ACR Anchor Method?

Care to post a link to the video? I can't find anything I'd consider "recent".

It was from 2012. Nothing new. But if you haven't seen it, here it is:

dmmclimbing.com/knowledge/k…

I've been trying to stay out of conversations about this until Mr. Titt posts his test results. It's been five years now and I'm starting to get old. Anything?

paulraphael wrote:I've been trying to stay out of conversations about this until Mr. Titt posts his test results. It's been five years now and I'm starting to get old. Anything?

Well to save you embaressment about your claims for the ACR I DID send you a review of them and pointed out the failings rather than publicising them. If you want me to post them here then I expect I´ve got them on file.

Mt. Titt, I never received anything from you that I can remember. The issue here isn't my embarrassment. This anchor method is something I submitted to the community for review, and if there are problems with it beyond the obvious limitations, then they should be made public in a form that's as objective and transparent as possible.

I'm glad you've shared your concerns here. But we should be able to review not just your conclusions, but the methodology and the scope of the tests that led to them. It's the only way people can make informed decisions about whether or not use this method, and under what circumstances.

And of course it would be helpful to see how other anchor methods performed under the same test circumstances.

Ah well, no matter. The short version of the tests is a 28 page pdf which I can´t put up here but here´s a very brief comparison of the slidng X/ACR/Alpine Equaliser.
The ACR is effectively the same as the sliding X in function, geometry and performance.

The load split over the three legs with an ACR and a sliding X is 58%/28%/16%, worse still is the Alpine Equaliser which gets 84%/16%/0%.
The Alpine Equaliser design does have the advantage though that when one piece fails there is no remaining twist (around the ring in your case) and it actually equalises as a two point anchor better, somewhere around 1.6:1.
"Connecting two pieces (a slightly tricked-out sliding-x. Works the same way as an x, but the ring creates an obvious powerpoint and greatly decreases friction)"
In your design the cord is wrapped around the ring 450° whereas on a conventional 2 karabiner sliding X the cord is wrapped 270°. The extra wrap angle around the ring makes equalisation even worse than a sliding X, and the ACR gives a load split of 28%/72% compared with a conventional X which gets 35%/65%.

Having the bends round a ring doesn´t seem to achieve anything worthwhile. Replacing a cheap, versatile piece of equipment (karabiner)with a single-purpose object makes little sense to me, since I wouldn´t use either system that´s fairly irrelevant though I guess.
The performance of the ACR is no better than a sliding X as a three-point anchor, worse as a 2 point anchor and the extension uncontrolled if any part fails.

That's helpful. I look forward to seeing the full report, so we can understand the circumstances that generated those numbers.

One point of concern is the assertion that extension is uncontrolled with the ACR. This would only be the case if it's poorly rigged. Did you read the paper on proper use, including extension-limiting knots and extending with slings?

I don't agree with your assessment that the ring has no benefits beyond what a carabiner offers. This has been discussed ad nauseum, here and in the instructions, but the ring has the following advantages:
-no weak axis
-no gate or locking mechanism that can fail if loaded improperly or if subjected to cord running over it
-no sharp edges that can damage a cord running over it
-allows pre-rigging of the twisted loop, to eliminate human error

And as other people have figured out, the ring doesn't weigh much, it doesn't get in the way if you want to use the cord as a sling or regular cordelette, and if you find yourself needing an extra rap ring ... hey there it is.

ACR

Identical to:-

3 point sliding X

Both tested by sliding weight, the anchor is set up on a horizontal bar with weight applied, the anchor initally equalised and then the bar tilted until the weight slides, in effect replicating the effect of an offset load on the anchor and forcing it to equalise. There are plenty of other methods all of which give the same results but this is one of the more convenient.
The computers connected to the strain guages on each point give the graphs above.

It was decided that publishing all the information (it is over 100 pages of obscure stuff) was pointless as half way through I found that the consquences of extension were vastly higher than had previously been thought and that no "equalising" system either gave useful and controllable load sharing or was immune from potentially catastrophic extension (with or without limiter knots). Since "equalised" anchors also gave spectacularly poor load sharing it was considered better to supply information to those best placed to use it and change the emphasis in climbing literature and training from equalisation to redundancy and no extension.

Thank you Mr. Titt for publishing these plots.

I have a few questions:

• what you've done is statically weight a bar with anchors, and then plot the distribution of weight on the 3 anchors, specifically, you are plotting % of total force at each anchor point vs. bar tilt angle.

-Right?

• Why is this representative of how the thing will perform with a fall? The DMM tests, for example, were with a fall, an impulsive force. The BD tests referred to in posts above also showed that knots always drastically reduce runner strength, but they made it look as though they would only fail at much higher force than gear. As the DMM drop tests show, the BD tests which did not simulate a fall (I think these are quasi-static, i.e. turn up the force until the knot fails) *really* missed the point. Falls will cause slings to fail at much, much lower loads.
• What is a 3-point magic X? Sorry, not familiar with it.
• Why didn't you compare to an "eyeball equalized" and tie-off, the obvious competitor?
• ** The DMM video and data clearly show that knots drastically lower runner strength. You don't address this. Perhaps a good question is, "which is worse, shock-loading a piece, or drastically reducing the strength of your runner"?
• If you put an ACR on 3 pieces, and you back it up with your rope on one piece, you have no issues of runner compromised by knotting, and you are redundant.

Any time you knot, you are shooting yourself in the foot. The DMM video/data make that absolutely clear. .. and if it's not a clove hitch, it's going to be hard to untie. I still don't see any clear winner here.

-TPC

thepirate1 wrote:Any time you knot, you are shooting yourself in the foot. The DMM video/data make that absolutely clear. ..

How do you figure that? How much strength are you wanting in you anchor arms and what are you pieces rated at?

How about just using the damn climbing rope if you are so worried.

thepirate1 wrote:Thank you Mr. Titt for publishing these plots. I have a few questions: * what you've done is statically weight a bar with anchors, and then plot the distribution of weight on the 3 anchors, specifically, you are plotting % of total force at each anchor point vs. bar tilt angle. -Right? * Why is this representative of how the thing will perform with a fall? The DMM tests, for example, were with a fall, an impulsive force. The BD tests referred to in posts above also showed that knots always drastically reduce runner strength, but they made it look as though they would only fail at much higher force than gear. As the DMM drop tests show, the BD tests which did not simulate a fall (I think these are quasi-static, i.e. turn up the force until the knot fails) *really* missed the point. Falls will cause slings to fail at much, much lower loads. *What is a 3-point magic X? Sorry, not familiar with it. * Why didn't you compare to an "eyeball equalized" and tie-off, the obvious competitor? *** The DMM video and data clearly show that knots drastically lower runner strength. You don't address this. Perhaps a good question is, "which is worse, shock-loading a piece, or drastically reducing the strength of your runner"? * If you put an ACR on 3 pieces, and you back it up with your rope on one piece, you have no issues of runner compromised by knotting, and you are redundant. Any time you knot, you are shooting yourself in the foot. The DMM video/data make that absolutely clear. .. and if it's not a clove hitch, it's going to be hard to untie. I still don't see any clear winner here. -TPC

That´s basically it, it tells us at what angle of offset force the anchor will start to slide to "equalise" and how well it has distibuted the load once it has.
I and others have done plenty of drop tests, they are useful but not as accurate as slower speed testing to see what the force distribution is.
The strength of slings with or without knots isn´t the object of the tests, only how the total force is distributed over the anchor points. You need to know how the forces are distributed to be able to judge if the knot will fail and there is plenty of other information available on knot strength. There is generally no need to knot slings anyway, I can´t remember the last time I did so.
A 3-point sliding x looks like this:-

3 point x
It is often made with two karabiners at the master point to reduce interference between the strands and possibly friction.
If you consider the one piece clipped with the rope is strong enough to provide a belay then the ACR is superflous, you can belay direct on the one "good" piece and connect the rope to the others as redundancy. This is one method being commonly taught nowadays.
The comparison between dynamically equalising systems and fixed ones (cordalette etc) have been extensively researched, neither have any particular advantage as regards load distribution (they are all equally as poor). As the dynamic ones have potentially dangerous issues with extension/redundancy the the fixed systems are to be preferred.

thepirate1 wrote:*** The DMM video and data clearly show that knots drastically lower runner strength. You don't address this. Perhaps a good question is, "which is worse, shock-loading a piece, or drastically reducing the strength of your runner"?

Please link to your sources so we don't have to go hunting for them.

Generally speaking: extension is considered far worse. Let's examine why:

1) Anecdotal evidence suggests that material failure of the webbing/cord in an anchor is extremely uncommon. I haven't heard of any reports where the material has failed solely due to the load a fall has applied to the anchor. Most anchor failure is from gear pulling. Meaning there isn't an observed problem.

1a) Test data shows the strength of the anchor is good enough. Extra strength on top of what is needed isn't useful.

1b) In a redundant anchor, the unlikely event of material failure will have significantly reduced the magnitude of the impact, reducing the overall force the remaining pieces will see. Think of it like crumple zones in your car.

1c) The DMM data highlights the importance of having dynamic elements in the system to reduce peak load. Mainly connecting the belayer to the anchor with the rope. In their test, they are falling directly onto the sling.

2) IF you are subject to extension, this means one piece of your anchor has already blown. In general, extension will result in the remaining pieces seeing a higher load than they would with no extension. If one piece has already blown, the remaining pieces can now be considered suspect (assuming you judged them all good to begin with), and you'd do well to limit the load they see.

simpler generally = better

Knots weaken the rope? Ok but you still have to tie in. Also, all knots do not weaken equally, bulkier knots use larger angles.

Is there any research that shows that fully loading an anchor one direction, then shifting the direction of load, accurately models what happens in any kind of fall?

This proposition seems counterintuitive. Certainly not a given.

Hey Jim do you have any data on the 2 point quad anchor? It is my understanding that the quad does a good job of distributing the load between 2 pieces, making it the only anchor setup to do so.

If you've got 3 solid pieces, you don't really need to worry much about equalization because you're pieces are solid. No extension is higher on the priority list in this case. If your pieces aren't solid, you need to consider the situation you're in and this risks involved. If you choose to climb on instead bailing, you're gonna want more than 3 pieces and you're probably better off building the anchor with the rope to get the most energy absorption as possible.

paulraphael wrote:Is there any research that shows that fully loading an anchor one direction, then shifting the direction of load, accurately models what happens in any kind of fall? This proposition seems counterintuitive. Certainly not a given.

It may seem counterintuitive to you but it is however accurate and completely logical.
We aren´t stupid enough to spend considerable amounts of time and money without knowing what we are testing so the various methods are compared for accuracy first and also because it is sometimes nescessary to use a different method anyway. The common 4 methods I use give the following results using an identical system (2 point sliding X).

Drop Test
1,69:1 Load Split

Slide Test
1,71:1 Load Split

Tilt Test
1,67:1 Load Split

Offset Load Static
1,70:1 Load Split

With 2-point anchors you can also load them any way you find convenient and draw a vector diagram instead of measuring the forces on each point.

It is also the case that in conditions where strength of anchors is most important there will be a belayer involved and any force on the belay is unlikely to not involve at least some loading from the belayer. None of which alters the underlying physics of friction in the belay system.

eli poss wrote:Hey Jim do you have any data on the 2 point quad anchor? It is my understanding that the quad does a good job of distributing the load between 2 pieces, making it the only anchor setup to do so. If you've got 3 solid pieces, you don't really need to worry much about equalization because you're pieces are solid. No extension is higher on the priority list in this case. If your pieces aren't solid, you need to consider the situation you're in and this risks involved. If you choose to climb on instead bailing, you're gonna want more than 3 pieces and you're probably better off building the anchor with the rope to get the most energy absorption as possible.

The two-point quad is comparatively the best, using 8mm cord I got 1.61:1 load split for the quad and 1.89:1 for a sliding X. Good is relative!
However human beings can make a good shot at equalising a two-point anchor anyway using a cordalette or whatever and can normally get as good as the quad, 3 point ones the tale is different!

the real question... has the crème de la crème been tested?

mountainproject.com/v/11209…

Greg Welsh wrote:the real question... has the crème de la crème been tested? mountainproject.com/v/11209…

Oh man, would'nt it be great if J Marsella's anchor came back with perfect equalization scores. HAHA. Jim, please test this.

Greg Welsh wrote:the real question... has the crème de la crème been tested? mountainproject.com/v/11209…

You don't test an anchor like that; you sacrifice a virgin or large animal to it and beg for its mercy.