Building and Using a Knot Testing Rig

Chit Chat
1

I have for some time now been planning a knot destructor – a rig in which I can subject knots to progressive tension in order to observe how the knots deform, and to be able to subject pairs of knots to enough force to destroy them so that I can compare the strength of one with another. Well, I have finally finished planning how, and what with, and have now build an extremely simple and economic rig.

I have opted for Static Comparative Testing. I have an old 5 ton hydraulic jack and an even older adjustable 10 ton axle stand. I have set these up about 8" apart with a length of 1" silver steel bar below them and another length above them. The two test knots are tied in either end of the test cord, the center is draped over the top bar and the bottom bar is fed through the two loops. Slack is taken up and the stand is adjusted until the top bar is horizontal and the cord is under tension.

I then very slowly pump the jack to progressively increase the tension. As the load is increased, I can watch as the knots flow under load. I haven’t set it up yet, but the knots can be photographed so that dimension changes can be followed and loaded deformation can be recorded. This might yeild information why some knots break before others if certain structures repeatedly form in weaker knots. Finally, one of the knots will fail with a thunderous bang.

By comparing a matrix of knots, I hope to be able to build a table of comparative strengths without needing to resort to the expense of utilising a certified calibrated testing facility. The rig is extremely low cost (mine cost nothing), but even if you bought the components new, you should be able to build one for ca ?20-?30, putting it into the range of many members. Then hopefully results can be reproduced by others, leading to a significant table of relative knot strengths and beyond that, examining how knot performance changes in differing materials.

If anyone is interested in creating a rig for themselves, I will post details of its extremely simple construction.

2

Materials:-

5 ton hydraulic bottle jack
10 ton adjustable axle stand (8" square base)
two off 1" dia Silver Steel bars each 14" long
two off 1" self centering bearings
three off 8" lengths of 2" x 3" timber (this length is set by the width of the base of the axle stand)
two off 17" lengths 5/8" x 4" timber
16 off decking screws.

3

Pretty neat; might i also suggest a dyno/pull scale; they can be pricey, but i hung out on ebay long enough to get 2 @ about $50 each. They tell you what the tension is; and also have another needle that holds the highest tension reading after relief of tension/knot breaking.

Other considerations would be how different knots, in different types of rope constructions differ; like 3-strand, double braid, regular braid etc. Ratings of break strength of the line itself can vary; and different batches of line can vary in strength somewhat. So, i’d suggest making your own control break strength from tensionless knot around a large , smooth device on each end; from the same cord length you will eventually cut the knot test pieces from.

Other tests might include measuring the static elongation; stressing it a few times like that, then a break test; as a cycles to failure test. Dynamic tests, and elastic absorption can be done too, and the dyno can sure help there. Static knot tests are good rules of thumb; but i think different constructions of rope, and different knots can give different handling of dynamic forces; then after static elongation too. So, a knot and rope construction that is better statically; might not be the best choice under dynamic loading; especially in a short line. For, line length is not a static force consideration; but certainly is a dynamic consideration; whereby the more rubberband length; gives more dampening qualities.

if using a mechanical advantage pulley pull with dynamic testing; it should be noted that a 3:1 can give more static strength; but reduces dynamic absorption qualities; as elastic deformation is partially dependent on percentage of the tensile loading capacity achieved in dynamic loading. This doesn’t make much difference to the primary line in use as 3:1; but it will make a difference to the connecting devices outside of that system/ to support/ anchors etc.; that might very well be a knot you aren’t focussed on testing(test knot would be on the other end)… So, statically a 3:1 can reduce anchor and connection loading because a 1:1 through a pulley will give double loading to pulley and connections; but a 3:1 will only give 1.3333 loading statically. But dynamically a 3:1 can give more loading than 1:1; especially in short and/or non-elastic lengths. A 90 degree pulley redirect of higher test line; can keep you out of line with the weaker cord/knot being tested when it breaks; and place slightly less than 2:1 on the pulley/anchor system.

Please, also include safety re-search; whereby if you don’t have the system in an isolated box; you have a catch tarp over it; or laced through a tree branch; that would catch line and hardware from flying at breakage; and of curse, don’t be inline with such forces. Also, we (should) now offer a dis-claimer, adult stupervision(witch definitely leaves me out :o); and this advice, as we recommend/invite others to participate in this endeavor…

Don’t let Dan the man catch wind of this; as he’d surely enthusiastically flood ya more than i!

4

Step 1 – make the base.

Do not worry that the timber specified will not be able to take the forces we will be working with. None of those forces get through to the frame, its function is only to hold the load bearing components in place, especially after the knot breaks and releases the forces into the rig. The sizes speck’d were for an axle stand with an 8" square base. If your axle stand base is different, then adjust sizes to suit.

First, take two of the 2" x 3" cross struts and cut a notch across the 2" face, 1" wide and 1" deep. Make the notch in the center of the strut. The first silver steel bar will sit in these notches directly beneath the axle stand and the bottle jack. The bar should be a snug but easy fit to allow the bar to be withdrawn and replaced each time the rig is loaded with a new test piece.

Next mark and drill the side timbers 1" down and 1" in , 7" in and 15" in from the LHS. (clamp the two sides together and drill though both at the same time). Then mark and drill 1.5" up and 1", 7" and 15" in from the LHS. Use a drill that will allow the decking screws to clear through.

Hold the first cross piece (the one without a notch) in place, flush with the LHS end of the face and 1/4 down from the top edge. Fix the first notched cross piece in place centered on the 7" screw holes and again 1/4" down from the top edge. Then fit the second notched cross piece centered at the 15" screw holes and again 1/4" down from the top edge. Turn the base over and screw on the second edge matching the first. It should look roughly like this :-

http://knotbox1.pbwiki.com/f/Base%20sml.jpg

Step 2 – Fit the force gear.

Stand the axle jack in place over the two LHS cross pieces and secure it by screwing it into the cross pieces.

Fit the bottle jack in place over the RHS cross piece, fixing it in place with screws if fixing holes are available. Alternatively, clamp it in place with two pieces of scrap wood, bridged from the front and back sides across to the bottle jack and screwed through to the cross piece, or securely lash it to the cross piece.

Slide the first Silver Steel bar into the slot beneath the jack and into the slot beneath the axle stand. This removable bar is the bottom tension point.

I use two self centering bearings to hold my top tensioning bar, but these are probably unnecessary. If you want to do away with them for simplicity and cost, then lash the second Silver Steel bar to the axle stand and the top of the bottle jack and there you have it – a 20,000 lb test rig (although it is only a 10,000 lb jack, it has a 2:1 advantage because it is forcing against one end of the bar while the cord is suspended in the middle). Finish the job off by getting rid of any splinters and rough edges and give the assembly a quick spray to help make it look half decent and you are ready to start putting knots to some serious tension.

Although the rig could take 20,000 lb I have no intention of going anywhere near there. I will start all of my comparisons with 2mm braid which hardly makes the rig grunt – I am sure that I have a lot to learn even working at these very modest (and relatively safe) tensions.

Step 3 – Safety

SAFETY is of course an ever important issue, both your own and anyone else’s who you might allow to watch you testing knots. If you do not feel competent to take full responsibility for safety, then do not build a test rig – just be content to read the results from others. On the other hand, if you do decide to build a rig and someone gets hurt, don’t consider for one moment that you might slap an action on me for damages, it will make me very upset and you might find bit of anatomy being tested to the full 20,000 lb and then we would both be very upset.

Seriously though – the energy stored in even a 2mm diameter braid is so great it can blind you in an instant. If you do not feel competent to work safely with forces and energy of this magnitude and are not prepared to take full responsibility for your actions, then DO NOT BUILD THIS TEST RIG. It is not a toy – take heed of the figures – 20,000 lbs – that the force of 100 big men – you would not want to be in the presence of even one uncontrolled man, let alone 100.

http://knotbox1.pbwiki.com/f/Rig.jpg

5

Hi TTS,

Thanks for the ideas and the important heads up re Safety.

As you can see from the pics, the rig is super simple and exclusively static / comparative. No force measurement, just - which one broke before the other - and possibly some examination of the debris after the bang. I appreciate that a lot of user relevant information can be gleaned from dynamic testing, but the whole subject is so complex (work to heat conversion, shock wave magnitude and direction etc. etc.), that I don’t think I will be going there just yet. If I manage to build a matrix of A stronger than B stronger than C etc. in just one cord, then I will have achieved a lot, especially if others can validate the results.

Having said that, I think the job ahead is still far from simple.

I have just put my first knots on the rig – Bowline loop vs Bowstring loop and Bowline loop vs Fig. 8 loop. Bowline vs Fig. 8 came as no surprise, the Bowline failed, but in the Bowline vs Bowstring, the Bowstring failed and I really did not expect that so I immediately questioned the uniformity of the cord I am using.

I have realised that even these two results will raise questions not just from myself but also of course from others not present at the test. So perhaps before I do any more testing it is time to consider what controls need to be in place for a recorded test and what information should be recorded over and above ‘which knot broke’.

Starting the list then with the test cord itself.

  1. What should be used and how do we ensure uniformity if a number of people set up rigs with the intention of verifying their results. One ideal would be for the Guild to buy a large quantity of a single source of cord (say 2mm braid) and members could then call off from this stock for testing to an agreed protocol. The alternative is that we each source our our test cord and forsake the opportunity to achieve the all important peer review and validation.

  2. What cord materials and constructions should be included? We already know that for example cotton will fail in a different manner to nylon braid, cotton being extremely sensitive to compression. If we were to have say three reference cords for testing - what would they be and why?

Next comes the section on what information should we collect.

  1. How was the knot tied and dressed?
  2. Should the knot be pre-tensioned and to what degree?
  3. Should the knot be photographed to confirm dressing?
  4. Should test knots be tied and independently verified prior to testing?
  5. How else should the test pieces be standardised?
  6. Should the knot be photographed under high load prior to failure?
  7. How / where should the results be recorded? Should they be compiled by the Guild?
  8. Should any attempt be made to record the nature and point of the failure?
  9. If it is intended to attempt to identify structures prone to failure or resistant to failure - what would need to be recorded?

As resolving the cord standardisation issue is likely to take some time (as the Guild Council would be involved) perhaps we could share views on the information collection problems (opportunities).

On reflection, it might be some time before an IGKT consensus might be reached (and I think it would be valuable for this project to be formally ratified, supported and sponsored by the IGKT). Consequently, I believe it could be of value to start running some test samples in oder to uncover as many potential issues as possible. If anyone has some samples they would like me to run on the rig, PM me and I will give you my postal address so you may send me the test samples you want destructing.

6

If anyone does want to send me samples, you will need to know the dimensions of the closed rig.

If you take two small pieces of ca 1" dia dowel and set these 11.75" apart at their centers, you will mimic the two bars in the ready to load position.

Tie one loop test knot to pass easily over the first dowel, pass the cord around the second dowel and then tie the second test loop around the first dowel so there is no slack.

Of course, if you want to compare two bends, tie two cords together with one bend, pass the cord around the two dowels and tie the second bend opposite the first to give a continuous loop which fits snugly over the dowels.

Then mail me the samples and I will wreck them for you.

Derek

7

Here is a question to put onto the back burner for later consideration.

“Does the length of the cord outside of the knot influence the knot strength or mode of failure in cored braid?”

The reason for posing this question is that in some of the early tests, the fractured end exhibits different lengths of core and braid. this might indicate that the core and braid expand at different rates, so in a longer test piece, far more tension may be built up in one over the other, while in short test pieces, there would be far less chance of this differential making a marked impact.

No answer yet but something to consider as understanding grows.

8

i would say in a single line test; especially metered/measured-no.

But, in a side by side test; a rope and or knot construction(of unlike knots or lines) that could deform more or have more line in it’s construction to pull slack from to relieve it’s tensions; could appear to be stronger than one that couldn’t slacken or deform; even giving or exaggerating to the point of ‘errant’ data. Also, cored lines could have internal frictions and slides etc. to the same end; reacting differently to their outer lacings etc.

Please check your messages; and don’t tell Dan how much fun we are halfing.

9

Progress.

I have decided to seal all the cord ends before tying the test knots in order to prevent movement of the core relative to the braid. I am doing this by applying a single spot of Superglue to the braid at the point to be cut, then cutting through the mid point of the glued zone.

I have been making the test loops 4" long in order to keep the spreading angle of the legs to a low value. This has been using 16" of cord to make the two loops. In total I have been using ca 3ft of cord per test. Not a great cost but I do hate wasting cord.

In the stronger knots (Fig. 8 and Lyndy) the cord has been exhibiting a lot of stretch (more than the 5.3" extension of the jack)

I have overcome this to a certain extent by hanging a large 'bina off the top bar. Because it is a smaller diameter it has allowed me to reduce the loops to 2" yet maintain the same spreading angle and overall reduce the amount of cord used to ca 1.5ft

A spin off from this move is – half as much cord - half as much stretch.

I will do a bit more ‘ad hoc’ testing, then propose a protocol which hopefully will attract a bit more response.

10

Since you are setting up testing w/o calibration of force but only checking some presumed
ordering of knots by relative strength, what sort of uniformity is it that you want to ensure?
–uniformity of results almost pre-supposes the point of testing!
How are you to ensure uniformity of knot?!! That is more to the point, and what one should
hope for from broad participation in such a venture then would be a picture of how various
knots behave in various materials.
Now, were one to devise some calibrated testing where forces were measured,
one would want to ensure that there was uniformity of measurement. This is a better
testing than of A-v-B; I’ve seen A-v-B results where AvB is 3:1, BvC is 3:0, and if wondering
about AvC you’d of course expect 3:0 but in fact it was 3:1. Were one to have seen some
measurements of break force, this seemingly inconsistent outcome might be understood
as not so surprising (e.g., were all forces close, or widely varied).

2. What cord materials and constructions should be included?
In what does one care about strength? (And what cordage might be more uniformly constructed than others--giving better assurance that it was the knot and not some cord deformity that made the result?)
Next comes the section on what information should we collect.
  1. How was the knot tied and dressed?

Absent knowing this, what is the point? (But this is never known.)
2. Should the knot be pre-tensioned and to what degree?
For ropes, the Cordage Institute IIRC has options of 20% or 50% rated tensile strength, which seems a little too nominal because of the basis of that nominal, vendor-provided value. But for the relevance of the testing to practical circumstances, I guess one can think of BOTH cases of need: a rockclimber's tie-in loopknot is slack until severely tensioned in a fall, never taking what pre-tensioning would do to it, really. (Although maybe some climbers take "sport falls" repeatedly sans adjusting their tie-in, and there it would be similar.)
3. Should the knot be photographed to confirm dressing?
Seems best, if possible. --once at the setting, maybe again at some near-rupture point (where "near-" can be a point where change is pretty much gone, the shape is set and tension just builds)
4. Should test knots be tied and independently verified prior to testing?
I guess you mean should a tester perhaps tie off a bunch of specimens, one per knot, and then first get a photo-review check that indeed the tying is done as expected, before running of repetitions of the test. Well, yeah, that would be pretty good. We might see that this being done by one person, with the photos available for review, would guide others (it would move to a matter of their judging their own work to this visible group-think agreement, as opposed to their judging work w/o such reference at all, or getting their own group-review).
6. Should the knot be photographed under high load prior to failure?
Yes, that is good information. One might see some materials allowing greater deformation from common appearances at setting time to rupture than others (nylon more than polyester, say).
8. Should any attempt be made to record the nature and point of the failure?
Certainly! One might see if premarking cord w/hash marks (and some way to identify each) will work, coupled with the photograph under tension to put marks at whatever positions. Another and less risky marking would be bits of thread sewn into the cord.

–dl*

11

“A journey of a thousand miles starts with the first step”

The journey to find the grail of – understanding how knots work and how they focus failure in their substrates – has been started many times. Reviewing the records of those expeditions shows us some of the complexity of that world and also tends to show that none of the expeditions seem to take us any closer or pave a way towards that grail with perhaps one small exception captured by Dan in this quote:-

I have to totally agree with Dan’s criticism. Trials to date have shown that we must start with consensus on the EXACT form of the knots being subjected to testing. Without this agreement and understanding we will be starting from the same point as all the others and will have learnt nothing from them.

I am thinking that this level of consensus can only be achieved by sharing physical samples of test knots between members of a volunteer Peer Group. If a test is to run say eight replicates, then perhaps duplicate test pieces should be sent to each of four peer auditors who will verify that the knots have been tied to the provided diagram and dressed and set to their expectations. If any of the reviewers do not agree that the knots have been tied correctly, then the volunteer peer group will review the disagreement and arrive at a consensus before the test proceeds. Notes of the review process will need to be kept with the test results. Yes, this will slow the test process, but members of the Peer Group will all have confidence in exactly what knot was actually tested.

Of course, this all depends on the availability of enough volunteers to make up the Peer Group.

Anyone interested?

12

Dan,

Issues which cropped up repeatedly in your reply were the fact that the calibrated measurement of the force needed to rupture the knot would be much more valuable and that dynamic testing would be far more relevant to the real world.

I do not dispute this, and were I to win the Lottery, I would be straight out and purchase an Instron Tensiometer with full NPL traceable standards and maybe even fund accreditation for the testing procedure. However, I haven’t won the Lottery and am limited to a test rig created from equipment and materials already on hand. Consequently, although much inferior to calibrated dynamic force measurement, I am limited to comparative static testing.

I do believe though that with an appropriate Protocol and adequate Peer Review, some valuable work can be performed which will at least allow us to rank knots in terms of deformation under extreme stress and their impact in reducing the cord breaking strength. We might even, with a little ingenuity, be able to cast some light upon the impact of specific knotting structures on performance.

Ingenuity and collaboration will be fundamental in any success we might achieve.

To this end your caution re marking the cord, and your alternative of using a sewn in thread is going to be a fundamental. Once the knots are dressed and set, we could pass a tell thread through the load strand immediately outside the knot and another 1cm away for calibration and yet another a further 1cm away to confirm that the tells were not weakening the cord. Similar tells incorporated into the loop strands will allow us to measure how much cord was drawn out of the knot during tensioning. In addition, a photograph just before rupture will show how much the load line tell has moved away from the knot and can be used to indicate where inside the knot failure reached.

"2. What cord materials and constructions should be included?"

In what does one care about strength? (And what cordage might be more uniformly
constructed than others–giving better assurance that it was the knot and not some cord
deformity that made the result?)

I did not mean to infer selecting cord for strength attributes. By materials and constructions I was alluding to the fact that say laid cotton cord might break with totally different characteristics than say a nylon braid. I have a little experience that points to cotton thread being particularly sensitive to compression and so might make compressive knots very weak. By careful selection of materials and constructions we might be able to verify or reject this perception.

With that in mind, what would you choose as your test cords?

13

Here is an outline for members to start with.

Nothing precious here, feel free to rip it to bits where you think it is weak or lacking and rebuild it to something workable that as many as possible are happy with.

Derek

Comparative Knot Strength Assessment Protocol

Equipment

The IGKT Forum Comparative Test Rig or any similar micro incremental rig able to exert up to 10,000lb force onto a test sample comprising two knots under test, the weaker of which should rupture first.
A knot tying jig.

Materials

IGKT supplied reference spools of test cord, or cord of known and verifiable provenance.

Records

Every test shall be recorded with a unique identifier comprising a unique tester ID followed by a unique and sequential integer.
Information for each test will include the following:-

Unique identifier
Date of test
Cord batch identification (IGKT cord ref)
Knot A identification and dressing details (including a schematic or photograph if appropriate, length of cord to tie knot and 80% rupture extension)
Knot B identification and dressing details (including a schematic or photograph if appropriate, length of cord to tie knot and 80% rupture extension)
The name of the Peer Reviewer who ratified the knots had been tied correctly and any comments from them or the review team
A photograph of the knots under load, prior to rupture to guage cord extension and extent of cord extraction from the knot
The name of the knot which failed
The distance from the tag to the break and the length of the break
The length of the break on the broken end.
The length of cord extraction both sides of the surviving knot

Samples/Evidence

The test samples and fragments will be appended to its record sheet

Sample Preparation

Select a reference sample spool of cord.
If the reel is new, measure 1m from the end and spot the cord with one drop of superglue. When set, cut through the center of the glue zone and discard the end 1m.
Tie one loopknot in the cord to give a loop 2" long and an end at least 2" long.
Remove any twists in the cord within the knot and in the loop.
Dress and set the loopknot in accordance with the agreed structure.
Place the loop over the tying jig and pass the line around the second peg and back to the 4" mark.
Add on the length of cord needed to make the second loopknot and mark the cord with a pen.
Place one drop of superglue onto the marked spot and when set, cut through the glue zone.
Tie the second loopknot then adjust the knot on the jig to remove slack from the connection cord.
Dress and set the second loopknot in accordance with the agreed structure.
Make a total of eight replicate test samples.
Make out a record card for every test piece and loop the sample through a punch hole on the card.
Send duplicate samples to each of four Peer Reviewers.
On return of the samples, confirm the Reviewer has authorised the knots and schedule the samples for testing.

Method

Enter the Unique Identifier into the test log and record date of test.
Load the knot to approx 80% of rupture extension achieved when testing the weakest knot against itself.
Release the load and unmount the sample for marking.
Mark the cord 1cm from the load line entry to the knot and 1cm from the loop entry to the knot.
Using a fine rounded needle, pass a cotton ‘tell’ line through the test cord at the marked points and 1cm further from them.
Reload the test piece and tension to ca80% rupture extension.
Photograph the tensioned knots.
Continue application of tension until one knot ruptures
Measure and record extension at rupture.
Remove the test pieces and photograph the ends of the ruptured knot
Measure and record the distance from the tell to the start of the rupture and the length of the rupture for both ends.
Record any observations re the nature of the failure.
Staple the completed test remnants to the test record.
Submit the results to the Peer Review Team
Enter results into the Testing Database.
Update the Relative Strength Table.

14

Hmmm, I mentioned the nature of loading once, and not to claim that dynamic loading
was (everywhere) more important. --reading w/agenda, methinks! Rather, I simply
said to consider what loadings most mattered in various applications, and try to model
that. But it might be argued that if we believe we have methods that are improvements
upon others, or otherwise care to compare our results with others’, than we should at
least make some similar testing for that sake (which we might do, e.g., to try to figure
out how two sources presumably using standard slow-pull testing arrived as such
different figures–we’d replicate the test method, but pay close attention to the possible
different knot geometries that might’ve been key factors).

In some practical cases, the significant aspect of “strength” might result from the way
the knot is battered about, where a not-terribly-high-strength but well protected knot
such as the Fisherman’s knot ultimately proves stronger than some other. But this would
be difficult testing; I just cite it as something to bear in mind in estimating relevance of
any testing to knot selection.

On your test rig, Derek, what concerns me is its elongation range–looks pretty short,
and so would be limited to relatively low-stretching materials (and knots that don’t feed
out a lot of material on compression).
And I’m still trying to be clever re getting some suggestion of calibration–crude, even–
into a simple tensioner. argh

–dl*

15

Yes Dan your concern is founded. Both the jack and the stand have a working extension of 5.5" I did not think that would be an issue but have been surprised by both the elasticity of the 2mm braid I have been experimenting with, but mostly with the feed out from some knots. The Fig 8 loop vs the Lyndy loop both fed so much cord out that I had reached the 5.5" before either had failed.

I overcame this by letting down the jacks and taking a turn around the top bar, this took a little over 3" out of the test piece length, then continued to load the piece. This time the feed out had ended and the fig 8 broke after only a few inches loading.

The test rig is slow and far from ideal, but with a little ingenuity I hope it will yield some useful information.

Derek

16

Excellent presentation & invitation–a start & directions.
I’m taking the bitwise reassembling approach, with or sans ripping noises.

[b]rationales (for the method) / goals:

G1) Focus on things that can be done rather simply, with variously available tools, methods, & materials.

G2) Seek uniformity (or best approximation of it) or at least clarity in what is being done
–specification of #1’s items, and of the tested knot.

G3) See this initial effort as just that–a firSTep, towards better understanding, and as part of maybe
an iterative, collective process that builds upon itself.

G4) Describe some extant testing / information in a critique that highlights why some new, better
test & reporting methods are needed–which should help cast our goals more clearly.
[/b]

----------------------- Remarks -------------------------------------------

R01. We introduced G4 in a separate thread, which has to date (2007-10-01) gone nowhere.
It’s a reasonable thing to do, and can yet be done.

R02. I’d not first seek uniform testing devices, but ask for INTEREST in the process by broadcast
announcement, with solicitation of devices that others might have available to them. Granted, one
stands to get various results by various means; but this in itself can be helpful, and in any case
useage will deliver various loading to knots. I think that there is much to be gained even without
such replicating testing as pure uniformity of methods would provide; deliberate replication esp.
of peculiar results should be possible as an iterative step, and possibly made efficient with some
sort of sampling (vs. fuller testing)–if sampling confirms … .

R03. “IGKT supplied reference spools of test cord, or cord of known and verifiable provenance.”
Here is an aspect I think is too bound to the goal of uniformity/replication. And as it’s a costly one
in terms of materials & administration, I’m happy to suspend it on 1st pass. By common sourcing
(e.g., buying Sta-Set from Samson (or is it Yale? !)), there might well be some uniformity. But folks
might have access to lots of used/discarded cordage which could be put to some use. We stand
to have those results questioned re effects from cordage, but still getting easily tested results I
think will help–cf. R04.

R04. We stand to gain a LOT by simply doing a MUCH BETTER job of recording where various
knots break, and by testing some greater variety of knots than is commonly done–a pretty easy
thing to do, given the limited nature of testing for most ROPE knots (anglers knotting is maybe
more diverse).

R05. “Every test shall be recorded with a unique identifier comprising a unique tester ID followed
by a unique and sequential integer.” Okay, we can have fun putting this together, and thereby
killing some potential confusion of reference.

R06. Hmmm, that list of data to record seems excessive, enough to put off potential contributions;
I’ll not discuss it further in this post, for time. (Again, consider R04.)

R07. “The test samples and fragments will be appended to its record sheet” --or tagged w/right
identifier themselves.

R08. “Select a reference sample spool of cord. If the reel is new, measure 1m from the end and spot the cord with one drop of superglue.
When set, cut through the center of the glue zone and discard the end 1m.” WHOA: IGKTers don’t discard new cord!! :o :smiley:

R09. The bit of careful sample preparation is sure to kill the idea, and esp. the mailing to & back
from some reviewers–Holy bureacracy, BatMan! Absent such keen observation, we, yes, might be
at some doubt as to what…, but in the interest of lifting the test balloon off of the ground, some
baggage must be left. Testers need some common guidance–an image, and some corresponding
words–for forming the specimen. It will help if they can photo-record their results, and then we have
the ruptured item (which might not be very helpful–i.e., might be quite ruptured).

R10. “Using a fine rounded needle, pass a cotton ‘tell’ line through the test cord”
I think that “cotton” is gratuitous; brightly colored (and maybe white is fine–matter of contrast) thread
is more to the point and spacing–which might depend upon dia. of test material (1cm in 3/8" rope
makes a relatively finer gradation than 5mm in 2mm cord). I think that the material can be threaded
PRE-tying (obviously, knowing where the tying will be), and thereby come throughout the knot;
to this, perhaps a separate–relative-knot-position marker (at entry, at end, e.g.) threads of different
hue (black, say) can be added. Photos will help show where things are. And a variety of colors
can better deliniate respectively marked points, though requires more fussing in marking (one
could simply knot/non-knot thread ends to distinguish order, if that even is helpful to do (maybe not)).

R11. “Reload the test piece and tension to ca80% rupture extension.”
Why? Why 80%, and how would one know what this (sorta “2 B Tested”!) is?
Moreover, how relevant to use are results obtained after this preparation,
as opposed to something more akin to reaLife™? Working rope should be
loaded no more than SWL(WLL) values; climbers will be loading knots that
variously are fallen upon or merely manually set. One risks following a standard
procedure that while uniform is also irrelevant, and potentially yields dubious information.
(Some reaches hither & yon re this, though, might show more linear changes in
values and suggest that perhaps easier & uniform methods do NOT so differ–don’t know, this, now.)

R12. “Photograph the tensioned knots.” A tester might get a good idea of when (tension-wise)
to do this, after a test or two, at least per knot. I suspect there gets to be a compression after which
not so much change occurs, at least esp. not to geometry–but maybe we see differences re this per knot.

R13. There might be some value in having some tests constucted such that the surviving
knot can be re-tested, against a newly tied one in the same cord (easiest (least use of material)
for a hitch), or against a calibrated tester. (Might we see e.g. old knot weaker than new?)

Okay, enough for now.

–dl*

17

Dan,

I was beginning to think that perhaps there was no interest in this project but your detailed response shows how much thought you have put into giving the project a strong foundation. You are also looking beyond this small beginning at the considerably larger picture and the need to paint this trial into that larger perspective. To address this need, perhaps we should put together a leading document of Rationale, Goals and Perspective, particularly emphasising that this pilot project is very much a learning exercise or as you put it a first step, towards better understanding, and as part of maybe an iterative, collective process that builds upon itself.

Would you perhaps be prepared to put such a brief together for the project?

The second thing I have drawn from your analysis and comments is that we will probably need a number of different highly detailed Testing Protocols, each designed to address a particular aspect of the bigger picture.

One lesson I learnt during my many years as an analyst is that humans are sloppy, bad communicators and excruciatingly prone to influence results to match their expectations, and that as a consequence if you want to even remotely approach Fact instead of Figment, then the testing method must be rigorously and pedantically defined and preferably should require the input of numerous people – one person tie the knots, one person verify the knots, one person destruct the knots and one person record/interpret the result. Striving for this approach in our situation has a number of advantages over just seeking meaningful results – in our case it also means the opportunity to share the workload and it gives members the chance to see and authenticate the knots for themselves in the way that a picture (even a good quality one) or a diagram never could. Of course, I doubt that we will ever get a team of four together, but even a team of two people will do a far more reliable job than one.

Whilst you are keeping in mind the bigger picture, I am proposing to take a very narrow (almost one dimensional) trip into essentially uncharted ‘KnotSpace’. The more we are able to nail variables then the more reliable and reproducible the results will be. By contrast, the more variables we allow to creep in, then the more blurred that line can become. Consider if we were lucky enough to have three people testing, yet none ever saw the others knots (i.e. no peer review of knots before testing). If the results in some way do not agree, then we could be fairly sure that it was a knot tying difference and go back to confirm if they were tied the same. If however, we also allowed a different cord to be used and now the results were different, then what would you look at to find out why they varied -knot effect - cord effect or combination of both?

Once you have one thin line reliably and reproducibly nailed in place, then you can start to vary other parameters and compare the differences. Different tying structures - just how much of an impact do they have? Different cord - does it effect some types of knot but not others? You can only start to make these comparisons once you have a reliable and reproducible baseline in place.

The variability referred to by Roo in a previous post on work done by Murlie is a testament to the need for this rigour – was that variability due to tying one knot to the next? was it due to dressing? was it due to twist? was it due to rope variation? was it due to variations in draw out? was it due to slippage? – so many questions and many more that cannot now be answered because we do not have access to any of the details necessary to interpret the cause of variation and we were not able to interface with the experiment while it was in progress nor now reproduce it in order to be able to ask new questions.

While a rigorous approach will put off some people from contributing, it will mean that the results we obtain will be far more lasting and valuable than something produced by a ‘crowd doing their own things’ without organisation or uniformity. We have a number of examples of the latter and drawing any conclusions from them has not so far proved possible, so a Guild based rigorous approach is going to be valuable is I would argue is long overdue.

18

Heh, I’m interested in this project.

However, as I’ve stated before, my interest is limited to what I call ‘life support knots’.

The only type of rope I’m interested in is the modern EN 1891 low stretch and EN 892 dynamic rope (also commonly referred to as Kernmantel rope). Its a 2 part core and sheath construction first introduced to Earthlings by Edelrid in Germany. Go here; http://www.edelrid.de/en/kat/uid_kategorien/0000125/id_matchcode/company/bop/0/chksum/55f6d1aa06e1aa533c69fc1c14fdd3b8/beetools.html for info on Edelrid and KernmantEL ropes (love that spelling!). Yes, I have a narrow field of interest… sorry.

Where are you at with this project DerekSmith?

Your test rig is innovative but as pointed out, has limited travel and scope.

I have mucked about with high lift jacks (as used by 4WD, SUV enthusiasts) and Tirfor winches… they give plenty of travel and pulling grunt.

I also have a portable (brick size) load cell - a ‘Dynafor’ 5 ton digital model). Runs on 4 x AA size batteries.

I often use this load cell and high lift jack combo to test permanent anchors in remote areas (eg at cliff sites) - where gear has to be carried in by foot.

Of the Kernmantel ropes, I have to add my two bobs worth by stating that Bluewater (USA) and Edelrid (Germany) are the most common ropes you’ll come across. Bluewater low stretch (ie static) ropes are widely used by rescue teams, defence forces and other outdoor recreation groups (in abseiling, sorry rappelling applications). Due to its widespread use, it makes sense to test Bluewater ropes.

11.0mm diameter rope is most common but, some personnel use half inch/13mm ropes (eg USA rescue teams and Fire rescue teams).

I suspect that due to my narrow field of interest (ie knots used in life supporting applications and kernmantel ropes), it may not be compatible with your objectives…

agent smith

19

AgentSmith, if you have such nice gear, how about doing some nice testing
with it for the climbing/etc. community by testing a low-stretch item in current
debate: HMPE slings & the so-called, VARIOUSLY FORMED, “Girth Hitch”?
Do you ever pull old slings from sport sites or trad rap stations?
See what e.g. Kolin Powik has done (cf. www.bdel.com/scene/beta/qc_kp.php#current
(and maybe you need to browse this a little, esp. for some other interesting testing)),
but recognize that, quite surprisingly (!), he ignores what I’d think is the most common
geometry of this broadly (mis-)named joint, which is asymmetric!!?

The obvious prospective solution? --reeve the tapes together SEVERAL times
(like making an extended Surgeon’s-like knot vice the minimal SquaREef knot),
which IMO should both increase strength AND render the joint easily untiable.

Derek, your points about necessary doubt in light of sub-standard testing are pretty
much agreed–it is part of our beef w/extant testing, afterall–, but that doesn’t much relieve
the distress to be felt by would-be testers; and, again, my point R04 about what can be
realized even without your rigor stands as partial promise of improved if not perfect testing.
A compromise might come if to some degree the different testers attempt to do SOME
replicative testing of others’, apart from what might be more natural to them; this would
amount to a sort of sampling, and if results were within an expected range we might
not be so remorseful about full replication, and if they differ, well, that would point to
an area needing scrutiny. On the flip side, the problem with your “one-dimensional”
testing is that it is so nearly perfectly narrow as to then raise the challenge of
relevance to much of anything else? And the cost to do more “else” is great, given
all the overhead you load on this testing of multiple peer review, etc..
The results might not be such as to give concerns (but, OTOH, this lack of concern
on getting expected results might itself be a blinding to mistakes!).

In short, I want SOMEthing, of some value, to be done, and to be done better
than the current state of the practice–which doesn’t take much, methinks.
Remember, a perfect line lacks thickness, and in some sense that’s
analgous to lacking meaning, practically. It’s a balancing act, to be sure.

Oh, and one should also ask what is to be gained (esp. compared with what
might be lost–opportunity cost) by gainsaying testing that doesn’t meet the
specified method? If someone has a BigTruck to bust real rope, but cannot
match your small-cord (necessary for small stroke of jack) method, who gains
by excluding her truck-busted results? Maybe we find more than one method
that can be done by two or more testers. (Hmmm, picture surging a trawler
with maybe a half-dozen 3/8" cordage specimens set in staggered lengths
to break in sequence, small weight bags nicely pulling exploded parts down
to water harmlessly.)

The variability referred to by Roo in a previous post on work done by Murlie is a testament to the need for this rigour -- was that variability due to tying one knot to the next? was it due to dressing? was it due to twist? was it due to rope variation? was it due to variations in draw out? was it due to slippage?
Given what I know of it, I'd say "probably not" as answering you in one swell foop! --though knot variability is possible, albeit by one tyer. Rather, it is likely in large part to the harder-to-nail-down vagarities of knotted cordage at its limit, and I think this is Roo's point re A-v-B testing versus calibrated testing per some means. E.g., if one looks at the actual data for the Lyon testing of some kernmantle ropes in various knots, you'll see different results for the same knot, and can match them (mentally) with those of other knots to get different sets of A-v-B results. (If each of one knot's results are paired IN ORDER with the other's, it might *win* 3-0; but if the best is paired with the other's worst, ... 1:2 (100v95, 90v85, 80v75 or then 100v75, 90v95, 80v85).

[quote[While a rigorous approach will put off some people from contributing, it will mean that the results we obtain will be
far more lasting and valuable than something produced by a ‘crowd doing their own things’ without organisation or uniformity.
[/quote]
Really?
In ReaLife™, are knots tied by committee w/such rigorous examination, etc.,
or by random individuals in various ways & materials & settings? I often remark
as folks taking e.g. some reported results for the fabled Fig.8 loopknot as Gospel
Truth when they haven’t even the notion of variability in what is denoted by that
knot name (to the tester vs. themselves). Well, yes, we would be mindful to report
that just THIS so-shaped structure in THIS material loaded in THIS manner led
to this result; but could be venture much further in even comparative assessment
vis-a-vis some pratical application?

I’m not advocating blindness on knot geometry in place of rigor–some hope for
getting knowledge of test specimens, but photo of actuals or reasonable assurance
that an exemplar was replicated–perhaps because we have presented the given
“knot” in several possible/common geometries and requested testing to focus on
just one or two, or to any but please identify which–i.e., that we have some belief
that differences are set in daylight and should be recognized.

Nor am I suggesting "a crowd doing their own things without organization … ".
(But if one does see variability, it should at least raise awareness of that aspect,
which is a step above the oft’-seen recitement of some results as Fact!))

–dl*