|
majid_sabet
May 15, 2009, 3:11 AM
Post #126 of 211
(7239 views)
Shortcut
Registered: Dec 13, 2002
Posts: 8390
|
In reply to: nevaarrrrrrrrrrr!! and i think i believe the exact opposite. he repeats almost verbatim what he picks up from other users, except usually misses one keep point to make any of what he says a fallacy. couple on top of that, no understanding of simple physics or engineering and you have disastrously poor imformation being spewed. i think it was last year that i snapped in a thread, and since then will never read his posts with anything but disdain. (maybe a little harsh) but i sure as hell won't be open-minded (making me immature). Actually, I want and prefer people like you to judge exactly based on what you see of me in RC and nothing more. keep it going
(This post was edited by majid_sabet on May 15, 2009, 3:12 AM)
|
|
|
|
|
jt512
May 15, 2009, 3:24 AM
Post #127 of 211
(7229 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
bill413 wrote: angry wrote: jt512 wrote: bill413 wrote: jt512 wrote: bill413 wrote: jt512 wrote: bill413 wrote: USnavy wrote: Well the one I have says it scans at 1000 cycles a second and the manufacturer said it’s an appropriate choice for capturing dynamic loads. Why are you not happy with your Daytronic strain gauge? Do you think 1k scans a second is not enough? Well, my recollection of basic information theory is that you want your sampling rate greater (by at least 2 times?) than the time of the event you are trying to capture. So, what is the duration of the peak load? It's instantaneous. I guess he's screwed. Jay So, that just means he needs something that reads twice as fast as 0.0000 seconds. Seriously - the peak load must occupy some finite time. Otherwise not even analog measuring devices would accurately capture it. Also, in a fall, the greater the kinetic energy (larger mass, longer fall), the greater time over which the load occurs. If the peak force is instantaneous (and I think it is), then you can't capture it. But if readings are taken 1000 times per second, then you can measure the force within 1/2000 of a second of the peak force, which, hopefully, is good enough. Jay Umm, within 1/500 th of a second. No. If the time at which the maximum force occurs t is in the interval [0, .001s], then max(min{ t, .001s– t}) = .0005s. That's 1/2000 of a second. Jay It's on like Donkey Kong!!! Jay, I think that your expression is the time during which the transient is invisible to the sampler. That is, if at time t0 we measure a value, and at time t1 we measure a value, anything that happens within the space t1-t0 is invisble to the sampling. All we know are the endpoints. That's fine. The times at which some consecutive pair of samples were taken will be the endpoints of an interval containing the maximum force. Thus, if the length of that interval is 1/1000 of a second, at least one sample will have been taken within 1/2000 of a second of the maximum force. Jay
|
|
|
|
|
Rudmin
May 15, 2009, 3:31 AM
Post #128 of 211
(7228 views)
Shortcut
Registered: Mar 29, 2009
Posts: 606
|
But Bill, that would still make Jay correct. A sample is taken every 0.001 seconds. The peak force will lie somewhere between two samples taken 0.001 seconds apart. The furthest time the peak force can be from one of those two measured samples before it starts getting closer to the other one is 0.0005 seconds.
|
|
|
|
|
jrathfon
May 15, 2009, 3:50 AM
Post #129 of 211
(7219 views)
Shortcut
Registered: Jun 5, 2006
Posts: 494
|
k.l.k wrote: jt512 wrote: If the peak force is instantaneous (and I think it is), then you can't capture it. This is pretty fun. But let's just Kevork it. Why don't you just tell the poor fucker that "peak load" is an abstraction from empirical data and be done with it? The armchair engineers would need weeks just to assimilate that claim. Hell, the ME majors at my University would need a few days and probably some minor recreational drug use. This thread is a lovely example of everything that is stupid about this site. word.
|
|
|
|
|
jt512
May 15, 2009, 3:53 AM
Post #130 of 211
(7214 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
k.l.k wrote: jt512 wrote: If the peak force is instantaneous (and I think it is), then you can't capture it. Why don't you just tell the poor fucker that "peak load" is an abstraction from empirical data and be done with it? Because it's not an abstraction from empirical data. Clearly, there is a peak force. It's just that if it's instantaneous, then the engineers are hopelessly consigned to underestimating it when they try to measure it. Jay
|
|
|
|
|
k.l.k
May 15, 2009, 4:13 AM
Post #131 of 211
(7197 views)
Shortcut
Registered: May 9, 2007
Posts: 1190
|
jt512 wrote: It's not an abstraction from empirical data. Of course it is. Although, in the context of this conversation, that may be a banal philosophical point. If you think of "peak load" as a "concept," then it's not an abstraction from empirical data-- until you want a number. If you want "peak load" to be "instantaneous," then you're inviting the engineers to respond with hopeless attempts to measure "duration" or "event" or any number of other ultimately theological concepts. In a purely theoretical sense any system can have a measurable "peak load." But then we are fudging in a number of ways. This is a philosophical issue that is not really relevant to the underlying impulse of the OP, which is why it is a perfect way to Kevork this thread. Or spin it out into an endless series of exercises in amateur metaphysics, which is the same thing. But the lurkers needn't worry: if we add in some fart and sex jokes, and some bad computer graphics of high school physics, we can build the paradigmatic rc thread.
|
|
|
|
|
patto
May 15, 2009, 4:50 AM
Post #132 of 211
(7185 views)
Shortcut
Registered: Nov 15, 2005
Posts: 1453
|
jt512 wrote: k.l.k wrote: jt512 wrote: If the peak force is instantaneous (and I think it is), then you can't capture it. Why don't you just tell the poor fucker that "peak load" is an abstraction from empirical data and be done with it? Because it's not an abstraction from empirical data. Clearly, there is a peak force. It's just that if it's instantaneous, then the engineers are hopelessly consigned to underestimating it when they try to measure it. Jay Not if the engineers interpolate between the two points properly. But all that is irrelevant because if the force peaked at 2000kN for 0.0005s then you still would unlikely to be breaking something. You need a non negligable amount of time at high force to break stuff. There isn't much macro physical phenomenon that has real sharp edge behaviour. Even 'shock loads' have a ramp up if you get the time scale small enough. I don't see why 1000samples per second would not be enough in this case. Any significant discrepency between the analogue and digital is likely to be blamed on the equipment. In contrast if I dropped a steel ballbearing onto a steel surface then I would probably would need much greater resolution than 1000samples a second to measure the impact.
|
|
|
|
|
jt512
May 15, 2009, 5:11 AM
Post #133 of 211
(7175 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
|
|
|
|
|
curt
May 15, 2009, 5:28 AM
Post #134 of 211
(7165 views)
Shortcut
Registered: Aug 27, 2002
Posts: 18275
|
jt512 wrote: patto wrote: jt512 wrote: k.l.k wrote: jt512 wrote: If the peak force is instantaneous (and I think it is), then you can't capture it. Why don't you just tell the poor fucker that "peak load" is an abstraction from empirical data and be done with it? Because it's not an abstraction from empirical data. Clearly, there is a peak force. It's just that if it's instantaneous, then the engineers are hopelessly consigned to underestimating it when they try to measure it. Jay Not if the engineers interpolate between the two points properly. How can you interpolate between two values to obtain a value that is greater than both of them? Probably by using polynomial interpolation--or a similar best fit to the data. http://en.wikipedia.org/wiki/Interpolation
jt512 wrote: In reply to: But all that is irrelevant because if the force peaked at 2000kN for 0.0005s then you still would unlikely to be breaking something. But the duration of the peak force is 0. Does that mean that it would be impossible to break anything (How am I doing, Kerwin?). Jay The duration of the peak force may be zero, but since we are modeling the climbing rope as a highly damped spring (by application of Hooke's Law) the first derivative of acceleration near the peak becomes quite small--i.e. the peak tension in the rope basically follows a sine function and does not change rapidly near the peak. For that reason, I would think (for the case of a climbing rope at least) that 1000 sampling points/second would have relatively little error in the measurement of peak force, even without interpolation. Curt
|
|
|
|
|
jt512
May 15, 2009, 5:48 AM
Post #135 of 211
(7150 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
curt wrote: jt512 wrote: patto wrote: jt512 wrote: k.l.k wrote: jt512 wrote: If the peak force is instantaneous (and I think it is), then you can't capture it. Why don't you just tell the poor fucker that "peak load" is an abstraction from empirical data and be done with it? Because it's not an abstraction from empirical data. Clearly, there is a peak force. It's just that if it's instantaneous, then the engineers are hopelessly consigned to underestimating it when they try to measure it. Jay Not if the engineers interpolate between the two points properly. How can you interpolate between two values to obtain a value that is greater than both of them? Probably by using polynomial interpolation--or a similar best fit to the data. Yes, you're right. I looked up the definition of "interpolate" after I posted my message, and found that it included methods that would permit estimating values outside the range defined by the endpoints. I was going to amend my post, but you replied to it while I was editing it. Jay
|
|
|
|
|
rgold
May 15, 2009, 5:54 AM
Post #136 of 211
(7142 views)
Shortcut
Registered: Dec 3, 2002
Posts: 1804
|
Some random comments. 1. Rocknice2 refers to the fall force calculator at http://www.myoan.net/...t/climbforcecal.html I've posted all over the place that this calculator is completely bogus, but it is hard to get the word out. (Some later posters noticed this.) Oddly, rocknice2 exhorts us to "ignore the FF on the calculator." The calculator's inability to get the FF right is the thing that provides the clue that it is worthless. For example, try calculating the "Shock (in kilo-Newtons)" for a fall in which the distance to the last anchor is zero, namely a FF 0 fall. 2. Fishclimb wrote, "The key factor in calculating the Falling Force is the time it takes to decelerate. F=ma. Shorter deceleration times increase the force in a fall." It isn't clear what "falling force" means, but the question is, in general, about the peak load to the anchor. If shorter deceleration times were the determining factor, then tiny falls (think one inch, say) would result in the highest peak loads. 3. Majid writes that there is no such thing as a fall factor 2 in climbing. This is incorrect; any time a leader falls onto the belay without intermediate gear it is FF 2. 4. In support of this, Majid cites the article by Dan Curtis in the College Mathematics Journal. http://www.cwu.edu/...rints/cmj135-140.pdf This article is a stunning example of perfectly good mathematics that is utterly rotten engineering. Curtis, for some strange reason, uses a non-standard definition of fall factor, in which the rope stretch is included in the distance fallen. He then shows (assuming the spring model), that the maximum force depends only on this fall factor. Mathematically speaking, Curtis' approach and the standard approach are equivalent; the equations for one can be transformed to those for the other. But Curtis' equations are useless for engineering, because you have to know the rope stretch to use them. The absurdity of this approach becomes clear when you realize that knowing the rope stretch is equivalent to knowing the maximum tension in the rope for the spring model, meaning that in order to use Curtis's equations, you already have to be in posession of the answer you are interested in. I might add that Curtis fails any reasonable test of scholarship, since his analysis has been done over and over again, and correctly, and he fails to reference any of the available accounts that predate his. (As far as I know, the original calculations were done by Arnold Wexler and Dick Leonard and were published in the Sierra Club Journal in 1947.) 5. Colatownkid speaks repeatedly of the fall factor as an approximation. I think this way of speaking confounds two different things. On the one hand is a mathematical model, based on viewing the rope, at least during extension, as an ideal spring. In this mathematical model, the fall factor determines the peak tension in the rope and so the peak load to the top piece. No approximations are involved. The model is, however, just that---a model, meaning that it includes some but not all of the features of reality. The results of the model are thus approximations to the real situation. 6. The ideal spring model is the simplest model and actually seems to be fairly good at predicting loads. But it is certainly possible to create more complicated models that include features of reality that aren't part of the spring model's assumptions. Most of the more complicated models include viscous damping in either simple or complicated ways. One of the problems with these models is that there does not seem to be an accompanying explanation of why a rope would be subject to viscous damping. The damped models can be fitted to experimental data, and one can hope that they provide predictive value for situations analogous to the ones in which the data was collected, but one also has to wonder whether such models can predict rope behavior when the situation is not analogous to the one used for the trials. (A good example: by how much will a screamer reduce peak load?) 7. Halifax keeps saying that the fall factor calculation is a poor model. I guess this would need some kind of quantification, and then an agreement about what kinds of answers one wants from a model. If the spring model regularly overstates the peak load in a consistent way that is not "too far off," (definition required), then it has some value in providing a useful upper bound. 8. Attaway, in the paper mentioned by ptlong, has an interesting take on the modeling process. He speaks of "system stiffness," and some of his experimental data show that fairly complicated systems with friction and damping exhibit "system stiffness," meaning that the whole kit and kaboodle together behaves (during dynamic extension) remarkably like an ideal spring. The problem, of course, is determining a priori what the spring constant for the system is. 9. I had the impression from Attaway that the system behavior for dynamic ropes was far more spring-like (which is to say that Hooke's law applies) for dynamic loading as opposed to slow loading. If this is true, it would suggest that models employing viscous damping are really not right, since by definition the effects of viscous damping are more pronounced at higher velocities. 10. As for measuring peak load, since the rate of change of tension with respect to time is near zero as the rope tension approaches its maximum value, the tension isn't going to change all that much in one of those tiny intervals near or containing the maximum value.
|
|
|
|
|
k.l.k
May 15, 2009, 5:58 AM
Post #137 of 211
(7138 views)
Shortcut
Registered: May 9, 2007
Posts: 1190
|
rgold wrote: Some random comments. 1. Rocknice2 refers to the fall force calculator at http://www.myoan.net/...t/climbforcecal.html I've posted all over the place that this calculator is completely bogus, but it is hard to get the word out. (Some later posters noticed this.) Oddly, rocknice2 exhorts us to "ignore the FF on the calculator." The calculator's inability to get the FF right is the thing that provides the clue that it is worthless. For example, try calculating the "Shock (in kilo-Newtons)" for a fall in which the distance to the last anchor is zero, namely a FF 0 fall. 2. Fishclimb wrote, "The key factor in calculating the Falling Force is the time it takes to decelerate. F=ma. Shorter deceleration times increase the force in a fall." It isn't clear what "falling force" means, but the question is, in general, about the peak load to the anchor. If shorter deceleration times were the determining factor, then tiny falls (think one inch, say) would result in the highest peak loads. 3. Majid writes that there is no such thing as a fall factor 2 in climbing. This is incorrect; any time a leader falls onto the belay without intermediate gear it is FF 2. 4. In support of this, Majid cites the article by Dan Curtis in the College Mathematics Journal. http://www.cwu.edu/...rints/cmj135-140.pdf This article is a stunning example of perfectly good mathematics that is utterly rotten engineering. Curtis, for some strange reason, uses a non-standard definition of fall factor, in which the rope stretch is included in the distance fallen. He then shows (assuming the spring model), that the maximum force depends only on this fall factor. Mathematically speaking, Curtis' approach and the standard approach are equivalent; the equations for one can be transformed to those for the other. But Curtis' equations are useless for engineering, because you have to know the rope stretch to use them. The absurdity of this approach becomes clear when you realize that knowing the rope stretch is equivalent to knowing the maximum tension in the rope for the spring model, meaning that in order to use Curtis's equations, you already have to be in posession of the answer you are interested in. I might add that Curtis fails any reasonable test of scholarship, since his analysis has been done over and over again, and correctly, and he fails to reference any of the available accounts that predate his. (As far as I know, the original calculations were done by Arnold Wexler and Dick Leonard and were published in the Sierra Club Journal in 1947.) 5. Colatownkid speaks repeatedly of the fall factor as an approximation. I think this way of speaking confounds two different things. On the one hand is a mathematical model, based on viewing the rope, at least during extension, as an ideal spring. In this mathematical model, the fall factor determines the peak tension in the rope and so the peak load to the top piece. No approximations are involved. The model is, however, just that---a model, meaning that it includes some but not all of the features of reality. The results of the model are thus approximations to the real situation. 6. The ideal spring model is the simplest model and actually seems to be fairly good at predicting loads. But it is certainly possible to create more complicated models that include features of reality that aren't part of the spring model's assumptions. Most of the more complicated models include viscous damping in either simple or complicated ways. One of the problems with these models is that there does not seem to be an accompanying explanation of why a rope would be subject to viscous damping. The damped models can be fitted to experimental data, and one can hope that they provide predictive value for situations analogous to the ones in which the data was collected, but one also has to wonder whether such models can predict rope behavior when the situation is not analogous to the one used for the trials. (A good example: by how much will a screamer reduce peak load?) 7. Halifax keeps saying that the fall factor calculation is a poor model. I guess this would need some kind of quantification, and then an agreement about what kinds of answers one wants from a model. If the spring model regularly overstates the peak load in a consistent way that is not "too far off," (definition required), then it has some value in providing a useful upper bound. 8. Attaway, in the paper mentioned by ptlong, has an interesting take on the modeling process. He speaks of "system stiffness," and some of his experimental data show that fairly complicated systems with friction and damping exhibit "system stiffness," meaning that the whole kit and kaboodle together behaves (during dynamic extension) remarkably like an ideal spring. The problem, of course, is determining a priori what the spring constant for the system is. 9. I had the impression from Attaway that the system behavior for dynamic ropes was far more spring-like (which is to say that Hooke's law applies) for dynamic loading as opposed to slow loading. If this is true, it would suggest that models employing viscous damping are really not right, since by definition the effects of viscous damping are more pronounced at higher velocities. 10. As for measuring peak load, since the rate of change of tension with respect to time is near zero as the rope tension approaches its maximum value, the tension isn't going to change all that much in one of those tiny intervals near or containing the maximum value. Dude, it's almost 11 on the West Coast. That's like, what, next week, East Coast Time? Kudos for the epic post. That was some serious early morning labor. Grading finals?
|
|
|
|
|
k.l.k
May 15, 2009, 6:01 AM
Post #138 of 211
(7135 views)
Shortcut
Registered: May 9, 2007
Posts: 1190
|
jt512 wrote: [Yes, you're right. . . . I was going to amend my post, but you replied to it while I was editing it. Jay Whoa, stop the presses! God on ya, JT! Good work here tonight, folks. We'll see you next week. Stop off at the receptionist on your way out, and schedule your next appointment.
|
|
|
|
|
jt512
May 15, 2009, 6:01 AM
Post #139 of 211
(7135 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
rgold wrote: 10. As for measuring peak load, since the rate of change of tension with respect to time is near zero as the rope tension approaches its maximum value, the tension isn't going to change all that much in one of those tiny intervals near or containing the maximum value. Stop being practical. You're a mathematician. The expected value of the maximum of a series of measurements of impact force is less than the true maximum impact force, no matter how trivial the difference, dammit. Jay
|
|
|
|
|
curt
May 15, 2009, 6:02 AM
Post #140 of 211
(7134 views)
Shortcut
Registered: Aug 27, 2002
Posts: 18275
|
Rich is well known to be a night-owl. I suspect rc.com is his usual treatment for insomnia. Curt
|
|
|
|
|
jt512
May 15, 2009, 6:14 AM
Post #141 of 211
(7126 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
curt wrote: Rich is well known to be a night-owl. I suspect rc.com is his usual treatment for insomnia. Curt Curt, when we make consecutive posts to a thread, the juxtaposition of our profile pics seems kinda' gay. Would you mind changing your profile pic? It's not like that problem has any historical significance, or anything. Jay
|
|
|
|
|
curt
May 15, 2009, 6:27 AM
Post #142 of 211
(7121 views)
Shortcut
Registered: Aug 27, 2002
Posts: 18275
|
jt512 wrote: curt wrote: Rich is well known to be a night-owl. I suspect rc.com is his usual treatment for insomnia. Curt Curt, when we make consecutive posts to a thread, the juxtaposition of our profile pics seems kinda' gay. Would you mind changing your profile pic? Well, if you would change yours to something a bit steeper, people might stop assuming that you are my bitch. Curt
|
|
|
|
|
rgold
May 15, 2009, 1:47 PM
Post #143 of 211
(7102 views)
Shortcut
Registered: Dec 3, 2002
Posts: 1804
|
Hoho Curt, insomnia is never the issue. I just naturally like to be up late at night, it seems to be a genetic disposition in my family. When I do go to sleep, I never wake up until I have to. In this case, k.l.k. is on to something---after a full day of grading finals, I really needed to think (but not too hard) about something else before bedtime. And Jay is right too, shame on me for considering tiny differences in tension to be zero for all practical purposes. To appropriate a now-famous quote of Eliot Spitzer, what was I thinking?
|
|
|
|
|
hafilax
May 15, 2009, 3:41 PM
Post #144 of 211
(7079 views)
Shortcut
Registered: Dec 12, 2007
Posts: 3025
|
rgold wrote: 7. Halifax keeps saying that the fall factor calculation is a poor model. I guess this would need some kind of quantification, and then an agreement about what kinds of answers one wants from a model. If the spring model regularly overstates the peak load in a consistent way that is not "too far off," (definition required), then it has some value in providing a useful upper bound. I may not have made it clear but I have been trying to keep my responses in the context of the original question: Common forces in real world falls. Once people start throwing around hard numbers from fall factor calculations it makes me want to point out the deficiencies of the model. If I am standing over a piece with a good estimate of just the fall factor will I be able to predict the impact force on my protection? For long, clean into air falls with a rigid belay it will do pretty well but how often does that happen? There are too many complexities added by the real world case. For example, we know that just changing the belay device from a Gri Gri to a tube makes a big difference. There is also rope drag, knot cinching, dynamic belays and even rope recovery. I agree that it is instructive in outlining some rules of thumb and one can roughly modify that knowing the properties of the added factors listed above. I keep thinking of a section from a thermoacoustics text book I've been reading discussing device design. There is an excellent computer program that predicts the behavior of ideal devices. The author cautions the reader against becoming enamored with the model and encourages building a device earlier than later. There are a lot of strange effects that can happen and will only be found in a real world device.
|
|
|
|
|
jt512
May 15, 2009, 4:46 PM
Post #145 of 211
(7059 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
rgold wrote: 7. Halifax keeps saying that the fall factor calculation is a poor model. I guess this would need some kind of quantification, and then an agreement about what kinds of answers one wants from a model. If the spring model regularly overstates the peak load in a consistent way that is not "too far off," (definition required), then it has some value in providing a useful upper bound. If we could determine that the error really was consistent, then we presumably could measure the error, and correct for it. Hence the the model would provide more than just an upper bound. Biased predictions are ok, if you know enough about the bias. But I actually do think that there may be a fairly serious problem with the model that results in underestimation of peak force when the fall factor is less than 2. When the rope is clipped through a top anchor, the section of the rope between the anchor and the belayer stretches less than section of the rope between the anchor and the climber. Intuitively, this strikes me as a serious departure from an ideal spring with a single modulus of elasticity. However, if my reasoning is correct, then perhaps the error is mitigated, at least for fall factors near 1.78, by using a rope modulus that is calculated from the rope's UIAA impact force rating. A related problem is that we seem to be treating the anchor as frictionless for purposes of calculating the peak impact force on the climber, but then incorporating a substantial frictional force when we calculate the impact force on the belayer. We then calculate the force on the anchor by adding together two numbers that were derived using inconsistent assumptions, a practice that strikes me as a little weird.
So, questions to rgold: Am I correct that friction through the anchor is not accounted for in the model, and that therefore the model substantially underestimates impact force for low-fall-factor-falls? If so, can we (where "we" = "you") improve the model? And if not, should we perhaps ignore the reduction of the impact force on the belayer's side of the rope by friction when calculating the total force on the anchor, to compensate? Edit: I have derived an equation for maximum impact force that incorporates the effect of friction at the top anchor, and will post it soon. Jay
(This post was edited by jt512 on May 19, 2009, 3:41 AM)
|
|
|
|
|
desertwanderer81
May 15, 2009, 5:42 PM
Post #146 of 211
(7019 views)
Shortcut
Registered: Sep 5, 2007
Posts: 2272
|
bill413 wrote: jt512 wrote: bill413 wrote: USnavy wrote: Well the one I have says it scans at 1000 cycles a second and the manufacturer said it’s an appropriate choice for capturing dynamic loads. Why are you not happy with your Daytronic strain gauge? Do you think 1k scans a second is not enough? Well, my recollection of basic information theory is that you want your sampling rate greater (by at least 2 times?) than the time of the event you are trying to capture. So, what is the duration of the peak load? It's instantaneous. I guess he's screwed. Jay So, that just means he needs something that reads twice as fast as 0.0000 seconds. Seriously - the peak load must occupy some finite time. Otherwise not even analog measuring devices would accurately capture it. Also, in a fall, the greater the kinetic energy (larger mass, longer fall), the greater time over which the load occurs. No, it doesn't. You're not reading peak time. You're reading "almost" peak time. On a 1/1000 sec frame rate, the furthest from the peak rate that you can possibly be is a function of the increase in force multiplied by the max amount of time away from the max. the furthest away you can be is the frame rate/2, because at no point in time can you be further than 1/2 of a frame from the peak. That means that you are at most 1/2000 seconds from the peak force. The closest you can be to measuring the peak force is 0. Therefore, on average, you are 1/4000 seconds from measuring the peak load.
|
|
|
|
|
ptlong
May 15, 2009, 6:16 PM
Post #147 of 211
(7008 views)
Shortcut
Registered: Oct 4, 2007
Posts: 418
|
jt512 wrote: So, questions to rgold: Am I correct that friction through the anchor is not accounted for in the model, and that therefore the model substantially underestimates impact force for low-fall-factor-falls? If so, can we (where "we" = "you") improve the model? And if not, should we perhaps ignore the reduction of the impact force on the belayer's side of the rope by friction when calculating the total force on the anchor, to compensate? Jay, The simple spring model does not include any terms for friction. As you note this gets tacked on without including its other effects. There are improved models. Both Martin Pavier and the Italian Alpine Club have created models that fully include the effects of friction from the carabiners in the system. These models incorporate damping in the rope itself. The Italian version also models the belayer mass and the belayer's hand (which they found to be important). Both models were compared to experiment. Of course with these models you can no longer write out a simple formula like the Wexler equation. You need a computer. I don't recall that in either case that direct comparisons were made between the simple and more sophisticated models. It would be interesting to see where and how they diverge. Maybe when rgold awakens he can tell us more.
|
|
|
|
|
adatesman
May 17, 2009, 10:30 PM
Post #148 of 211
(6945 views)
Shortcut
Registered: Jul 13, 2005
Posts: 3479
|
|
|
|
|
|
patto
May 17, 2009, 11:54 PM
Post #149 of 211
(6931 views)
Shortcut
Registered: Nov 15, 2005
Posts: 1453
|
adatesman wrote: Only if they do their scanning digitally. The beauty of analog is there's no such thing as scan rate and capturing the peak is fairly easy. Think old-school mechanical dynamometer with a second needle that will mark the furthest travel of the readout needle. Doing it electrically is a bit harder to do, but certainly possible and built in to the Daytronic 4077 strain gage indicator. Love your work adatesman, but I gotta disagree with this. While a sytem may be analogue you got to be aware that it still has limitations in measurement. Even assuming precise measurement AND a measuring device that doesn't affect the quantity being measured you still need some finite amount of time to perform the measurement. The very fact that an instantaneous peak has effectively no energy indicates that it is impossible for such foce to influence a measuring device. All devices have some sort of inertia to overcome be it the inertia of a guage or the of the actual sensor.
|
|
|
|
|
JimTitt
May 18, 2009, 7:48 PM
Post #150 of 211
(7645 views)
Shortcut
Registered: Aug 7, 2008
Posts: 1002
|
Well not really, the strain guage is measuring constantly or at least as fast as those electrons can move! Normally you then sample, any decent PCI board is running at about 200,000 a second per channel and pretty standard are ones at 500.000 samples per second for single channel. BUT this has to be processed so then you either wait 20 years for your PC to do something, buy a Cray or write an algorithm to turn this into something useful. Industry standard for climbing gear is pretty well 1000 a second. Just to record a peak load as Daytronic does it is relatively simple and elegant with the time lag being pretty small-like how long did it take for those electrons to move. This is your finite time to perform the measurement (more precisely to record it). Whether or not this if all of pure academic interest is another thing!
|
|
|
|
|
|
|
|
|