Edelrid MegaJul Belay Device

Jon H wrote: Your assertion is the exact opposite of reality and the scientific method. You HAVE to eliminate the human variable in testing. Without a controlled baseline, results are meaningless.

Um, no. That is not how the scientific method works. Human behavior is an important part of the efficacy of things which involve humans.

When you are discussing the inherent properties of the device itself, it is often preferable to eliminate as much measurement error (and I mean that in the statistical sense) not pertaining to the device itself as possible. When you are looking at the efficacy of a device in the field, human interaction is important, and scientifically valid. You are simply asking different questions which require different experimental designs, some of which may strongly require human involvement.

Edit to note: this statement stands on its own, and neither supports nor rejects any of the opinions in this thread.

well the grigri is one of the better tested "locking" devices on the market ..

yet folks get dropped with em all the time

;)

I own one as well Brian. In the 41 years I've been at this, as a gear whore I may have owned most of the devices ever produced (I'd even paid for the last DMM ATC style model Jim designed that was never produced). And I have even worn some of them out (no big feat for the origonal reverso). What Jim and Rich say is right on the money, however I would add that once you factor in "ease of use", for me, that is the nail in the coffin for why the Megajul is a waste of scratch. There is no function the Edelrid has which is not better served by many other devices and that's why it's not worth buying.

Except looks. I love the way it looks. But that's pretty much it. If you like it, good for you. Around here, you'd be in a minority.

Jon H wrote: Your assertion is the exact opposite of reality and the scientific method. You HAVE to eliminate the human variable in testing. Without a controlled baseline, results are meaningless.

Well yes and no.
We know the performance of belay devices in general depends on several factors. the plate itself, the rope, the human input and the fall severity.
The human input is easy, you test loads of people to see how well they can grip a rope, get the pattern and then test to cover either all of them, the mean or the 95% percentile. The DAV use the mean so their baseline is 250N, the CAI the 95% so 160N, the BMC from 50 to 207N and I test up to 300N.
Then design a test method and present the results to cover all the eventualities which is why my test results go all the way from 0 to the reasonable maximum for that particular rope. The user can themselves determine where they fit into the grip profile.

The rope is the bugbear for certification as there are so many different ropes and they change all the time and a certification test requires reproducable results all down the line. For comparison testing it´s easy, just pick a few typical ropes and perform the same test with them on all the devices. You don´t get a finite result but a very accurate and reproducible comparison which is what climbers need.

The fall severity is easy to extract from other information we have from fall testing and theory since it is a purely physical function. You do need a drop test for the more complex devices such as the GriGri and Cinch as their characteristics aren´t linear.

Billcoe wrote:I own one as well Brian. In the 41 years I've been at this, as a gear whore I may have owned most of the devices ever produced (I'd even paid for the last DMM ATC style model Jim designed that was never produced). And I have even worn some of them out (no big feat for the origonal reverso). What Jim and Rich say is right on the money, however I would add that once you factor in "ease of use", for me, that is the nail in the coffin for why the Megajul is a waste of scratch. There is no function the Edelrid has which is not better served by many other devices and that's why it's not worth buying. Except looks. I love the way it looks. But that's pretty much it. If you like it, good for you. Around here, you'd be in a minority.

I've been a gear whore with a bunch of belay devices as well starting with a Sticht plate and even some obscure ones like a Cassin Logic. I recently had a Mammut Smart and Smart Alpine for a while. What I didn't like about them was the ability to feed rope out and they are clunky large. It was too easy to get short roped or short rope a leader. I have heard people complain about feeding rope with the MegaJul. I have not experienced that at all. I find it as easy to feed rope as an ATC. The only issue I have with it is using it in guide mode it is much harder to pull rope through it than an ATC Guide especially with thick diameter ropes.

"The only issue I have with it is using it in guide mode it is much harder to pull rope through it than an ATC Guide especially with thick diameter ropes. "

Yup, that's typically the same for me. Unless I'm planning on getting in a lot of lead pitches, I don't have a thin rope. When I have a thin rope I want the lightest device I can find so I haven't gotten the correct combo of rope+biner down that would optimize the Megajul. Plus the rapping has been heinous. :-)

Take care

Patrick Shyvers wrote: You just have to watch a few people mis-using the GriGri to understand what he means. For example, a popular way to pay out slack is to pinch the cam with your thumb. However, when your climber falls, the natural instinct is to tense- which leads to clamping your hand down on the cam, and preventing it from locking!

I agree that it is easy to misuse a Gri Gri but I would suggest that misuse is more difficult with a Mega Jul. Even if a fall occurs while rope is being fed with a hand pulling up in the yellow loop, the device still locks. I'm sure there's a way for a belayer to screw up but I reckon it takes more work.

Well the guys at Climbing Technology were keen to have their product put through the wringer and in fact gave me a couple to play with, they are like that as a company.
Early days on the testing as I only got home yesterday and had to remove my test rig this morning to use the overhead crane to turn a boat over.
Ignore the unlabelled curve above the MegaJul one.

Roughly follows the same general pattern, the more assist at the start the less increase with hand force. It´s a bit weak if only one strand is loaded but otherwise powerful enough.
In dynamic mode (usefully this is provided so we can direct see how effective the device is without the locking assist) it´s too weak with this diameter of rope and virtually every conventional plate would be better.
Just got to cut half the stuff off, change the rest and it might go on the cliff with me!
And yes, it´s a fairly steep learning curve with about a dozen modes I didn´t want.

Shouldn't you have been testing the Micro Jul with that skinny rope instead of the Mega Jul? That is the extreme low end of rope diameter for the Mega Jul and the mid range for the Micro Jul. The Micro Jul is clearly more suitable for that rope.

Edelrid specs:

Mega Jul -- Suitable for 7.8 - 10.5mm diameter ropes

Micro Jul -- Suitable for 6.9 - 8.5 mm double and twin ropes

Out of curiosity, how are you generating those smooth curves? Why not show the raw data instead of obscuring it with smoothed lines?

Brian wrote:Shouldn't you have been testing the Micro Jul with that skinny rope instead of the Mega Jul? That is the extreme low end of rope diameter for the Mega Jul and the mid range for the Micro Jul. The Micro Jul is clearly more suitable for that rope. Edelrid specs: Mega Jul -- Suitable for 7.8 - 10.5mm diameter ropes Micro Jul -- Suitable for 6.9 - 8.5 mm double and twin ropes

7.8mm is within the specified rope diameters for the MegaJul and the MicroJul won´t take my 9mm ropes.

shoo wrote:Out of curiosity, how are you generating those smooth curves? Why not show the raw data instead of obscuring it with smoothed lines?

My computer generates the curves. Rope testing is notable by it´s jerkiness with all sorts of bumps and hiccups as the rope gains tension, slips and so on. The raw data is anyway enormous even scanning at 50ms, each point used to create the curve is made from 3 pulls creating 3-400 data sets so maybe 1,000 data sets and there will be up to 12 of these depending on how far up I go with the weights. The curve is then created using a cubic spline but I doubt anything would change using another method.

Jim, if a mega jul is set up correctly with an appropriate biner and rope diameter, your graphs just don't make sense. You cannot convince me that when a mega jul is fully locked, it holds less load than an ATC. To me, that's ridiculous.

Syd wrote:Jim, if a mega jul is set up correctly with an appropriate biner and rope diameter, your graphs just don't make sense. You cannot convince me that when a mega jul is fully locked, it holds less load than an ATC. To me, that's ridiculous.

What the graphs say is that with a pair of 7.8mm ropes, which satisfy the diameter limits specified by Edelrid, when a MegaJul is "fully locked" and subjected to a load in excess of about 200 lbf, someone with a grip strength of about 35 lbf will experience rope slippage, whereas the same belayer with the same ropes and the same grip strength will be able to withstand about 700 lbf. before slippage with the ATC XP.

One of the problems with the assisted lockers is the "appropriate biner" proviso. How many biners are you supposed to buy and test out before you've got one that is "appropriate," and given the failure of ordinary field testing to reveal much if anything about plate behavior under higher loads, how would you even know if you had an "appropriate" carabiner?

None of the slip thresholds in the graph are high in terms of more serious falls. 7.8 mm twins will be running through all of these devices according to the graphs, but the 200 lbf threshold is indeed surprisingly low, and if correct indicates the MegaJul is simply mis-rated for 7.8 mm ropes.

As for "appropriate rope diameter," I've been ranting for a while that something like the middle third of the manufacturer's ratings are sensible to use. That surely looks to be the case with the MegaJul.

rgold wrote: What the graphs say is that with a pair of 7.8mm ropes, which satisfy the diameter limits specified by Edelrid, when a MegaJul is "fully locked" and subjected to a load in excess of about 200 lbf, someone with a grip strength of about 35 lbf will experience rope slippage, whereas the same belayer with the same ropes and the same grip strength will be able to withstand about 700 lbf. before slippage with the ATC XP. One of the problems with the assisted lockers is the "appropriate biner" proviso. How many biners are you supposed to buy and test out before you've got one that is "appropriate," and given the failure of ordinary field testing to reveal much if anything about plate behavior under higher loads, how would you even know if you had an "appropriate" carabiner? None of the slip thresholds in the graph are high in terms of more serious falls. 7.8 mm twins will be running through all of these devices according to the graphs, but the 200 lbf threshold is indeed surprisingly low, and if correct indicates the MegaJul is simply mis-rated for 7.8 mm ropes. As for "appropriate rope diameter," I've been ranting for a while that something like the middle third of the manufacturer's ratings are sensible to use. That surely looks to be the case with the MegaJul.

How much rope slippage are we talking about? Will it slip a meter before locking up? Will it slip until the climber hits the deck without locking at all like an ATC that someone lets go of due to rope burn? Will it not slow down the climber due to the "rope squeeze" of the device? Has this been tested? If it slips a bit then locks that is a good thing. If it slows down the climber's fall enough to potentially recover control of the rope that is a good thing. Edelrid does suggests an "appropriate" carabiner (of course they sell it). I agree with you that testing the extreme low end diameter rope is not that useful, because as you state, a lot of devices do not perform as intended at the extreme ends. If you specifically use that skinny diameter rope than you should buy the more appropriate MicroJul which is specially designed to use skinny ropes.

Once the rope starts slipping under tension, how much slippage occurs depends not on the fall factor but on the height of the fall, and of course on how much resisting force the rope is slipping against. So giving a slippage number, even theoretically, requires more data.

I don't know what happens with the assisted locking devices if the rope is already slipping through it (this does appear to be inevitable for big impacts) and then the climber lets go. Neither the MegaJul nor the Smart seem to provide much in the way of hands-off resistance; the only device that appears to do that is the Alpine Up. Until further evidence comes in, I'd guess the assisted lockers will drop a falling leader if they start to slip and the leader lets go. So you still need belay gloves, maybe even more so than with an ATC-XP.

There is this video of the Alpine Up catching a factor 0.8 fall with a 70 kg weight on double 8mm ropes unattended and without much rope weight on the brake side, which would make one think it can withstand a higher almost-zero-grip impact load then the 463 lbf (2 kN) in Jim's graph, so that makes one wonder whether there might be something different about dynamic performance as opposed to slow pulling. But we don't know how many times they tried this and what the percentage of successes was. Maybe it a stop like this is in reality very unlikely...

youtube.com/watch?v=nFn4xzV…

someone needs to go out and do drop tests with real belayers in the system for these assisted locking devices

and publish the results

DAV/CAI ... im looking at you

;)

Syd wrote:Jim, if a mega jul is set up correctly with an appropriate biner and rope diameter, your graphs just don't make sense. You cannot convince me that when a mega jul is fully locked, it holds less load than an ATC. To me, that's ridiculous.

To me, believing something despite being presented with clear and credible evidence to the contrary is "ridiculous."

Jim, back to my previous comment. If you are willing to share, I would be curious to see (and play with) the raw data from these tests. I get the feeling that the results being shown are might be sensitive to the model assumptions inherent in the way you've created your dataset and made your regression lines, or at least that just showing the lines doesn't tell the complete story. I have a pretty thorough statistics background (though not in engineering, for what it's worth). Also, I totally understand if you're not willing to share, so don't feel any obligation whatsoever.

shoo wrote: I get the feeling that the results being shown are might be sensitive to the model assumptions inherent in the way you've created your dataset and made your regression lines, or at least that just showing the lines doesn't tell the complete story.

Jeez, we aren´t building spacecraft here!
I just pull out the relevant bit from the data logger by eye (around the highest value) which will look like this 208,5215606689
216,7815246582
223,4453125
229,6317749023
236,1815032959
241,8124389648
245,1723022461
249,2980651855
252,1116790771
255,6405487061
258,0233764648
261,1440429688
262,6145019531
265,6962890625
266,0627441406
268,7139892578
267,8781433105
267,3043212891
268,7522277832
268,0309143066
266,5695800781
266,5272216797
265,8539428711
268,6813049316
270,4434814453
273,2039489746
274,9021911621
276,3335571289
276,3335571289
277,1043395996
279,8782653809
280,6866455078
282,9700012207
282,7061157227
282,267791748
279,8988647461
279,7368774414
280,4504699707
281,7590637207
281,7858581543
280,3583374023
280,3613586426
280,4682312012
277,8180847168
270,4512023926
272,6824035645
272,8060302734
274,1416625977
275,5802001953
275,0513305664
274,2037963867
272,5997924805
270,4722595215

and either let excel make a curve or pick out the highest value myself which in this case is 282.97kg (the rest of the digits are meaningless in this application since I´m not calibrating that low, on a max load of 500kg the accuracy is around .5kg for my normal work).
I´d round that to the nearest kg then run the test again twice with another section of rope each time and do the same again. Average the three results and use that as on load point on the curve. A few kg variation here or there is normal, used ropes aren´t the most consistant things in this world.

bearbreeder wrote:someone needs to go out and do drop tests with real belayers in the system for these assisted locking devices and publish the results DAV/CAI ... im looking at you ;)

And what´s wrong with YOUR national federation or do you expect us Euro´s do and pay for everything? :-)