess these values are sufficient to describe the proportions of all possible splices of this type (excluding tapering and stitching/whipping options).
I agree all these should be included in the table, except CC', which is unknown.


I read the spreadsheets in this light.
Row Taper/Crossover means the same as R↔︎C.
No. I think it is the same as EC, not RC.


Row Reference is equal to R↔︎E.
Agreed


Row core eXtract is L↔︎R + R↔︎E.
Agreed


This row is not really needed if Reference (R↔︎E) is already given, length of eye can be whatever, and the formula is always "R + eye".
Just an observation here. The length of the eye is not technique dependent, but I think this is actually important when the ultimate strength of the splice is being examined. This is part of the total bury of the core in the cover. The length of buries in the various techniques have hopefully been designed to cope a wide range of eye sizes.


Row Z is L↔︎X.
Agreed

Row I refers to L'. This row has no value since the position of L' is determined by practical means (e.g. next to L after milking the rope). If this is always the case, this row may not be needed (or maybe for mentioning if rope was properly milked or not).
Agreed


Row II is L'↔︎C'.
Agreed


In the previous paragraph I used C↔︎C' to describe the position of C'. When making the marks in the rope, R'↔︎C' is needed.
You've lost me here . I will have to think about what you are saying.


Value C↔︎C' is not intended for making the splice but for evaluating the resulting splice
Yes, but we have no way of measuring CC' unless we dissect the splice, so there is no point having it included.


Row III is C'↔︎X'.
Agreed


The spreadsheet thus covers all the key values that I described above (and thus all possible variants of the splice). My R'↔︎E' can be derived from row III / C'↔︎X' (R'↔︎E' is equal or slightly shorter than L'↔︎C' + C'↔︎X').
I don't think R'E' has anything to do with L'X'.

The exit points (X, X') are needed when making the splice, but they can be forgotten when evaluating the final product.
X' can be forgotten, as long as it was long enough to leave enough room for bury. The position of X is critical though, as it determines how much tail core is buried below the loop (the core tail is cut off at X).

I note that the values (lengths) may be slightly different when making the splice, and when evaluating the final product (because of changes in length due to different radius' in different parts of the rope).

In a loop the load is divided in two parts. Therefore it may be ok to cut half of the threads away, as for example from the core in the Samson video of mail #1.
Agreed

There are at least two kind of stitching. In the Samson style stitches cover a considerable length of the rope. In the Sailrite style stitches are at one spot but from many directions. If the stitches are expected to carry part of the load too, they should be somewhat loose at the start, or at least flexible so that the thread can move so that the load will be shared evenly by multiple stitches when under load. The Samson style may make better use of the whole length of the rope ends (good milking needed).
Just my gut feeling, but I think the stitches should be primarily there to stop the splice shaking loose with flogging and other cyclical loads. The splice should hold on its own merits otherwise. For this reason I personally like Samson's style.

[Juho]

Originally Posted by Seaworthy Lass

Hi Juho

I have gone back to read this earlier post of yours to make sure we were on the same wavelength regarding what columns would be useful in Alan's table.

I have found a discrepancy with how we regard the marks on the core.
As I understand it, the marks on the core such as L' (and therefore R') do not necessarily (and I think are unlikely to) line up with their counterparts on the cover when the splice is completed. I have responded in blue below as it would have taken too long to split it properly.

SWL

Some responses to your comments.

> If L' is the mark on the cover before it is removed from the splice, then for most of the methods that I have viewed, it is very difficult to imagine how it ends up there in the completed splice.

I'd like to define L' as "next to L after the core has been extracted and the rope milked" (and hopefully next to L also in the final product). I refer to the Samson video that did it this way. If one does not do good milking, the mark might move in some direction, but that would probably not be intentional but just a mistake.

> R' is actually marked at the junction after being milked into position there, but it may shift further into the loop as the splice is finally milked.

Also here I refer to the Samson video. They marked R' very late in the process (through a small opening in the cover) in order to start tapering from that point. I think it was very close to its final position. I have not yet seen other uses for R' (that would require marking it earlier in an uncertain place).

> In all the methods I have looked at, they aim to make C and C' as close to each other as possible, including the Premium method.

I'm uncertain if Premium intended C and C' to stay next to each others or move apart at final milking, stretching and loading (I addressed this problem in mail #36, comment 2). I have also not seen any obvious sign of someone planning to intentionally separate C and C' further than required (for technical reasons).

> although we can speculate according to how much extra core is cut off than "expected"

Yes, I tried to watch for such indications in some videos but couldn't find anything conclusive.

> I agree all these should be included in the table, except CC', which is unknown.

I used CC' since in the drawing that describes the final product since CC' is a final product oriented measure. For making the splice one should use e.g. distance between C' and L' (good for marking the position of C'). The table describes ways to make the splice. Therefore I agree that CC' is not useful there. One can e.g. just explain how C' is determined in each method.

> I think it is the same as EC, not RC.

In the Samson video they mark C starting from R, and (one version of) the spreadsheet has value "R - 8 strands". Although the base of measurement originally comes (was started from) E, the final step in the Samson video and spreadsheet seem to refer to RC.

> I don't think R'E' has anything to do with L'X'.

Yes, it seems that I mixed parameters of cover and core. R'E' can be derived from LX. I should have talked about RE that is equal or slightly shorter than L'C' + C'X'.

> X' can be forgotten, as long as it was long enough to leave enough room for bury. The position of X is critical though, as it determines how much tail core is buried below the loop (the core tail is cut off at X).

Yes. One can also refer directly to E' (that was renamed S' in the latest versions of the drawing). Milking, stretching and loading could cause E' (S') to move away from the vicinity of X.

> Just my gut feeling, but I think the stitches should be primarily there to stop the splice shaking loose with flogging and other cyclical loads. The splice should hold on its own merits otherwise. For this reason I personally like Samson's style.

Samson made quite a number of stitches. That could mean that they prepare for having considerable load on the stitches, not only light flogging. Or maybe they are just safe rather than sorry since thread is cheap . I guess stitching could be done also to add strength, if one wants to avoid bulky splice at all cost. But that is probably unusual.

[Juho]

I missed two.

> You've lost me here . I will have to think about what you are saying.

The drawing shows the final outcome. The alternative approach is to study the methods that lead to the outcome.

The values that I used are thus related to the final outcome, like how much of cover is buried in the splice (RE, or RS in the latest version) (is there enough friction?), how much core in buried in the splice (R'E', or R'S' in the latest version) (is there enough friction?), or did the crossover point get loose (CC').

In the methods we are more interested in questions like "how can I find the spot where the cover should be buried", which leads to using parameters that can be marked and used one by one when making the splice. The spreadsheet describes different methods, so it definitely sees the world from this second point of view.

> Yes, but we have no way of measuring CC' unless we dissect the splice, so there is no point having it included.

Yes. It is hard to measure CC', and it is not very useful as a parameter for some particular splicing method. It just tries to describe the end result. It is "practical" in the sense that you can say that most methods probably aim at making CC' as small as possible (I'm still unsure about Premium).

Among the characteristic values that I listed also RC is quite trivial since in most splicing methods I expect that value to be "short fid" or some other trivial small value.

[Seaworthy Lass]

Originally Posted by Juho

Some responses to your comments.

Thanks. It helps if we are all on the same page .
The only issue outstanding is how L' is defined (when that is decided, R' will follow automatically).

Originally Posted by Juho

> If L' is the mark on the cover before it is removed from the splice, then for most of the methods that I have viewed, it is very difficult to imagine how it ends up there in the completed splice.

I'd like to define L' as "next to L after the core has been extracted and the rope milked" (and hopefully next to L also in the final product). I refer to the Samson video that did it this way. If one does not do good milking, the mark might move in some direction, but that would probably not be intentional but just a mistake.

I understand. But then I think we need to designate another point on the core that is the spot marked on the core before it is extracted from the cover at R. I did this initially and called this P for "pull". We can go back to nominating a seperate point like this if you like.

The reason I think we need this is because L' and P (or whatever we call it) are very unlikely to coincide, even when mistakes are not made.
This is due to the following:
- The cover below R is undisturbed when L' is marked.
- The cover below has bunched up when the splice is completed.
- The core in this section will shorten too, as it passes over the tail of the cover (and this portion of the cover may be elongated as it is empty, so more core than expected may be needed to cover it.), but it may not be by the same amount that the cover shortens.
Also:
- The point where the core is extracted is not consistent between methods (sometimes extra core is pulled put first)
- The distance between this point and C' has some kind of "fudge factor" applied that varies between methods.
- Lots of methods cut off so much core that it must be being pulled from lower down, shifting the marked point up from R.
Finally:
- C' is marked relative to the point on the core where it is extracted, so we do need to know where it is.

I think I have exhausted the arguments .

I understand It would be useful in the discussion to have L' denote where the core lies next to R in the final splice, but for all the reasons above, I think we should then have a seperate point on that denotes the mark made on the core before it was pulled out at R.

Originally Posted by Juho

> I think it is the same as EC, not RC.

In the Samson video they mark C starting from R, and (one version of) the spreadsheet has value "R - 8 strands". Although the base of measurement originally comes (was started from) E, the final step in the Samson video and spreadsheet seem to refer to RC.

I don't have a big preference which way the values are presented (Alan presented that column as EC though). Given we know ER, as long as either EC or RC are noted, the other can be worked out.
The only value of actually presenting EC is that this is the amount the cover tail is buried in the core if no slippage of the core occurs and this is interesting information.
This is a minor issue.

SWL

[Juho]

Originally Posted by Seaworthy Lass

I understand. But then I think we need to designate another point on the core that is the spot marked on the core before it is extracted from the cover at R. I did this initially and called this P for "pull". We can go back to nominating a seperate point like this if you like.

The reason I think we need this is because L' and P (or whatever we call it) are very unlikely to coincide, even when mistakes are not made.

Yes, I agree. I made some shortcuts there, and mixed terminology of the final product and the intermediate step of pulling the core out.

This is due to the following:
- The cover below R is undisturbed when L' is marked.

The cover below (towards the long end of the rope) of L is undisturbed, right?

- The cover below has bunched up when the splice is completed.
- The core in this section will shorten too, as it passes over the tail of the cover (and this portion of the cover may be elongated as it is empty, so more core than expected may be needed to cover it.), but it may not be by the same amount that the cover shortens.

Yes. The cover is expected to shorten somewhat more than the core.

Also:
- The point where the core is extracted is not consistent between methods (sometimes extra core is pulled put first)

For exact results, I guess the rope should be milked well and marked (P'?) after the core has been extracted. That's the only solid reference point available that can guarantee that the rest of the markings on the core will be at the correct places. If that mark is not in the correct place, we can expect either cover or core (and don't know which one) to pull in when the splice is loaded.

- The distance between this point and C' has some kind of "fudge factor" applied that varies between methods.
- Lots of methods cut off so much core that it must be being pulled from lower down, shifting the marked point up from R.

down = towards the long end
up = towards the loop end

If you take the milking knot (K) as our fixed point, then P' probably moves slightly down, and that is (mostly) because part of the cover was inserted inside the core. But at the same time also L (and R as a result) is likely to move a bit more down (because both core and cover itself were inserted inside the outmost core). P' is thus likely to move up if measured from point R. And P' is likely to be little bit to the right of L' (of the final splice), as you drew it in mail #72 (except that R' should read L' there).

I guess the "fudge factor" means that we need a method and rope specific correction to the distance from P' to C'. Usually the intention is to bring C' as close to C as possible. The loop should be constructed so that R is exactly next to L. Measurement from R to C is quite accurate (since there will be one core (not tapered) inside that segment of cover, and RC was maybe marked when there was still a core inside (I tried to address this in my mail #74)). P'C' should thus be RC + CC' + L'P'. L'P' is the required additional correction. CC' depends on the diameter and other properties of the rope.

In the Premium video they seemed have something like half fid for CC' + L'P', which may be on the high side, if they plan to keep CC' as short as it was initially made in the video. In the Premium video P'C' = fid, RS = fid (when measured, finally a bit longer), RC ≈ half fid, and that leaves about half fid for CC' + L'P'. If taht value is too high, CC' may grow, and the end of the cover may come partially out of the core when the splice is milked and stretched.

Finally:
- C' is marked relative to the point on the core where it is extracted, so we do need to know where it is.

Approximately at (marked) P'C' distance from P', where P' is at a method and rope determined constant to the right of L'

I think I have exhausted the arguments .

I understand It would be useful in the discussion to have L' denote where the core lies next to R in the final splice, but for all the reasons above, I think we should then have a seperate point on that denotes the mark made on the core before it was pulled out at R.

Yes, I agree. If we want to formalise the treatment of rope and method related deviations from the ideal model, including P' in the model is a good approach.

[Alan Mighty]

See the latest version of work in progress 'Class 1 double braid eye splice.xls' below.

Still working on it, obviously.


Use the Tabs to move to different worksheets.


Tab 1: Standard Marks
- has a fairly final table giving International Standard Names of marks etc with definitions
- have not added P and P' yet
- Juho's diagrams
- the very start of a table of International Standard Names of dimensions (we need both dimensions used to setup the splice (first stage) and calculated dimensions (after the splice) that can be used to compare splices and then can be used in reverse engineering to determine what the first stage dimensions should be to deliver the calculated dimensions desired - that's why we are using a spreadsheet, not a simple table)


Tab 2: Concordance
- just a bare start at a look-up table, with ISNs (International Standard Names) equated to the marks used in published splice techniques


Tab 3: Canonical
- my name for the published splice techniques that use most or all of the marks used by Samson Rope and Brion Toss
- yet to have all the ISNs added, yet to reformat the measurements so they can be used for dimension calculation
- yet to have a table that does the calculations of what I call those calculated or derived dimensions


Tab 4: Other
- at the moment a garbage can for published splice techniques that do not use all the marks used by Samson Rope and Brion Toss (in other words, splice techniques that I currently think cut a corner or abbreviate a step and make me think that they are not preserving balance between cover and core).


For SWL, Juho, Jedi and any other CF member who want to contribute to the spreadsheet - I think for the sake of CF we ought move the working versions of the spreadsheet off the Forum. For several reasons (the working drafts will be confusing in the future - although SWL could use her moderator privileges to tidy up and remove working drafts; uploading files must be some load to CF; CF forces us to use *.xls rather than *.xlsx; now that I've added Juho's quality graphics, the file size is bigger). If you accept this idea, send me a PM and we can exchange email addresses.


If we exchange working drafts of the spreadsheet, we are going to need a naming convention, such as appending a timedate group and an author bi- or tri-nominal to the file (AM, SWL etc).


I've some other obligations for now (as always), hence the working draft now.

[Seaworthy Lass]

Juho and Alan, I am off to Zwolle shortly, so I'll respond to the last two posts later.

Alan, I am happy to exchange email addresses.

SWL

[Alan Mighty]

Originally Posted by Juho

CC' depends on the diameter and other properties of the rope.

In the Premium video they seemed have something like half fid for CC' + L'P', which may be on the high side, if they plan to keep CC' as short as it was initially made in the video. In the Premium video P'C' = fid, RS = fid (when measured, finally a bit longer), RC ≈ half fid, and that leaves about half fid for CC' + L'P'. If taht value is too high, CC' may grow, and the end of the cover may come partially out of the core when the splice is milked and stretched.

My guesses agree with these two points:


1. CC' may depend on diameter and the nature of the weave of the cover (eg Touwfabriek Langman's Newport double braid has a dense 32 strand cover and a 12 strand core, versus other double braid with 20 or fewer strands in the cover and 6 or fewer strands in the core); and


2. CC' may creep longer or may need to creep longer as a way of getting cover and core back into balance. Brion Toss (and a few others, including Bob Norson) advocated stitching the crossover. When I think about the likely function of the L'C' dimension (I-II in Samson speak), I note that Brion Toss makes it L'C' 1 Ø diameter longer than Samson. I like to think that Brion may have dissected a splice and decided he needed that extra length for cover/core balance.


For now, the dimension is out of reach in splices without a stitched crossover unless (a) a post-mortem dissection allows measurement; or (b) a trick, such as sewing one metal thread or a dab of a metal-rich paint into the cover at C and another into the core at C', followed by an X-ray or by employing a metal detector etc.

[Seaworthy Lass]

Hi Juho
I have nothing significant to raise with your post, but I have attached a few comments:

Originally Posted by Juho

For exact results, I guess the rope should be milked well and marked (P'?) after the core has been extracted. That's the only solid reference point available that can guarantee that the rest of the markings on the core will be at the correct places. If that mark is not in the correct place, we can expect either cover or core (and don't know which one) to pull in when the splice is loaded.

I think P' should be marked through a parting in the cover before it is extracted, as the extracting procedure is going to shift it by a variable amount if care is not taken.

Regarding the initial milking procedure, I have been thinking about that. What do you think the reason is for it? Some methods do this, some don't. Some just pull out the core a little without specifying how much.

I think that the purpose is to align the core and cover in a position you can return it to after you disrupt it during the splicing procedure. Otherwise you wont be able to milk the splice fully.

In some methods I have been able to estimate that the core moves out about 2-3 Ø during this action.
It must be very dependent on how much effort you put into the action. Some seem a bit halfhearted. It may also depend on the type of rope (ie manufacturer) and maybe the diameter of the rope.

I wonder if, with a bit of care, you could not always return to the same cover:core relationship? If so, the initial milking is a bad thing to do.

I need to play with this.

Originally Posted by Juho

I guess the "fudge factor" means that we need a method and rope specific correction to the distance from P' to C'.

Ideally, yes .
What hits me is that all the methods use quite varying lengths for P'C'.
This may be rope manufacturer specific, but it may also be rope diameter specific, and if so that complicates this a hell of lot. It strikes me that it seems to be wildly estimated given it is varying so much between methods

Originally Posted by Juho

Usually the intention is to bring C' as close to C as possible.

Yes, all the methods I have seen try and do this either by temporary taping or by stitching.

Originally Posted by Juho

The loop should be constructed so that R is exactly next to L. Measurement from R to C is quite accurate (since there will be one core (not tapered) inside that segment of cover, and RC was maybe marked when there was still a core inside (I tried to address this in my mail #74)). P'C' should thus be RC + CC' + L'P'. L'P' is the required additional correction. CC' depends on the diameter and other properties of the rope.

Agreed, except I can't see why CC' can't consistently be close to zero initially with a bit of care taken. I have tried a wide variety of diameters and I think the junction can always be jammed close together.

Originally Posted by Juho

In the Premium video they seemed have something like half fid for CC' + L'P', which may be on the high side, if they plan to keep CC' as short as it was initially made in the video. In the Premium video P'C' = fid, RS = fid (when measured, finally a bit longer), RC ≈ half fid, and that leaves about half fid for CC' + L'P'. If taht value is too high, CC' may grow, and the end of the cover may come partially out of the core when the splice is milked and stretched.

The Premium method is a nightmare to analyse, as they give the end position of the cover tail in the core, not the start as other methods do.
I agree though that CC' may grow and if so the amount of cover tail buried will therefore decrease.

[Seaworthy Lass]

Originally Posted by Alan Mighty

See the latest version of work in progress 'Class 1 double braid eye splice.xls' below.

I will look at this in the morning (12 hours from now). Busy day today with the trek to the home town of DSM .

Originally Posted by Alan Mighty

Tab 2: Concordance
- just a bare start at a look-up table, with ISNs (International Standard Names) equated to the marks used in published splice techniques

Where can I find the published standard names for the marks?

Originally Posted by Alan Mighty

For SWL, Juho, Jedi and any other CF member who want to contribute to the spreadsheet - I think for the sake of CF we ought move the working versions of the spreadsheet off the Forum. For several reasons (the working drafts will be confusing in the future - although SWL could use her moderator privileges to tidy up and remove working drafts; uploading files must be some load to CF; CF forces us to use *.xls rather than *.xlsx; now that I've added Juho's quality graphics, the file size is bigger). If you accept this idea, send me a PM and we can exchange email addresses.


If we exchange working drafts of the spreadsheet, we are going to need a naming convention, such as appending a timedate group and an author bi- or tri-nominal to the file (AM, SWL etc)

I agree we should move the working versions of the spreadsheets offline, primarily because I am strugging to attach a version that can be modified using my iPad .

The discussions, however, I think should be kept here. Members may be able to chime in with useful comments. Jedi has already contributed useful posts. The strength of doing this via a forum is the potentially useful group input.

SWL

[Seaworthy Lass]

Before we embark on this, I think we should decide what we are trying to achieve.

Eye splicing is a complex matter with many variables such as:
- Measurements used for the marks
- Procedure (including whether initially the core is pulled and milked, if tapering or stitching is used etc)
- Rope diameter used (eg the same proportion of cover and core thickness may not be maintained, and maybe the best procedure will also vary with rope thickness)
- Rope manufacturer (all double braid poly ropes are not the same)
- Skill of the person making the splice and care taken
etc

The variables are so great that it could takes years of research testing hundreds of splices (with enough trials of each to have statistically significant findings) before we can decide what the "ideal" method is. This is way beyond our scope. So what are we trying to achieve?

This is what I think our initial aim is:
To compare the lengths various splicing methods instruct are used (in units of Ø).

If there are current standard notations for markings on splices, then we should use these. Juho, our work has not gone in vain in that area. I think it has given us both a far better understanding of these marks and I think we also now know if current notation (whatever it is) is adequate for what we want to tabulate and discuss.
Alan, where can we find "standard" notation?


From there, the sky is the limit .
Lots of reading initially needs to be done, as research (good and bad) has probably been done into many aspects of rope strength and splicing. This may take months to do well. I have read very little yet, partly as I am not sure where to access the papers that have been written.

---------

I think the aim of a good splice is twofold:
- For the splice not to come apart before the line has broken
- To retain the line's usual breaking strength

The effect of the splice on the rest of the line is the easiest bit. Someone must have tested how much altering the length of core:cover affects the breaking strain of the rope. I think this is the simplest place to start, as that is about the only thing done to the rest of the line below the splice, apart from the vigorous handling done to milk the splice into place (overkill, in my opinion, but I will discuss this later).

If studies have shown that the length of core:cover needs to be maintained as it was originally to retain strength, this is a very basic thing to aim for. We should try to disturb the line as little as possible below the splice, primarily only cutting off as much core as the excess amount after all the changes in length of cover and core have occurred in the splice itself (these will be mainly in the neck) plus the difference between the length of tail bury for the core and cover.

Shall we try and aim for this to start with, then tackle the harder aspects of splice strength itself if we have the enthusiasm to do so?
Unless something obvious pops out from reading the research that has been done already, I think load testing equipment is needed to achieve much else.

-----

To anyone reading this and wondering why we are trying to reinvent the wheel after six decades of double braided poly use and lots of studies and practical experience, I believe improvement is still possible in this area.

Eye splices do fail. For this reason they are not used in critical situations like climbing, where knots are preferred.

It is possible nothing will come of what we are doing, but I suggest if we read that minimising the disturbance to the main body of the line is an aim, and if we can see how this can be done when using Brion Toss's method, we should put it forward to him. I have not read that this has been critically looked at.
Rope manufacturers and videos come and go. Brion has written a fabulous text on rigging "The Complete Rigger's Apprentice" and the second addition has just come out. This is the text people are turning to, and possibly the splicing technique he describes used most frequently (Sailrite have adopted this too). If we are to achieve something that has lasting benefit, if we come up with anything, shall we see if he is willing to try modifying his method rather than proposing a new method to add to the multitude out there already?

SWL