Quad anchor failure

I do not know anything more about the accident, nor the airman who fell to his death. But this accident highlights the importance of not extending in serene. I was not there, I am not familiar with the area even less the route. The accident happened when one piece of a two point trad anchor faile. The failure of one nut and the afterwards shockload on the anchor made the cam in the other end of the anchor come out. Link to the article: Link.

My condolences to all affected

Edit, link to the full accident report: ​Accident report ​​​

The intro says it was gear failure, not a failure of the quad. Apparently the nut and cam both pulled from the rock. 

I suppose they were rapelling on dynamic rope ? Then there was no shockloading. Just 2 bad pieces that both came out during a rapell.. meaning they were so utterly shit pieces that they did not even hold bodyweight.

A terrible accident. The accident highlights the importance of placing good gear in good rock, and choosing the appropriate number of pieces. There's a reason why S (solid) and R (redundant) are the first two letters in the acronym.  If a 6" - 24" body weight extension causes your anchor to fail, your primary connection points are the issue at hand.

I have climbed extensively at the Black Cliffs where the accident occured. The rock on the top of the columns is of very poor quality, often large and angular fractures of un-attached and weathered basalt chunks with the density and competence of a surficial basalt lava flow. The cracks are also often filled with sand, soil, and lichen, and therefore prone to shifting and breaking. I could see how someone without great experience could mistakenly believe protection was in solid rock here- it would be fairly easy. Whenever I rappelled in the Black Cliffs from anchors I built on the rim I was always very cautious and always nervous. IMO soil stakes are the better anchor in these environments. It's a tragic accident.

Condolences

Very scary and sad accident. I've been at the top of the Black Cliffs before (and other columnar basalt cliffs like it). It seems like the rock on top of those types of cliffs is always very crumbly and disjointed. I've only had to set a trad anchor a few times in rock like that, and I've never felt that comfortable with the anchors, so I usually end up with at least 4 pieces. If I can find a good horn to sling, I'll count that as a couple pieces.

I feel like the rock on top of volcanic cliffs hasn't been solidified nearly as much as the rock lower down on the cliff. It seems like there's more air mixed in, which leads to it being a lot weaker.

Basalt will get you! Shout out to his brothers that stepped up and immediately attempted to save his life.

"Quad anchoring system used in the mishap with nut still attached (left), however, the cam was not recovered (the cam depicted here (right), which would have been attached via quickdraw, is for reference for the way it was likely setup during the mishap climb)"

It wasn't attached to the quad?

Find it hard to believe the officials would quit looking for the missing cam. Sounds like they failed at due diligence. 

M Appelquist wrote:

What do you mean by “not extending in serene?”

The SERENE principles are Secure, Equalized, Redundant, Efficient, No Extension. There are variations of SERENE (and also ERNEST), but the last two letters usually stand for No Extension, i.e. if a piece of the anchor blows, the master point shouldn't move very far. If it does, it can shock load the remaining pieces and cause them to fail.

From the testing I’ve read about it appears:

If an anchor equalizes (eg a sliding X or quad) it will have some extension. So the acronyms are not possible in reality (It won’t equalize and not extend). An anchor without extension may have some load sharing but likely not full equalization.

An anchor that extends will not shock load the remaining piece enough to cause it to fail unless it is a poor piece to begin with. 

The youtube channel "hownottohighline" has some good videos on this topic.
They convinced me that the need for No Extension is not very important. Shockloading is just not something you should worry about much. Based on that, I agree with the assessment that in this case the gear placement was the likely culprit.

https://www.youtube.com/watch?v=nr3YBDnOI8Q

Dylan Pike wrote: The intro says it was gear failure, not a failure of the quad. Apparently the nut and cam both pulled from the rock. 

The gear pulled due to the anchor system failing.  I wouldn’t rule out the quad configuration playing a role though it seems poor placement/rock quality is most important.  To me the biggest concern is how anchor systems are taught (especially on Youtube) with an overemphasis on rigging systems over placements.  The fact that he placed only two pieces seems to me that he was overly focused on the quad configuration (incorporating more than two is possible but a pain) and perhaps relying on the mystical powers of equalization to make up for questionable pro.  It’s also possible that extension in the system played a role; the rope would prevent shockloading of the climber but I’m skeptical that it would do so for the anchor pieces, which were all attached to static cord.  Perhaps those who know more can comment on it but for me the takeaway is that Quads are not magical and will not make up for poor placements.  Would a pre-equalized anchor with 3 or 4 pieces and no extension have held?  Possibly.  To those who say extension and shock loading don’t play a role: why did the second piece fall AFTER the first extended?  If they were truly “equalized,” shouldn’t they have failed at the same time?

Its kind of amazing that MW3 was apparently not injured after falling 15 feet on rappel!

Ted Pinson wrote:

Would a pre-equalized anchor with 3 or 4 pieces and no extension have held?  Possibly.  To those who say extension and shock loading don’t play a role: why did the second piece fall AFTER the first extended?  If they were truly “equalized,” shouldn’t they have failed at the same time?

One solid placement would have held (one good cam placement may hold about 2,000 lbs) In all likelihood both placements were bad (likely related to poor rock quality). 3 or 4 pieces may have helped because the 3rd or 4th piece may have been a good placement. Or perhaps they were all bodyweight only placements and with enough distribution between all of them they would have held. But that's not the type of anchor I would ever want to trust or have ever had to use.

I don't say extension absolutely didn't play a role, but I do believe the failure was probably mostly or all the result of bad placements.

They could fail in series (not at the same time) if they were equalized and maybe one piece was capable of holding 125 pounds and the other piece was capable of holding 90 pounds. If the guy weighed 175 pounds it would hold at first. but then the piece capable of holding 90 pounds shifts or the rock moves and it releases, then all the weight is on the other piece, which is not capable of holding the entire load on it's own and the other piece fails.

I think many of us are saying that with good placements it doesn't really matter what anchor configuration you use, so primarily focus on good placements.

michael sershen wrote: The youtube channel "hownottohighline" has some good videos on this topic.
They convinced me that the need for No Extension is not very important. Shockloading is just not something you should worry about much. Based on that, I agree with the assessment that in this case the gear placement was the likely culprit.

https://www.youtube.com/watch?v=nr3YBDnOI8Q

So probably worth noting here:

According to the report, they were using Sterling Power Cord. I've only used Dyneema, but my understanding is that it is similarly static and it can similarly lead to a much harder impact force on a piece than nylon if there's extension.

The piece was still probably marginal to pull here, but materials could have also contributed.

Story seems fishy.  How was the cam not still attached to the cord along with the nut?  I call B.S.

michael sershen wrote: The youtube channel "hownottohighline" has some good videos on this topic.
They convinced me that the need for No Extension is not very important. Shockloading is just not something you should worry about much. Based on that, I agree with the assessment that in this case the gear placement was the likely culprit.

https://www.youtube.com/watch?v=nr3YBDnOI8Q

Shockloading is not as big of a deal as people make it sound if dynamic things are in the system.  But if the pieces are bad even a little shockloading is a problem.  Also, even if you had a BFK between these pieces, the direction of pull will change if one comes out (because no it is pulling that piece straight down as opposed to an angle) and that can be a problem as well.  

It highlights nothing other than the need for getting good gear. If you're afraid of 1 foot of extension with a dynamic rope in between you and the anchor then I would ask you how you feel taking a 1 foot fall lead climbing as the scenarios are similar. 

Ben Silver wrote:

So probably worth noting here:

According to the report, they were using Sterling Power Cord. I've only used Dyneema, but my understanding is that it is similarly static and it can similarly lead to a much harder impact force on a piece than nylon if there's extension.

The piece was still probably marginal to pull here, but materials could have also contributed.

If the rappel is on dynamic rope, shock loading will be reduced proportionally to the amount of dynamic line between the load and the anchor. If the rapeller is only a few feet out from the anchor, the dynamic rope will provide very little stretch and the anchor will undergo greater shock. If the rappeller is far down the line, all the rope to the anchor contributes stretch and thus reduces shock load (by increasing the amount of time over which loading occurs, and reducing max loading) during extension of the system. So the risk is higher when the rappeller is closer to the anchor and there is not as much rope to stretch. As someone mentioned above, taking a one foot fall is worse than a longer fall, because at very short lengths the behavior of dynamic rope approaches that of static ropes. I would say that is most noticeable in the 1-3 foot range, and many of us have probably experienced that at some point. 

In this case the rappeller was 38 feet down the rope, which, if dynamic, should have reduced shock loading considerably. However I just looked back at the full report and I didn't see the type of rope distinguished as static or dynamic. Neither of those words are found within the report. 

The use of dynamic software in the anchor wouldn't help much because of the short lengths involved within the anchor. And remember the first piece failed under static load conditions. 

I agree with others that it was likely the gear placements were of poor quality. The tops of basalt flows are notorious  for that. There is increased vesiculation (gas bubbles) in the lava at the tops of flows, as well as the common presence of thinner flows mantling the early thick cliff-forming flows. Both thin bedding and vesiculation permit more rapid weathering of the rock which degrades its suitability for anchors. If the rock was of better quality it would be part of the cliff face. Since it is of lesser quality, it weathers recessively, which means it retreats from exposure by weathering away or breaking up. It is likely an inspection of the gear placement features would have improved understanding of the failures and that these geologic factors contributed.

Regarding the missing cam... the reports says this:

“During the course of administering life-saving measures, a MT member unclipped the MA’s personal anchor system from the quad and the cam from the anchor.”

I’d guess the cam was misplaced after it was removed.

Also, even though the photo of the anchor shows a BD #4 cam, the report says the size is unknown.

The first piece in the anchor failed under maybe twice bodyweight and the tests linked above suggest that the second piece failed at under 3 kN (so under 675 lbf), making this an inadequate anchor by any standard.  The additional load caused by the two-foot anchor extension might have been critical ingredient in the failure of the second piece, but given the low failure levels for that scenario, the extension isn't the point to focus on.  The anchor was a terrible anchor, that's the point.  These comments are retrospective, but the picture of the anchor in the report makes it clear that the anchor was inadequate in principle, consisting of a very small nut and a cam of some sort.  The cam was lost  during the efforts at life-saving.

Regardless of the size of the cam, this anchor would have been a worrisome component of a retreat in a Patagonian storm.  At the top of a 50 foot cliff with all the time in the world to arrange multiply redundant gear in questionable rock, it was a disastrously poor decision with tragic consequences for the rigger.  The situation is made even worse because there were good rap bolts just at the edge, so that one strand of the rap line could have been clipped (with a long sling so that the rapellers wouldn't have to navigate around it), and this precaution would have turned a fatality into something perhaps not trivial but not life-ending either.  Moreover, the anchor shifted just as the second rappeler went over the lip to another ledge.  At this point, a now obviously dangerous anchor should have been reinforced in a hurry, or at least one strand of the rap lines clipped to the unused bolted rap anchor, but instead the rigger "inspected" the anchor and concluded that the effect of the shift was that the pro was now "more securely gripping the rock."  The anchor failed when the first person was near the ground, pulling the rigger off the top of the cliff.

I know these are harsh things to say about someone whose choices were fatal and who is now deceased; I wouldn't normally have commented.  But the fact that someone in charge of the safety of others (not to mention himself) made a a series of egregious rigging errors needs to be noted and processed.  No one in a recreational crag situation should be rapping off a #4 DMM nut and a cam, even in excellent rock, obvious backups were available and not used, and no action was taken when clear evidence of anchor instability surfaced.

A scary thing to note for those retreating in Patagonian storms is that the anchor was tested "five or six times" and then by the successful rappel of the first person down. and failed on the second person when they were near the ground.  Having trad pieces fail after being tested has been, in my experience, the result of an expanding crack.  This can be hard to judge in some cases, in others it is or should be obvious. This reminds us that getting pieces into separate features can be critical.