The physics toolkit

[jaydeehess]

NIST didn't have the camera. Are you suggesting all of their data is crap ?

Any image distortion resulting from camera optics will be consistent across frames. The graphs presented are all relative changes between frames.

Don't use the data or any results derived from it if you are not satisfied with its validity.

Bye

NIST was not attempting the precision for a minute forensic treatment of this video that you are.

How do you explain faster than free fall?

If careful measurements show a vertical acelleration of g+k then the results have to be at least +/- that k factor. They could also be +/- 2 times k, or 3 times k, or 5 times k.

Until you can justify accuracy that includes this k factor at the very least, then your results are simply not forensically valid, sorry. Perhaps you have work to do but then again the skid mark Chandler didn't bother with it either before jumping to his politically motivated conclusions.

 

[femr2]

So it was either vertical and horizontal flexing or it was a slower collapse period..

I'll rephrase. Strongly suggesting flexing. I qualified why (which you've snipped). As the vertical trace returns to zero in advance of release, the pre-release movement cannot be actual descent, unless you are suggesting that before descent the building got bigger.

As I pointed out elsewhere, any representation of faster than freefall indicates that there is at the absolute very least and error as large as the difference between the local acelleration due to gravity and this calculated acelleration.

You should read back through the thread.

Unless some TM member can point out the machinations in effect that allowfaster than gravitational collapse.

Drop the *TM member* rubbish please.

NIST traced the movement of the roofline near the *kink*. Their acceleration fit reaches 34ft/s².

Other members, such as W.D. Clinger, have also speculated about the possible reasons why faster-than-freefall acceleration of certain building features would be possible.

Assuming that the NIST trace is wrong, and that the location traced is at freefall (rather than exceeding freefall), then comparison between movements at that location and the NW corner clearly show that the NW corner reaches a higher acceleration. Possible reasons for this are obvious, as the centre of the building drops in advance of the NW corner, and so would act in a *slingshot* manner.

I am suggesting that a transition to very near free fall can indeed take place without explosives.

There is little (and probably zero once I've finalised scaling metrics) doubt that portions of the roofline attained at least freefall.

What is your opinion on this detail?

Further analysis to determine the arrangement of perimeter columns over the first 13 or so floors is required to make a conclusion.

You state that the building may well have been twisting

Most certainly twisting.

and that this is most visible at the top floors.

I didn't say that, but it's a reasonable assumption.

So your own interpretation of this data suggests that this twisting rather than explosives at or below the 12th floor are responsible to these windows shattering.

No, that is your interpretation. You keep mentioning explosives, not I.

But you get faster than free fall and thus there is that absolute minimum error margin that is patently obvious.

No. Again, read back through this thread.

I expect that they may be able to assist you in this if you cannot do it on your own.

I have no intention of seeking assistance from such groups.

Remember, it is the TM that is trying to do a forensic investigation on this minutia.

No, it is me providing accurate trace data on building feature movements. Your repeated mention of the *TM* is becoming a little tedious.

 

[femr2]

NIST was not attempting the precision for a minute forensic treatment of this video that you are.

They used techniques to determine movements down to +/- 1 inch, followed by interpretation of that movement. Have you read the last few pages of this thread ? If not, suggest you do. All of theail above about moire is directly related.

How do you explain faster than free fall?

See previous post, and last couple of pages of this thread.

Until you can justify accuracy that includes this k factor at the very least, then your results are simply not forensically valid, sorry.

Sorry, but you're going to have to prove the data and scaling wrong, and also contact NIST asking why their acceleration curve reaches 34ft/s².

I'll be including scaling metrics fairly shortly.

 

[pgimeno]

That's a small image. Here's a bigger one...

Thanks. Trying my best to fit the pixels instead of just trying to find an upper and lower bound, got me that the NE distances were 7.3% smaller than the NE corner ones (or equivalently, the NW ones were 7.9% bigger). With the previous pic, the figures were -7.0% / +7.5%. I am more or less confident that it's something between 6.5% and 7.7% reduction (equiv. between 7.0% and 8.3% magnification). Or in maybe simpler terms, that the factor is between 0.923 and 0.935 for one direction or their inverses in the opposite. But it's estimated by eyeballing and using whole pixels all the time, so don't take it for granted.

Distances ? The camera is quite a distance away, so distance-to-camera differences are slight. It's more the skewed view that would have an effect.

I think distance and skew are two different ways of stating the same thing. Note that I am thinking in terms of computer graphics, where the transformation from a 3D scene to the 2D screen needs division by distance from the camera to the point to represent, and that's where perspective deformation is generated (not dividing gets you an isometric view).

Have spent far too long (very sad) working out various metrics for the view, and will post here with the data in real-world units when I can get around to finishing it.

Just in case it's not clear, I'll remark that the reason I started discussing about the distance factor between the NW and the NE corner was in relation to this sentence you wrote:

NW Corner movements are larger than the NE corner movements (especially horizontally) which may be an effect of camera perspective, as the NW corner is closer to the camera.

With that factor in hand, you could rule out these effects.

I use various methods.

The question is if you can trace the straight line determined by one row of windows or even a building's edge with sub-pixel accuracy. If so, you can find a better factor than mine.

I'm not sure if you got the idea on how to find the factor to use for compensating perspective. I'd say you didn't since you answered "Cool" to my offer for a diagram and further explanation, but I'd like to be sure before I spend time on any.

 

[femr2]

you could rule out these effects.

I'd hope the result is a scaling factor to change the relative magnitudes. The scaling factor for the East side will be different to the scaling factor for the West side, which'll result in the relative size of the curves changing.

The question is if you can trace the straight line determined by one row of windows or even a building's edge with sub-pixel accuracy. If so, you can find a better factor than mine.

It's only movement of features that benefits from the sub-pixel tracing techniques, but viewing the image in, say, SynthEyes, allows the cursor area to be blown-up *8 with Lanczos3 filtering, realtime contrast, gamma and wotnot applied. I'd suggest *placement* could be sub-pixel (eyeballed), but still-image dimensions can only really be stated as pixel-accurate (at best). The graphs posted earlier of window spacing were to illustrate using multiple measurement points and then using that data to generate linear fits to try and reduce the margin of error (noise).

I'm not sure if you got the idea on how to find the factor to use for compensating perspective. I'd say you didn't since you answered "Cool" to my offer for a diagram and further explanation, but I'd like to be sure before I spend time on any.

Yeah, got it, but there's nowt wrong with placing a description in the thread. The more differing ways the merrier. If they all come out with the same results, the results are reinforced.

ETA: Thanks for the metrics. I'll see how they compare to mine fairly shortly. Imagine they'll be very similar. Will then regenerate all graphs in real-world units.

 

[pgimeno]

Yeah, got it, but there's nowt wrong with placing a description in the thread. The more differing ways the merrier. If they all come out with the same results, the results are reinforced.

Well, this shows the measurements I've been talking about, where the range came from:

It's an animated gif (2 frames, 3 seconds per frame), each frame showing how the corresponding interval was measured. In this case, instead of a row of windows I chose the top border of the building, but I don't think that was a good idea. With the bigger image I chose the top row of windows.

Any two rows of windows can be selected, but the farther they are from each other, the better the precision will be.

The green lines are the ones that are measured. For extra safety I measured from the bottommost part of the topmost pixel to the topmost part of the bottommost pixel to exaggerate towards small, and vice versa, as can be seen in the green lines when the image is zoomed.

The drawback of all these safety measures was the awful precision.

 

[femr2]

Well, this shows the measurements I've been talking about, where the range came from.

Thanks. Similar to my own methods, though I generate a grid of lines over the surface (doesn't really help much tbh).

For vertical axis, I get very similar results... ~0.92
For horizontal axis...~0.89

Draft values of course.


Applying those multipliers to the original graphs in question doesn't result in the curve magnitudes evening-up.

I'd estimate maximum NW corner horizontal motion is still double that of the NE corner. More than double in general.

Minor error in the multipliers wouldn't make much of a dent in the magnitude differences.

 

[pgimeno]

For horizontal axis...~0.89

If I'm understanding you correctly, horizontal axis doesn't make sense with this method, since the horizontal lines are not "almost parallel" to the camera as are the vertical ones. The value you get will depend on the height of the window row you pick, while the other ratio doesn't depend on it (or does very slightly). You have to apply the same factor to both axes. That way of proceeding (using the same factor) assumes that the distance to the camera of the point being studied doesn't vary substantially with the movement in either axis, which is a safe assumption.

A rough estimation on the factor to apply to top vs bottom is that it would be around 1% or less, and it's a result of the slight tilt of the camera upwards. A visible effect of that tilt is that neither edge of the facade is exactly vertical in the image, but the left edge is slightly tilted to the right and the right edge to the left. Simplest thing to do here is to neglect the camera tilt for this purpose.

The correct way would be to use the four points obtained with the intersection method I explained, to calculate the plane in which the facade lies, and use that plane to find the ratio wrt the (x, y) position as a whole, not just on x or y separately, since these are just approximations. I'd have to search for information on how to do that, though. I'm guessing that knowledge of the untransformed distance ratio between the horizontal and vertical lines (e.g. by knowing the real-world measures) will be necessary for that purpose.

Applying those multipliers to the original graphs in question doesn't result in the curve magnitudes evening-up.

Don't apply the horizontal axis one. Use the vertical one for both axes. That's assuming I've understood you correctly on how you calculate the horizontal one (I'm assuming that you measure the length of the horizontal lines and divide them to obtain the ratio, and that's incorrect; only the vertical lines ratio is meaningful).

Minor error in the multipliers wouldn't make much of a dent in the magnitude differences.

The only conclusion I can draw then is that they don't move uniformly. That's something I already expected after calculating the ratio: looking at the graphs, the differences didn't look to be around 7% but much bigger. Maybe it's a pivoting action on a non-centered pivot point what is influencing the movement.

 

[femr2]

If I'm understanding you correctly, horizontal axis doesn't make sense with this method

I'm not using the same methods.

The horizontal multiplier is determined by placing (almost) vertical lines aligned with each (horizontal) inter-window-space, then determining variation in horizontal distance between each windows. Plotted and linear fit. Tried it with per-5 and per-10 window multiples with very similar results.

Will have to determine an additional multiplier for lower down the building, but not required for the upper corner multipliers in this context.

The only conclusion I can draw then is that they don't move uniformly.

I concur. Scrubbing through the video many times reinforces that impression.

Maybe it's a pivoting action on a non-centered pivot point what is influencing the movement.

Would make sense. Will have to clarify with a rotoscope.

 

[tfk]

femr,

Sorry, been busy.

First off, you are welcome & invited to show off your "kinematic visualization".

And having scanned over the dialog in the pertinent threads, I am perfectly comfortable having that conversation stand as a prime example of both your & my relative engineering experience.

First, the trivial.
Your absurd precision claims for heavy steel plate & massive welded steel assemblies.

Based on overlaying absurdly toleranced models onto (75 dpi?, 150 dpi?) scans of 40 year old (1:50? 1:100? scale) architectural prints.

Anyone wants to read the humor of it all can find it here:

Your original post:

Column stacking and relative placement is accurate to 0.0001 inch ...

When I commented:

... You don't need, and you don't WANT, 0.0001" accuracies on parts that are this massive.

You replied:

[re:] "You don't need, and you don't WANT, 0.0001" accuracies ..."

Oh I definitely do. ...
...
The purpose of stating accuracy is to clarify the care used to build the thing from the information available.

You don't "clarify care used to build" your model by claiming absurd, ridiculous precision. You demonstrate care used in building your model by claiming reasoned, justifiable, and accurate precision.

By claiming absurd precision, you demonstrated massive carelessness. You demonstrate a lack of comprehension of the parts & assemblies that you are modeling. You demonstrate a lack of understanding of the purpose & reliability of the architectural prints.

You would have demonstrated carefulness by saying things like:

"Sorry guys. These are really lousy scans, but they're the best info that I've got.

Since I'm not going to try to do any stress, deflections, etc. analysis, the dimensions, thicknesses & tolerances of the columns & their webs are irrelevant.

For column placement, the scans are so bad that I can't even read the dimensional call-outs on the print. So I'm forced to position the columns based on doing overlays. Yeah, I know this is famously risky. And the prints themselves have been folded, stuck in a draw, were never dimensionally stable, etc. And are not even guaranteed to be correct.

And with a (maybe) 75 dpi scan, the very best that I can do is ±1 pixel, which produces a scaled tolerance is about ±.013". And at the drawing scale, this turns into a little over about ±1". Under the best of assumptions.

But it's the best I can do. So take the absolute locations with a big grain of salt. Since we're looking at gross motions, these issues are probably minor issues in the big picture, anyways."

THAT would have impressed me with your knowledge of the design & documentations process, with the parts as fabricated, as assembled, as transported, as hoisted, as, as pulled, jockeyed muscled and welded into place.

Instead, you said "Column stacking and relative placement is accurate to 0.0001 inch".

Followed by:

If anyone has any suggestion of how placement accuracy could be improved, please let me know asap.

And, when I pointed out this silliness, you (& the gang) then followed it up with 6 pages of ... childish crappola.

All right. That's enough of the fluff. In my next post, I'll get to the meat of the discussion: The engineering & art of simulation.

[Yup, femr. "Simulation".]


Tom


PS. BTW, femr, re:

Stick to confusing 'stacked tolerances' with noise in real world data.

Part tolerances (i.e., deviations from "ideal" dimensions) are precisely "noise in real world data". In every way imaginable.

Too bad you're not able to get a handle on that...

 

[jaydeehess]

I'll rephrase. Strongly suggesting flexing. I qualified why (which you've snipped). As the vertical trace returns to zero in advance of release, the pre-release movement cannot be actual descent, unless you are suggesting that before descent the building got bigger.

Nah, the improbability drive is no where in evidence

Drop the *TM member* rubbish please.

Well it was not specifically directed at you but you certainly must be aware that Chandler's and now your forensic investigation of these approx 3 seconds of Sept.11/2001 will be or are being used by those who do identify themselves with the 911TM.

NIST traced the movement of the roofline near the *kink*. Their acceleration fit reaches 34ft/s².

Other members, such as W.D. Clinger, have also speculated about the possible reasons why faster-than-freefall acceleration of certain building features would be possible.

okie dokie then, if its an indication that a slinging effect on the upper storeys occured due to vertical flexing if the structure causing a greater than gravitational initial force on the north side of the struture. Obviousy that initial force could not continue throughout the fall. Unless this 'spring' is releasing that energy over the course of two and a half to three seconds.

Assuming that the NIST trace is wrong, and that the location traced is at freefall (rather than exceeding freefall), then comparison between movements at that location and the NW corner clearly show that the NW corner reaches a higher acceleration. Possible reasons for this are obvious, as the centre of the building drops in advance of the NW corner, and so would act in a *slingshot* manner.


There is little (and probably zero once I've finalised scaling metrics) doubt that portions of the roofline attained at least freefall.

Except that now, with the inclusion of a slinging effect you really would have to quantify it to determine just what it means. Well, maybe not you personally since you are not concerned with what the conclusions, drawn by the 911TM, from this are.

Further analysis to determine the arrangement of perimeter columns over the first 13 or so floors is required to make a conclusion.

Yep.

Most certainly twisting.

I believe I was allowing for that assumption.

I didn't say that, but it's a reasonable assumption.

I know you did not say it, sorry if I was putting words in your mouth but as you say, if its twisting due to some destruction in the lower core then it will be amplified at the far top end of the columns.

No, that is your interpretation. You keep mentioning explosives, not I.

I know you get your back up whenever explosives are mentioned but there are basically two choices here (discounting space-a-beams) , fire/mechanical damage and a horizontal progression of core failure leading to global collapse,
OR,
explosive severing of columns(either first destroying the core then the intervening column/perimeter columns, or blowing it all at once) in the lower dozen or so floors.

Thus, I mention explosives, as much for other viewers of this thread as for you.

I have no intention of seeking assistance from such groups.

Then if its a minute forensic investigation you wish to do on this 2-7 seconds I guess its up to you. Personally I cannot blame you for not wanting to get into bed with those groups. It would make my skin crawl too.

No, it is me providing accurate trace data on building feature movements. Your repeated mention of the *TM* is becoming a little tedious.

As much as you wish to keep your distance from the *TM* your work IS going to be used by them to draw conclusions. I would expect that Chandler himself would see your work as bolstering his conclusions.
Its been the bane of all science and technical research to be used by others to draw conclusions that the researcher does not intend or even want.

Certainly you knew this going in and I am sorry if you keep hearing TM this and TM that but like it or not you are now linked to the *TM*.

As for the twisting and slingshot effects , it should be noted by anyone wishing to draw conclusions about the cause of the global collapse that the actual rate of collapse means nothing at all as far as requiring the use of explosives or not since such effects are of course, unquantified.

It is this twisting and slinging that is being used to explain results of faster than free fall acelleration and thus if this is that significant a force then it must be accounted for, quantified, for any forensic work that is going to be used to render conclusions in the manner that the 911TM will inevitably be trying to arrive at.

Concluding the use of explosives from femer's work or that of others is a leap to conclusion driven primarily by a political world view that demands large scale conspiracies and subterfuge.

My OP in this thread wanted an accounting of the error margins since Chandler had been using this to conclude that explosives must have been used. Now you have done a more minute examination of these few seconds of the collapse and , arguably by some, you have decreased those error margins but simply added more complexity to the description of the collapse making a leap to conclusion of explosives as great or longer.

 

[tfk]

The meat: Simulations 101

femr: [paraphrase] "... it's not a simulation. It's a 'kinematic visualization'."

Yeah, and you can call a dog's tail a leg. And then claim that "dog's have 5 legs."

Don't make it true.
___

What is a simulation? It's some sort of a representation, which attempts to elucidate some tiny group of features, of "the real thing"

Some are elaborate. Like 3D FEA.
Some are simpler. Like CAD drawings.
Some are simpler still: Like pencil sketches, stick figures & napkin drawings.

Some are dynamic. Like physics based element collision models & videotapes.
Some are static. Like force diagrams & photographs

Some are concrete. Like parts taken from the same lot build as "the real thing".
Some are abstract. Like equations, verbal descriptions & all analyses of real world parts & events.

They are ALL "simulations".

And so is your "kinematic visualization".
___

Furthermore, none of the above simulations show "the real thing" precisely.

Every single one of the simulations mentioned shows nothing but a minuscule subset of all characteristics of an object or a process. They ALL have limitations, flaws, errors in their portrayals, descriptions of the real thing.
___

The art of simulation is weeding out the insignificant factors & retaining the important ones.

And a good sim identifies those characteristics & processes clearly, and then attempts to get as many as the most critical ones as accurate as needed.

You did the first thing. You said clearly

To help visualise various [collapse] initiation mechanisms, it will be pretty useful.

You later suggested to Tony Sz that it'd help him "visualize his column to column contact" during collapse. [paraphrased]

The problem with your sim is that you modeled the WTC core columns "as built". This model bore little resemblance - in any way that was important to collapse - to the "real thing": the one that did collapse. That is, it had none of the damage.

So, your stated goal of "visualizing various collapse initiation mechanisms" had little hope of elucidating those mechanisms.

I pointed this out to - politely - here. And provided what I believe to be valuable suggestions as to how to improve your model.

The doo-doo promptly flew into the fan.

Ah, well.


Tom

So, tell me. What did you guys ultimately learn about collapse mechanisms from your 'kinematic visualization'?

 

[jaydeehess]

"kinematic visualisation" is better than how the 911TM characterizes NISTs renderings. IIRC they refer to them as 'cartoons'.

 

[W.D.Clinger]

what slingshot effect?

okie dokie then, if its an indication that a slinging effect on the upper storeys occured due to vertical flexing if the structure causing a greater than gravitational initial force on the north side of the struture. Obviousy that initial force could not continue throughout the fall. Unless this 'spring' is releasing that energy over the course of two and a half to three seconds.


Except that now, with the inclusion of a slinging effect you really would have to quantify it to determine just what it means. Well, maybe not you personally since you are not concerned with what the conclusions, drawn by the 911TM, from this are.


As for the twisting and slingshot effects , it should be noted by anyone wishing to draw conclusions about the cause of the global collapse that the actual rate of collapse means nothing at all as far as requiring the use of explosives or not since such effects are of course, unquantified.

It is this twisting and slinging that is being used to explain results of faster than free fall acelleration and thus if this is that significant a force then it must be accounted for, quantified, for any forensic work that is going to be used to render conclusions in the manner that the 911TM will inevitably be trying to arrive at.

I believe it was femr2 who, in post 282, first suggested some kind of "*slingshot*" mechanism in this thread. Although some weird mechanism is possible, I have not seen evidence for anything I would describe as a slinging mechanism.

In posts 216 and 231, I pointed out that the northwest corner doesn't begin to move until some time after the roof has begun to fall. That implies some kind of rotation of the roof and a positive initial component to the downward velocity of the roof's center of gravity at tfk's t=0, when the northwest corner began its descent. That in turn implies that the vertical fall of the northwest corner should be modelled by the equation

y(t) = v₀t + ½gt² - f(t) - g(t)

where f(t) and g(t) are functions that model the vertical components of the (continuing) rotation and any resistance being offered by the compromised structure. (Both of these functions are presumably non-negative until some time after t=0.) The faster-than-freefall meme comes from using the equation

y(t) = ½gt²
which ignores the roof's downward velocity v₀, acquired before the northwest corner began its descent.

Another way of looking at the problem is that tfk erred in his choice of time origin. That error is understandable, because he was looking only at the sampled position data for the northwest corner as provided by femr2. The graph in femr2's post 222 clearly shows that the northeast corner and "near kink" began their descents before the northwest corner. He offered a related graph in post 243.

I just now noticed that femr2 may have provided a link to numerical data for other roof features in his post 271, but those spreadsheets are so poorly commented that I can't figure out what the columns mean, so I'll just ignore them until femr2 provides an explanation. The graphs in post 272 give some idea of how misleading it would be to calculate an acceleration for the roof as a whole using nothing but the position of the northwest corner.

I'm kinda bored by the bickering that has dominated this thread since post 201. I see no great mystery here, nor do I see any major problems with femr2's data. If and when femr2 explains the spreadsheet data he linked to in post 272, I'll run the numbers for the model I described in posts 216 and 231.

 

[jaydeehess]

So, tell me. What did you guys ultimately learn about collapse mechanisms from your 'kinematic visualization'?

Well, , if I may step in here tfk;

From my POV I asked for a quantization of the margin of error in Chandler's work. Not sure if that ever came forth. Instead femr provided a more minute examination of the event. He has claimed to have greatly narrowed , and at least quantified that margin of error. In that his work on this makes Chandler look like he was playing in a sandbox.
OTOH, femr although femr has claimed that the faster than free fall acelleration is bolstered by his work it begs the question that inspired my original request for a margin of error-
How does a structure collapse faster than gravity could propel it?

To that end, and knowing that it requires and extra force, femr and others have speculated that the building underwent a slingshot effect, in essence , it fell, bounced, and sprang back down. On that second downward travel 'stuff' broke. Since it had now started that downward travel with not only gravitational force driving it, its acelleration was also greater than that due soley to gravity.

The lower limit for that spring (slingshot) acelleration (s) would be the difference between the acelleration of collapse (c) and 'g'.
c=g+s (lower limit)

This would assume no resistance whatsoever from the lower structure

If there was some resistance then there is essentially an upward acelleration(actually a force but let's assume equal mass at all times)
c=g+s-r
IIRC c is being calculated to be (in round terms) 1.01g
1.01g=g+s-r

0.01g=s-r
Since we do not have a calculation for this slingshot effect's acelleration perhaps it was as high as 0.5g, or more, we just don't know. For the sake of arguement let's put it at 0.5g

0.01g=0.5g-r

This would mean that the lower structure offered a resistive force almost half that of the force of gravity on the mass of the upper structure.

"WHAT?" cry some in the 911TM, "this structure held the entire building up for years!!Now its only , suddenly , capable of supporting half that mass??1!1!!!1!!!1eleventy"

This ignores the fact that the south side of the structure was badly damaged, reportedly creaking , swaying, leaning, from a time very shortly after WTC 1 fell, and that this structure, at the time of onset of global collapse, had suffered an interior collapse that had seen the penthouse and rooftop structures fall into the center of the building.
In other words, at the time of the onset of global collapse it was a long way from the pristine structure that had held up for decades.

 

[jaydeehess]

What slingshot effect?

I don't know, it was postulated by femr2, but he credited W.D.Clinger with the idea, didn't he?

I really don't give a rat's rectum who came up with it. My OP was inspired by the faster than free fall acelleration. Since its pretty obvious that large rocket packs are not driving the collapse, nor does anyone in this thread (so far) believe in sci-fi space beams, then there is either another mechanical explanation
OR
it is the result of an as yet unquantified margin of error in the calculations.

Basically three choices;
1) a mechanical storage and release of energy
2) an measurement error
3) magic(sci-fi)

If there are other choices I'll be back in a week or two.
(out of town for work, then vacation)

 

[W.D.Clinger]

I don't know, it was postulated by femr2, but he credited W.D.Clinger with the idea, didn't he?

No, he did not. He noted that I had "speculated about the possible reasons why faster-than-freefall acceleration of certain building features would be possible". In the following paragraph, femr2 offered his own speculation about a slingshot mechanism. It appears to have been jaydeehess who incorrectly attributed the slingshot idea to W.D.Clinger.

I really don't give a rat's rectum who came up with it.

Fine. I hope you will stop blaming me for it, however.

Basically three choices;
1) a mechanical storage and release of energy
2) an measurement error
3) magic(sci-fi)

In posts 216, 231, and 294, I have suggested a fourth choice: an over-simplified model of the roof's motion (or, if you prefer, tfk's unfortunate choice of time origin). If you don't understand how neglecting the roof's initial velocity at t=0 can lead to an incorrect conclusion of faster-than-free-fall acceleration for the roof as a whole, then you have a little homework to do before you rejoin this conversation.

 

[femr2]

I have not seen evidence for anything I would describe as a slinging mechanism.

Perhaps not a great choice of word, but just an attempt to describe a side-effect of attachment to other parts of the building which move in advance of the NW corner...which is probably not a great description of the movement either.

those spreadsheets are so poorly commented that I can't figure out what the columns mean, so I'll just ignore them until femr2 provides an explanation.

No worries Raw data columns are (frame, pixel x, pixel y, fom). Will upload a copy with better presentation soon.

I'll run the numbers for the model I described in posts 216 and 231.

Cool.

 

[femr2]

t=0

Always a problem. Similar tracing for the Sauret footage has allowed for much improved t=0 definition, but the *flexing* of WTC7 adds variation to the vertical trace masking potential early vertical creep.

 

[tfk]

Hey guys,

There was nothing sacred about that time=0 selection that I chose.

I needed a starting time reference in femr's data to fit a curve.

I arbitrarily picked the last positive value for y before the y-value went negative and stayed negative. And therefore was unequivocally part of the descent. See below. I shifted all the data points by setting this point's coordinates to {0,0}.

By definition, the real start of the descent begins with a zero vertical velocity, and the slope of the position vs time curve needs to be zero.

From looking at the filtered data below, two things are pretty clear:

1. the initial value for the vertical position is not equal to 0, but approximately 0.6.

2. The previous t0 I picked is therefore too late. The slope is not zero at this t0. A reasonable t0 in femr's data should be around the {4.3, 0.56} point that I've indicated.

I'm not at all sure what the prominent dip is before the descent. It looks like the roof line lurched as something broke just before the roof gave way.

Tom