Does anyone know the strength of an ATC when used in a multi-pitch set up? It occurred to me that the entire weight of a climber is resting on the strength of the ATC clip and not on any carabiner. I know it is not likely, but say if the belayer did not take in slack while the climber was climbing and took a large fall, how much force would be exerted on the ATC?

I've looked at Black Diamond, and Petzl, neither of which have KN ratings on their multi-pitch belay devices. Any clue why?



  • @Sue it's just the Petzl Reverso 4 product page in case it moves again.
    – michael
    Commented Mar 11, 2016 at 22:37

5 Answers 5


Aha! I knew I'd read something about this somewhere, and here it is. Scroll down to the test done by Jim Titt.

First Fail Mode:

The trapped rope escapes sideways from under the tensioned rope and gets trapped between the tensioned rope and the side of the slot.This is very difficult to free off and you have to dismantle everything and twist the locking krab brutally to release the rope. Take your Prusiks.

Second fail mode:

Apply yet more load and the trapped rope where it crosses the tensioned rope goes down through the slot with a bang. At this point the holding power drops off considerably but not catastrophically, though pretty near! Easy to release, just unclip the krab when unweighted. Still need to take your Prusiks!

ATC Guide. 10.2 Mammut, used, non-treated. First fail mode 4.8kN. No second fail mode, rope sheath cut at ca 9kN.

ATC Guide. 9mm Edelrid, used, non treated. First fail mode 2.96kN. Max fail load 5.58kN. Residual load 1.6kN

ATC Guide. 8.2mm Edelrid, new,treated. First fail mode 2.05kN. Max fail load 4.06kN. Residual load 1.2kN

Reverso�. 10.2 Mammut, used, non treated. First fail mode 3.68kN. No second fail mode. Rope sheath cut ca 9kN

Reverso�. 9mm Edelrid, used, non treated. First fail mode 2.25kN. Max fail load 3.60kN. Residual load 0.9kN

Reverso�. 8.2mm Edelrid, new,treated. First fail mode 1.6kN. Max fail load 2.38kN. Residual load 0.7kN All with Petzl Attache 12mm round profile karabiner.

Not my idea of a reliable roped-solo device!


He pull tested an ATC in guide mode. Quick summary: the device itself doesn't fail or suffer damage. With a 10mm rope the ATC jams severely at 4.8 kN, breaks the rope at 9 kN. I don't know how to work out how much of a fall the second would have to take to get to 4.8 kN, but probably quite a bit; s/he certainly ain't getting to nine.

With an 8mm rope it's a little more alarming; device jams at 2 kN, locking fails at 4 kN. I imagine 2 kN would be pretty easy for a dropped follower to achieve, so anybody belaying two followers simultaneously on skinny double ropes evidently needs to pay attention.

  • That's a good point. I hadn't thought of two followers.. Thanks for answering my curiosity. :)
    – michael
    Commented Nov 20, 2013 at 23:42
  • What do you mean by the device "jams"?
    – imsodin
    Commented Mar 11, 2016 at 21:44
  • In Jim's testing he describes two failure modes. In the first, the loaded strand jams alongside the brake strand (you have a jammed mess, but the climber doesn't get dropped). In the second, the loaded strand does not get stuck alongside, but "pops" past the brake strand with a bang (here the braking force goes away). Neither case results in device breakage.
    – requiem
    Commented Mar 11, 2016 at 22:13
  • The link is broken.
    – Qudit
    Commented May 20, 2018 at 18:13

Welcome to outdoors.SE!

It occurred to me that the entire weight of a climber is resting on the strength of the ATC clip and not on any carabiner.

This is not quite right. My first quibble is that in the situation you describe where there was slack in the rope, the load is dynamic, so it's much greater than the weight of the climber. (Sorry, I'm a physics teacher, so I get pedantic sometimes.) The second issue is that it's not true that there is no load on any carabiner. The carabiner that is clipped into the ATC is acted on by a downward force L equal to the load of the falling climber, an upward force A from the anchor, and a small force B (probably upward) from the brake strand. The ATC is designed to make the rope jam while B is still very small. So to a pretty good approximation, Newton's first law applied to the carabiner says that A=L, and the carabiner feels a stress L, i.e., it's fully loaded with the dynamic force of the falling climber.

I think it's true that in this situation, where the second falls and the belayer has left slack, the force transmitted to the anchor can be considerably greater than the force that would be transmitted to the belay station in a normal lead fall. To clarify this, let's consider a normal lead fall.

In a normal lead fall, the rope forms an inverted U around the top piece of protection, and in this U configuration, the friction in the U causes the tension in the belayer's strand to be lower than the tension in the climber's strand by a large factor. The exact factor depends on coefficients of friction, but sources I've seen for typical friction between a steel biner and a climbing rope put it at about 5 to 13. Although this is nice for the belayer because it greatly reduces the tension in the rope at the belay station, it comes at a price, because force is bring transmitted to the top piece of protection, which could pull out.

So in your situation with the second falling on a slack rope, the good news is that he's not loading some crappy, non-redundant nut or cam that was placed somewhere because there was no opportunity to do any better. He's loading the anchor at the top of the pitch, which is presumably redundant and carefully constructed.

I guess you were concerned about the stresses in the block of the ATC itself, and whether they would crack the block open. The thing is, these are forces exerted on the ATC by the rope, and by Newton's third law the ATC is exerting equal forces back on the corresponding parts of the rope. The block of the ATC is made of metal, and the rope is made of nylon. If something is going to fail in this part of the system, I think it's the rope, not the block of the ATC.

  • 1
    "The block of the ATC is made of steel" .... erm, some kind of aluminium alloy. Your general point is right though, failure of the body of the ATC is simply not a concern.
    – AlanL
    Commented Nov 20, 2013 at 12:10
  • Well hello there Mr. Physics Teacher! :) So it sounds like the more likely scenario is that the rope will slide through the ATC can could cause a rope burn/ human failure. I really was just wondering if the companies had load tested the devices on those clips used for guide mode (carabiners all have rating on them so I was wondering why those didn't). The weight is technically passing through the ATC's guide clip on to the other anchors etc. so hopefully they designed them well so that they are not the weakest link in the chain, so to speak. Thanks.
    – michael
    Commented Nov 20, 2013 at 23:52
  • @BenCrowell Do I understand you correctly that by the factor 5 or 13 you mean that pulling 5N or 13N on one side would result in 1N on the other side? I'd be surprised if the efficiency was so low. Do you remember what sources you have these numbers from? I think in rescue/rope access they usually assume that the efficiency of a 180° redirect implemented by a carabiner has an efficiency of 50%, (or more). This means that pulling 1N on one side results in 0.5N (or more) on the other side. See following sources:
    – flawr
    Commented Apr 15, 2016 at 8:01

say if the belayer did not take in slack while the climber was climbing and took a large fall, how much force would be exerted on the ATC?

I assume we're talking about a follower being belayed with ATC in guide/autoblocking mode? Then even the most determinedly incompetent/malicious belayer cannot produce more than a factor one fall. Are you familier with the concept of fall factors? In this case, the unfortunate second climbs all the way up to the stance, without the belayer taking in any rope whatsoever, then falls all the way back down again. The distance s/he falls is equal to the total amount of rope in use: fall factor 1.

Unpleasant for the falling climber, especially if s/he hits something on the way down, but nowehere near the worst case scenario for force on the ATC. That comes with a factor two fall, where the leader falls off above the belay, without any intermediate gear, and falls twice as far as the amount of rope in use.

Now, in a factor two scenario I personally do have serious concerns about the safety of an ATC, not because of the device itself breaking but because of the rather limited braking force it provides on modern slick, skinny ropes. IMO it is seriously questionable whether many belayers would be able to hold a factor two fall on an ATC. The direction of pull is down whereas people are used to holding upwards pulls. Some rope is definitely going to run through, and most people these days don't belay with gloves, so ropeburn is pretty much a given. Etc.

For these reasons I'm seriously looking for an alternative with semi-automatic assisted braking for multipitch. The Alpine Up is big, heavy and looks dauntingly complicated; the Edelrid Mega Jul is nice but I've read about some quality control issues. Will probably buy the second generation Jul when Edelrid have ironed the bugs out.

(Otoh many people argue that the ATC is advantageous with marginal trad gear because its relatively soft braking action reduces the impact force. This is understandable, but based on some tests I've seen it seems to be exaggerated. Top Canadian ice climber Will Gadd for example is on record favouring a grigri for ice climbing these days.)


First off I had this same question Michael, How can a belay device used in "direct" or "Guide" mode not have a load rating. After all would you rappel from a figure 8 with out a load rating on it, I would not? As this situation creates a single point of failure in the climbing system I started to research this..

Here is a web site to calculate fall ratings and load. http://ferforge.tripod.com/Srt002.htm

So with my scenario and weight and a five foot fall while seconding a climb, I can generate 5kn load on the direct belay device, If I'm doing the math correctly. Seems like a bad idea to have a possible 5kn load on a single point of failure. This is calculated in a five foot fall also so your belay would not be having the attention to the rope handling they should. Also what about two climber simultaneously being belayed in "direct" belay mode, double the load?

Sending some emails and some research on the web, this is what I got back.

"Thank you for your email. The UIAA/CE does not necessitate a strength requirement for manual belay devices quite like you’re thinking, hence no strength is advertised or stated. We do know how strong our belay devices are but this information is proprietary and kept internal. I can tell you it somewhat approximates a carabiner in strength."

The UIAA testing requirements for braking devices can be found here. http://www.theuiaa.org/upload_area/Safety/Standards/Safety-Standards/UIAA_129_Braking_Device_March2013.pdf

From a post on a UK forum about DMM Pivot belay device made in Wales.

The DMM "Pivot" guide mode anchor connection point is rated to a minimum of 16kN.

I hope this helps. Regards Simon Marsh DMM International

I for one will always have a backup to my "direct" belays to an anchor system now. For standard "indirect" or from the harness belays I know the carabiner is load tested and I'm very comfortable with this setup.

Regards Chris

  • I am pretty sure you are doing the math wrong. The load is NOT on the belay device, but rather the carabiner the belay device is attached to. The load on the belay device is something like 2x the force you apply with your break hand.
    – StrongBad
    Commented Mar 11, 2016 at 21:33
  • Point up for the research. There is also a report of testing that included the original ATC and GriGri (and many devices) located at caves.org/section/vertical/nh/51/Hang_Em_High_Final.pdf. I suggest looking at the load testing chart of all the devices they tested, and also the failure mode of I believe the Cinch. Note the UIAA standard, for the manual braking devices, states that at the tested loads "The device shall not release the loaded rope, shall not show any visible damage, and shall continue to function properly. The rope shall not break."
    – requiem
    Commented Mar 11, 2016 at 21:38
  • @StrongBad In guide mode/belaying a second you clip the tuber at its link into a carabiner and the rope into another carabiner. So the whole force is on this first and and the carabiner in it. So it is very much relevant.
    – imsodin
    Commented Mar 11, 2016 at 21:56

here ya go the crew at Hownot2 did a bunch of tests on belay devices in a variety of configurations including the actual cables and pulling the eye to destruction basically someone else pulls stuff apart for you so you don't have to.There will be a lot this video that will answer your questions.


Bogdan Cardoş asked Petzl how strong the metal eye was on the atc guide since it is a single point of failure. They replied with "There is no standard requirement to communicate and breaking strength of the attachment point. The product has been designed to be used as described in the technical instruction. It is certified according to the EN standards so we don't have to communicate the breaking strength of the Reverso. Thank you for your TRUST"

Petzl is right, they don't have to tell you and it is probably super good enough, but trust? Trust is overrated, so I took my personal Petzl Reverso 4 and pulled it to destruction. Also, because we had them, we tested the Black Diamond ATC Guide and the smaller Black Diamond ATC Alpine Guide.

Our results: NOTE: The 10.0mm static BD rope used in the following tests breaks in a figure 8 at at 17kn ►PETZL REVERSO 4 (https://amzn.to/3kzl4hW) *Normal belay mode: slips while holding tail at 2kn and 2.5kn *Normal belay mode: breaks rope at 10.64kn with tail at 3.12kn *Double rope "rappel": breaks ropes at 19.59kn with tail at 6.49kn *Wire breaks at 7.62kn *Guide eye breaks at 12.41 ►BLACK DIAMOND ATC GUIDE (https://amzn.to/3ze9D3i) *In guide mode, rope slips at 5.5kn and 6.94kn *In normal belay mode, slips at 2.44kn and breaks at 9.45kn *Wire came out at 6.65kn *Eye broke at 29.07kn ►BLACK DIAMOND ATC ALPINE GUIDE (https://amzn.to/2V0PwH2) *Eye broke at 30.11kn

  • I know that video and it is certainly helpful. However, it would be good if the answer lists the results like outdoors.stackexchange.com/a/4903/17053 does. Text is a lot more concise than video
    – Manziel
    Commented Nov 23, 2023 at 9:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.