|
adatesman
Mar 31, 2008, 10:22 PM
Post #1 of 148
(25753 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
j_ung
Mar 31, 2008, 10:34 PM
Post #2 of 148
(25736 views)
Shortcut
Registered: Nov 21, 2003
Posts: 18690
|
Dude, I just have to say, you and your muh-cheen fuckin' rock. I think this is an awesome addition to the site.
|
|
|
|
|
irregularpanda
Mar 31, 2008, 10:54 PM
Post #3 of 148
(25722 views)
Shortcut
Registered: Mar 13, 2007
Posts: 1364
|
OK
Um, regarding rappelling: a fatty 10.5 tied to an 8 mil or less with an EDK and with a figger8 follow through.
Double bowline attached to a harness.
(so expensive I'm sure it won't happen) an OP link cam placed in a flare and a trango max-cam placed in a flare.
|
|
|
|
|
taydude
Mar 31, 2008, 11:24 PM
Post #4 of 148
(25693 views)
Shortcut
Registered: May 25, 2006
Posts: 531
|
do a pull test on one of those aluminum rap rings to shut up that paranoid guy.
|
|
|
|
|
adatesman
Apr 1, 2008, 12:32 AM
Post #5 of 148
(25668 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
irregularpanda
Apr 1, 2008, 12:42 AM
Post #6 of 148
(25657 views)
Shortcut
Registered: Mar 13, 2007
Posts: 1364
|
adatesman wrote: You never know... we come up with a particularly interesting test someone might pony one up. What kind of flare / load you have in mind? Outward or downward flare? Direction of loading? Binding on any of the lobes? A downward flare could probably be done without much damage to the cam. I suspect an outward would be more likely to damage the cam due to the lowered surface area (side note- I recall seeing a patent about this recently and will see if I can dig it up).
You know, I'm not a huge physics geek, so to be honest, that should probably be up to you. My gut reaction is this: Outward flare. Downward load. The problem with an outward flare is that there could potentially be so many different angle of flare that it becomes a question of "at what point will the OP link cam or the trango max cam break?"
OOOOHHH new thought: an outward flaring placement with a downward direction of pull could be fascinating for standard friends, camalots, and chocks also. Imagine one of those big honking WC rocks placed in the tapered position with perfect surface area. Do you have a hi-speed camera?
So many variables, not enough toys.
|
|
|
|
|
adatesman
Apr 1, 2008, 1:01 AM
Post #7 of 148
(25647 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
moose_droppings
Apr 1, 2008, 2:27 AM
Post #8 of 148
(25592 views)
Shortcut
Registered: Jun 7, 2005
Posts: 3371
|
A cheap simple one, 7mm tied together with a water knot.
If testing has been done on this, point me to it.
Thanks
|
|
|
|
|
marde
Apr 1, 2008, 3:24 PM
Post #9 of 148
(25541 views)
Shortcut
Registered: Sep 3, 2006
Posts: 169
|
irregularpanda wrote: OK Um, regarding rappelling: a fatty 10.5 tied to an 8 mil or less with an EDK and with a figger8 follow through. ...
that might answer your question
http://www.gudelius.de/spst.htm
at least the most important stuff is in english, too.
|
|
|
|
|
acorneau
Apr 1, 2008, 4:10 PM
Post #10 of 148
(25529 views)
Shortcut
Registered: Feb 6, 2008
Posts: 2889
|
WTF... deka-Newtons?!? 
(This post was edited by acorneau on Apr 1, 2008, 4:11 PM)
|
|
|
|
|
acorneau
Apr 1, 2008, 6:09 PM
Post #12 of 148
(25491 views)
Shortcut
Registered: Feb 6, 2008
Posts: 2889
|
Yeah, I know. But when the rest of the world uses kilo-Newtons....
|
|
|
|
|
marde
Apr 1, 2008, 6:20 PM
Post #13 of 148
(25481 views)
Shortcut
Registered: Sep 3, 2006
Posts: 169
|
come on,
kilo, deka are just prefixes used by the si system
why do you use half inch and quarter inch
when everyone is using inches???
I guess you are able to divide or multiply by ten
|
|
|
|
|
tradklime
Apr 1, 2008, 6:26 PM
Post #14 of 148
(25477 views)
Shortcut
Registered: Aug 2, 2002
Posts: 1235
|
I'll definetly send you the cam and the quick links, and anything else I dig up that may be of interest. I have a couple of old SMC rings sitting around I'll send you too.
Regarding the quicklinks, how many do you think would be needed to make the results somewhat representative? 5? 10?
|
|
|
|
|
adatesman
Apr 1, 2008, 6:46 PM
Post #15 of 148
(25469 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
cushman
Apr 2, 2008, 10:26 PM
Post #16 of 148
(25338 views)
Shortcut
Registered: Mar 24, 2008
Posts: 21
|
adatesman wrote: 6. 4/2/08 - Christmas comes early thanks to Mark Cushman! (Sorry, he didn't give his RC username...) All sorts of goodies to destroy
Hey, that's me. I shipped out the package to you today via USPS, you'll probably get it early next week. Have fun breaking the stuff!
|
|
|
|
|
shoo
Apr 3, 2008, 12:12 AM
Post #17 of 148
(25309 views)
Shortcut
Registered: Dec 22, 2006
Posts: 1501
|
adatesman wrote: tradklime wrote: I'll definetly send you the cam and the quick links, and anything else I dig up that may be of interest. I have a couple of old SMC rings sitting around I'll send you too. Regarding the quicklinks, how many do you think would be needed to make the results somewhat representative? 5? 10? Yikes, I'm no statistician; just an ordinary guy who likes to break things.  I'll break whatever you send, so perhaps there's someone else around with experience in the field that could weigh in on the subject? -a.
Statistician here (actually economist, but in this case it'll do). I can do a couple of things for you with regards to determining the number of links that are needed to get "representative" results.
1. Calculating a 95% (or any other % for that matter) confidence interval is really easy given the results. This basically gives you a range which you can be 95% sure (sort of) that these biners will fail within a certain range. You can do this with any number of biners tested. The more you test, the smaller the interval.
2. If you know the range you want (like, I want to test my biners so I can expect with 95% confidence that other biners of the same type will fail within 1kn of the average tested) and the variance (a measure of how different the results for each biner were), I can tell you how many biners you need to test to achieve this. Unfortunately, that means you already have to do some testing to figure out the variance.
3. If you want to get really fancy and you have lots of data, I can run some regressions for you to do some more in-depth analysis. Example, you want to test how much different knots slip with different materials. You test a bunch of knots with a bunch of different materials. I take the data and can approximate how much each factor contributed to the resulting slip.
Unfortunately, there really isn't a way to tell you how many biners you'll need to break to get representative results without knowing the variance, standard error, standard deviation, or something along those lines. The easiest thing to do is to give me the results and I'll make a 95% confidence interval. Obviously, the more you test, the better.
(This post was edited by shoo on Apr 3, 2008, 12:23 AM)
|
|
|
|
|
|
|
|
adatesman
Apr 3, 2008, 1:00 AM
Post #19 of 148
(25283 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
adatesman
Apr 3, 2008, 1:22 AM
Post #20 of 148
(25270 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
shoo
Apr 3, 2008, 1:39 AM
Post #21 of 148
(25262 views)
Shortcut
Registered: Dec 22, 2006
Posts: 1501
|
adatesman wrote: Hey Shoo, So if I'm hearing you correctly (and vaguely remembering what little statistics I slept through in college many, many years ago), obtaining the degree of certainty has more to do with having samples with similar results than having a huge pool of samples?
It has to do with both.
*Turn away if you don't like math*
<Range of predicted values> = <average value> +/- [t * s/sqrt(n)]
where t is 1.96 at 95% confidence, s is the standard deviation of the sample, and n is the size of the sample (i.e. number of observations). The latter term is half the size of the confidence interval (see the +/- there?)
Basically, what you should do is break a bunch and message me the results for each biner. I'll throw you back some statistics, like standard deviation, the 95% confidence interval, etc. If everything breaks consistently, you might get some good results. Who knows.
(This post was edited by shoo on Apr 3, 2008, 1:41 AM)
|
|
|
|
|
Jbitz
Apr 3, 2008, 2:28 AM
Post #22 of 148
(25236 views)
Shortcut
Registered: Nov 14, 2006
Posts: 124
|
I would like to see the double bowline with a double fisherman backup tested with and without the yosemite backup.
The only data I have seen in the past is on the single bowline often comparing it to the figure-eight.
(This post was edited by Jbitz on Apr 3, 2008, 2:30 AM)
|
|
|
|
|
adatesman
Apr 3, 2008, 2:33 AM
Post #23 of 148
(19440 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
JesusNut
Apr 3, 2008, 5:05 PM
Post #24 of 148
(19357 views)
Shortcut
Registered: Mar 29, 2008
Posts: 7
|
Im a little curious about this one, maybe no test needd and its just common sense, but, Is it better to join two slings with a girth hitch or a biner? Lets say its a top rope and there was some slack when I fell. Or its off a lead and I extended a piece with two slings and I fell? What would work better or safer? (Two slings being equal in all aspects)
|
|
|
|
|
basilisk
Apr 3, 2008, 5:16 PM
Post #25 of 148
(19350 views)
Shortcut
Registered: Oct 1, 2005
Posts: 636
|
JesusNut wrote: Im a little curious about this one, maybe no test needd and its just common sense, but, Is it better to join two slings with a girth hitch or a biner? Lets say its a top rope and there was some slack when I fell. Or its off a lead and I extended a piece with two slings and I fell? What would work better or safer? (Two slings being equal in all aspects)
You may be interested in the tests Black Diamond did:
http://www.bdel.com/...p_archive.php#110906
This may interest you too:
http://www.bdel.com/...p_archive.php#052107
|
|
|
|
|
|
1
