New FDR Decode

reply to post by tomk52


Wow Tom, thanks for opening your ignorant "internet mouth" again.

Let me see if I can recap the list of errors you have made in the last
two posts:

1. There is nothing fancy about the word "electric". It is what, it is...


2. Unlike your blind incorrect assumption about the aneroid altimeter,
I have made a factual statement about the 757-200 complete with
credible reference.

You on the other hand have run around hell's half acre with this lie about
"50 foot error" and you didn't even have a CLUE about the aircraft
WHICH YOU ADMIT IN YOUR POST ON PAGE 33! Incredible!

What is this, the second coming of "anti-sophist"?

3. There is NO error associated with the device as you claim because
it's not even the SAME FREAKIN' INSTRUMENT!
Furthermore as you have so conveniently overlooked, if PA is faulted
THEN VSI WILL REFLECT! Have you taken the time to browse the data
file before incorrectly 'assuming'? No. YOu have not. Look in the NTSB
data, and if Warren's decode provides VSI, please show us the last four
seconds worth.

4. Do you know what an ABSOLUTE PRESSURE SENSOR is? I really
did expect you to understand the difference between absolute and gauge values.

Hint: It has nothing to do with "absolute readings"

content.honeywell.com...

Yes IT MUST reflect because the SAME SENSOR produces signals for
VSI and PA from the AIR DATA COMPUTER.

If one reading is off, then the other MUST BE.

IE: A change in altitude in one second MUST reflect the change in
vertical speed in one second.

I will refrain from stooping to your level as these last replies from other
members have taken a more civil approach.


[edit on 26-11-2009 by turbofan]

Originally posted by cesura
As for the divergence between pressure altitude and my
calculated altitude: Either curve can be moved up or
down, without changing its shape, by applying a simple
additive correction (which, for the pressure altitude,
is needed anyway to adjust for barometric pressure).
So the curves can be made to coincide at either end of
the graphed interval, with the divergence at the other
end.

Yes, as I mentioned, I get the same shapes when I plot them in Excel, so we are on the same page relative to those.

I just wanted the upper and lower range of 'true altitude' based upon terrain and RA. That gave me something to plot my correlation points to for fine tuning. I am pleased to announce that thanks to that effort and one 'anomaly' in the chart above I was able to fine-tune my radar fit significantly. Although not perfect, it is about the best that can be done and remain data-driven. Your point 0 is an elongated range of +/- 0.25 seconds along the path-of-flight, which translates to ~0.033 (+/-) nautical miles at the final 483 knots recorded.

The revised offsets are:

Latitude offset = -0.0004
Longitude offset = 0.001

Based on this correlation, the final subframe ends at either the generator or Pentagon itself. The final subframe begins midway between the pole strike events. Although I see no delineated single pole strike, there is a definite change in long acc between the first recorded value and second value. This corresponds with the pole's #4 and #5 encounter in the correlation model. I'll leave it to you guys to sort that out. I've been three years just getting the positional data correlated and don't feel like taking on any more mind-numbing challenges

To those of you in the US: Have a nice Thanksgiving.
I'm off to cook some Texas chili.

Same to you sir!

Update: I know some people work better with pretty pictures than with numbers, so here is a graphic to make it easier for others to understand.

[atsimg]http://files.abovetopsecret.com/images/member/509a38790aea.jpg[/atsimg]

First, two time offsets are important to note. The NTSB CSV file uses a time stamp that matches IAD ASR radar time + 1 second. The Warren decode and P4T RO has a time stamp that matches the DCA and ADW ASR radars (IAD + 5 seconds). This accounts for the 4 seconds offset between the RO and NTSB CSV. The actual value is 4.7 seconds between IAD and DCA (one ASR radar sweep). Matching by altitude gave good results for for 4 seconds (IAD time) and 5 seconds (IAD + 1). I started with the lower value of 4 until I started fine-tuning the final seconds and found that 5 seconds gives the best fit for position and altitude modeling. Keep in mind I started this years ago with +/- 2 seconds.

Note at position (yellow arrow) -18 in the correlation (stole the scaling format from censura to match his efforts) is almost a dead-on for the ADW return position (brown dot) at 13:37:34. That corresponds to subframe 151350 and RO time of 13:37:34. A perfect fit with the tme correlation model. Position -15 corresponds to IAD (pink dot) 13:37:32 and DCA (blue dot) 13:37:37. The RO time (subframe 151353) of 13:37:37 is again a match for the time correlation.

Just can't get it any closer than that with the available data. The final position using this correlation is dead center of Rt27 headed for the Pentagon.

[edit on 26-11-2009 by 911files]

reply to post by turbofan


Aren't you talking about actual air pressure versus rate of change of air pressure?

The two are fairly independant as virtually any rate of change is possible at any altitude but I agree that there's some correlation in terms of any measurement error having an effect on both. How much of a correlation and whether it has any bearing on the discussion is in question though. IE the altitude could be in slight error with virtually no effect on the rate of change measurement.

VSI is recorded in the raw data file once per second in feet/minute but neither the NTSB or Warren's decode have extracted this figure but it's one I'm interested in checking. I was even considering re-compiling Warren's code to extract it as he was kind enough to supply the source files.

reply to post by 911files


The impact points on the poles would supply a very exact measurement of the plane's altitude and angle of bank at that instant if the measurement of the poles' remains was actually done. With a major building in ruins and a large number of casualties, general hysteria going on I don't know why anyone would go to such extremes in terms of documenting the very minor collateral damage but it is possible that it was done I suppose.

Are you aware of any such measurements being available?

Originally posted by Pilgrum
reply to post by 911files

 

The impact points on the poles would supply a very exact measurement of the plane's altitude and angle of bank at that instant if the measurement of the poles' remains was actually done. With a major building in ruins and a large number of casualties, general hysteria going on I don't know why anyone would go to such extremes in terms of documenting the very minor collateral damage but it is possible that it was done I suppose.

Are you aware of any such measurements being available?

No sir, I am not aware of any. I am sure the FBI has that forensic information, but they just ain't sharing.

reply to post by Pilgrum

Read the many other threads about the light poles and you will see that there is a disturbing lack of evidence about them.

As far as light pole one is concerned, it is a government myth that it was ever knocked down. You will only find media driven 'information' about the event.

Meanwhile, I patiently wait to see how the alleged FDR data is going to show five distinct light pole strikes with absolute certainty.

Originally posted by tezzajw
As far as light pole one is concerned, it is a government myth that it was ever knocked down. You will only find media driven 'information' about the event.

A cab driver's testimony is a 'government myth'. I find this shocking.

Originally posted by 911files
A cab driver's testimony is a 'government myth'. I find this shocking.

Lloyde has discredited himself as a witness, 911files. I have little doubt that you have watched his interviews with CIT and that you already know this.

Besides, show me one formal interview or transcript for where Lloyde was officially interviewed by the government?????

Anyway, there's other threads about Lloyde's contradictions. Pilgrum can view these threads about Lloyde and the lack of light pole data. There's no need to drag this thread off topic.

Originally posted by tezzajw

Originally posted by 911files
A cab driver's testimony is a 'government myth'. I find this shocking.

Lloyde has discredited himself as a witness, 911files. I have little doubt that you have watched his interviews with CIT and that you already know this.

Besides, show me one formal interview or transcript for where Lloyde was officially interviewed by the government?????

Anyway, there's other threads about Lloyde's contradictions. Pilgrum can view these threads about Lloyde and the lack of light pole data. There's no need to drag this thread off topic.

You are so funny. Thanks for the laughs. Have a great Turkey Day!

Originally posted by PilgrumAren't you talking about actual air pressure versus rate of change of air pressure?

No, not exactly.

Look at it this way, the Pressure Altimeter [PA] reads altitude based on the
plane's vertical position at any given time. If the aircrafts drops altitude,
over that one second of time PA will update with a new value. PA does
not tell you the time required to change altitude (rate of descent/ascent).

This is where VSI comes into play. VSI does not tell you the altitude, but
it tells you the rate of change in feet/min. By looking at VSI, you can calculate
the amount of change in altitude based on the previous scan.

Even though VSI samples the pitot port, the static port is used to determine
the delta in pressure (which is also read by the PA absolute pressure sensor).

As I mentioned before, I would expect to see these values "locked" in
their trends, and therefore would confirm Pressure Altitude whether it
were a 'real' file, or fabricated by flight sim.

Pilots, correct me if I'm wrong, but VSI is only used to tell the pilot how
fast the aircraft is moving up, or down to indicate safe limits of operation.

P.S. John, I know you're not much for the CIT boys, but you should really
watch Lloyd's interview (the most recent) before coming to a conclusion.

[edit on 26-11-2009 by turbofan]

Hey ces,

I ran into your bio online. Nice.

Originally posted by cesura

As for the divergence between pressure altitude and my calculated altitude: Either curve can be moved up or down, without changing its shape, by applying a simple additive correction (which, for the pressure altitude, is needed anyway to adjust for barometric pressure). So the curves can be made to coincide at either end of the graphed interval, with the divergence at the other end.

I think that you'll ultimately find that there is also an approximately 1 second skew in the radar altitude height (delay) with respect to the rest of the data, including the PA. It seems that the RA is (more or less) the last datum to be sampled as the data words are assembled. This means a LEFTWARD shift of PA data by about 1 second is necessary to get the PA to align with the RA.

Furthermore, you can correctly (and justifiably) align the PA ultimate point (-99', IIRC) with the ultimate RA point (4', IIRC), to get a meaningful representation throughout the rest of the curves.

Originally posted by cesura
To tomk52:
The Nyquist criterion is double the lowest frequency that you don't need to see or to distinguish from even lower frequencies.

It appears to me that "the highest frequency that you WANT to see" (my terminology) is just slightly less than "the lowest frequency that you DON'T want to see" (your terminology).

You're suggesting a Nf approximately equal to 2.1x as opposed to my 5x to 8x.

I recognize the theoretical Nyquist criterion. I am neither a mathematician nor a scientist. I'm an engineer. I'll stick with 8x whenever possible. And move carefully down to 5x when forced to.

If you know what you are looking for, (and you can make an argument that you do, with GIF terrain maps), then with some careful (laborious) work, you can cheat down a little further. But I would not trust anyone going down to 2x without understanding these details exquisitely well.

Originally posted by cesura
The cause of the Moire effect in Figure 5 is that our visual systems choose an incorrect low-frequency interpretation for inputs whose correct interpretation involves frequencies above the Nyquist limit.

I disagree.

The Moire effect is registered just fine by film, by cameras, by photoelectric cells, etc. Not only by biological visual systems. Therefore it is not caused by our visual systems. It's an objective phenomenon that is a spatial analog to "beat frequencies" in superposed oscillating systems with slightly different frequencies..

Certainly the spacial distribution of our rods & cones on our retina can also produce interference patterns that register as Moires as well.

[End Part 1]

[Part 2]

Originally posted by cesura
Applying that to the radar height, we might interpret a sawtoothed ridge that generally rises in height as a smooth decline, or we might interpret two distinct buildings as a single building, but we won't see radar heights that aren't really there. Is your point that the geographical elevations used by 911files may contain too much detail, and should have been smeared before they were compared against radar heights for his statistical analysis?

This is comparable to the old bed of sliding nails that you pressed your hand or face into and produced a coarse image on the other side.

If you were to take one of those things that had 50 x 50 nails in it (that's far denser than they were made, of course), and remove all except one row of nails. And then take out 3 out of every 4 of the nails remaining in the last row. Then you were press it at an angle (to represent the descent angle) onto some unknown part of a person's face, and register 12 points off of it.

Two things become obvious.
1. There will be some pattern on the nails that truly reflect the corresponding points on the person's face. So the "heights really are there".
2. Unlike the pattern of 50 x 50 nails which would be instantly recognizable, the 1 x 12 pattern will be pretty indecipherable.

Now your job is to find the line on the person's face that the nails represent.

As I said before, I believe that this can be done, if done carefully and somewhat laboriously. I'd do it by making a whole bunch of "locks", representing the ground terrain maps at various small angles diverging from the line of the known approach of the plane. And one "key", which is the (PA - RA) of the plane. Then I'd physically or mathematically slide the key around in the locks until I found a fit.

This is a fundamentally different form of "pattern recognition" than looking at either the "key" or "lock" alone.

Someone should not expect to look at the data points & think that they can see the ground terrain directly. There is an almost irresistible urge to draw spline curves between the data points. This is misleading.

As I said above, the plane was traveling between 700' & 800' between samplings. By simply looking at the data points, you should expect to see details that have an extent of about 4000'. Nothing much smaller. By this criteria, you can look for geographical large features. You should not expect to see small ones. Especially not buildings.

Give it a go & prove me wrong. I'd love to see it.


Tom

Hope y'all had a good Turkey day.

Originally posted by turbofan
P.S. John, I know you're not much for the CIT boys, but you should really
watch Lloyd's interview (the most recent) before coming to a conclusion.

[edit on 26-11-2009 by turbofan]

I have (all of them). One of the worst cases of eyewitness harrassment and leading I've ever witnessed.

reply to post by tomk52


Tomk52, I listened very carefully to your feedback and I think you misunderstand the concept. The graph is not a 'scatter shot'. It is a plot of two populations, one based on the upper limit of the terrain elevation within an already defined and known location of the plane (not a haphazard sampling). The other population is the lower.

Since each location at an instance in time is known based on radar measurements independent of the FDR, actual terrain elevation MUST be between those boundaries. Not even close to your nail impression analogy.

My goal was to determine what the populations (upper and lower) were doing over time. The real R. Mackey and others have speculated that air flow and air density at the extreme speeds impacted the accuracy of the PA. I have no opinion on that hypothesis except that it seems qute reasonable. If I understand your hypothesis, the PA is 'lagging' in reference to time. Same deal, I don't know enough about the system to have an opinion, but sounds like a reasonable hypothesis to me as well. Seems to me both would produce similar results.

I am not using my 'eye' or anything else to reach my conclusion. That is why you see a calculated linear fit for both populations. Those fit lines are the only thing I am concerned with. The tell me that the difference between the raw PA and calculated elevation (RA + terrain elevation) decreases over the final 18 seconds by ~100 feet. That is my interest, what the data is doing, not why it is doing it. That way I can make some reasonable data based predictions to improve my correlation. If it happens to support yours or Mackey's hypothesis is for others to debate.

My purpose was to improve my time correlation model and resultant positional data. That chart helped me refine my lat/long offsets and gave me a much better fit with the radar data once completed. If you don't find it useful, then I apologize. But I would suggest that instead of nit-picking you come up with a better model to understand what you are asserting, and I certainly would appreciate any data driven assistance. For now though, I don't think the model can be refined much further. Yet I still anxiously await what censura comes up with.

[edit on 26-11-2009 by 911files]

Remember, everyone that 911files posted this comment:

Originally posted by 911files
There is only confusion among CT forum 'experts' who have no clue what they are talking about.

Now consider the above quote, in light of the quotes below:

Originally posted by 911files
Tomk52, I listened very carefully to your feedback and I think you misunderstand the concept.

Not even close to your nail impression analogy.

I have no opinion on that hypothesis except that it seems qute reasonable.

I don't know enough about the system to have an opinion, but sounds like a reasonable hypothesis to me as well.

If it happens to support yours or Mackey's hypothesis is for others to debate.

But I would suggest that instead of nit-picking you come up with a better model.

Yet I still anxiously await what censura comes up with.

I've already admitted that I'm easily confused, 911files. Perhaps it's time for you to admit that it's not just the 'CT' people, reading this thread, who are also confused.

It's ok to admit that it's not just the 'CT' people who are confused, 911files. You can hardly approach this thread stating that only the 'CT' people are confused, when some posts that you type are filled with statements showing how unsure you are, or how unsure someone else is.

I've certainly demonstrated that cesura was confused on how to correctly model the flight path, when he claimed that a single point and a velocity vector would be sufficient - which is patently false. None of you leapt to cesura's defence on that one and he didn't try to defend himself or admit his error.

Collectively, you are all so funny. Thanks for the laughs!!!

EDIT: Have a great Turkey Day!

[edit on 26-11-2009 by tezzajw]

Originally posted by tomk52

I think that you'll ultimately find that there is also an approximately 1 second skew in the radar altitude height (delay) with respect to the rest of the data, including the PA. It seems that the RA is (more or less) the last datum to be sampled as the data words are assembled. This means a LEFTWARD shift of PA data by about 1 second is necessary to get the PA to align with the RA.

I think that you'll ultimately find that you're dead wrong AGAIN and continue
to speak out of ignorance.

Vert Accel: 8 Hz starting at word 2, 10, 18, 26, 34, 42, 50, 58
Pressure Altitude, 1Hz, word 62 [CAPT], 63 [F/O]
Radio Height , 1 Hz, word 55, in each subframe
Vert Speed, 1Hz, word 64

256 Words per frame
125 milliseconds between Vert. accel captures.

Hardly 800' between vert. positional data!

[edit on 26-11-2009 by turbofan]

reply to post by 911files


What do you mean by leading? Lloyd said his car was never on the
bridge even though the photos proved otherwise.

Originally posted by tezzajw
It's ok to admit that it's not just the 'CT' people who are confused, 911files.

See there ya go getting all confused again. There is a difference between serious debate and discussion of empirical data results and 'confusion'. TomK raises some serious issues and may have thought I was taking sides with one hypothesis over another. I understand both, so no confusion going on here.

As censura pointed out, we are working the probem from two directions with different goals in mind. Most people in this thread are interested in the last dozen or so seconds in the flight path. My interest is the entire flight path. I started my correlation at Dulles (IAD) at take-off based on M3 altitude values. I continued until there was no beacon data available. The correlation done then, could be applied to the IAD data once the plane was reacquired. The one constant among every site and data source was a clock. Each clock is offset (just like it is rare to have your watch, bedroom and wall clock exactly set together), but the underying assumption is that each measures one second as essentially the same unit. Early on, I found that was not the case with the NEADS data. So I looked for a standard to align all of the many clocks. IAD was the closest to real-time, so that is the standard I selected.

Now that we are at this stage, there is a new data source. Refining the model generated to-date I soon found two problems. The first is that I only sought +/- 2 seconds accuracy in my earlier efforts (a rough model). Since I settled on RO time = IAD time for my rough model which gave a good fit for my roughing, I found it no longer worked upon fine tuning at the EOF. So, I used IAD + 1 second and it resulted in a slightly better fit at the beginning of flight and significantly improved match with the DCA/ADW data.

Then, the INS systmen drifted during the 330 degree turn. So I started with positional offsets that improved the model further. Both of these helped me refine the model to +/- 0.5 seconds. That however did not address the at what point in the RO file did the pane encounter the poles question (something you and others desperately wanted to know). That is what the RA + terrain modeling was for. It revealed one point that could be used to refine my positional offsets. Now, that is just pulling something out of the hat if it does not result in improvements when compared to other data sources. It did, improving the EOF accuracy for the final 18 seconds to +/- 0.25 seconds.

So no, that is not 'confusion', that is working your way through the problem.

Originally posted by turbofan
reply to post by 911files

 

What do you mean by leading? Lloyd said his car was never on the
bridge even though the photos proved otherwise.

Okay, since you asked and I guess it is somewhat relevant to the OP (not to mention I am bored), I will answer. Lets apply your standard and discredit a few more eyewitnesses to the pole area.

SGT Lagasse - He said the poles were downed at the same intersection CIT convinced Lloyd he had to be on. He said he was at a certain pump, but changed his testimony when Craige challenged him with the Citgo video. He was leaning in his car talking to his K-9 when the plane passed with his back to the area he claims to have seen a plane, but in his words, he doesn't have "eyes in the back of his head", so he could not have seen it. Discredited.

Robert Turcios - He too was at a different pump than he said he was. He also could not see the location he said the plane was from where he was. Not only that, he said he ran up the hill for a better look, but in reality he ran into the station scared witless. Discredited.

SGT Brooks - He claimed the plane was a United plane. In his earlier recorded statement he said he saw the poles hit. Discredited.

Ed Paik - He said the plane hit the VSP tower and bent an antenna which was replaced the next day. He went into detail about how it was bent over 90 degrees and saw it. His brother backs up his account. Consider them both discredited.

Terry Morin - He claims he was between the annex wings, but in 2001 in his own words he says he had just walked out from between them and turned towards the guard shack. Discredited.

Sean Boger He said the plane went NoC. But in 2001 he said he saw the plane enter the building and heard the sound of metal crushing. Discredited.

I could go on....

Come on John, I thought we were on a course on honest research?

Let me correct and clarify your points below:

Originally posted by 911filesHe said the poles were downed at the same intersection CIT convinced Lloyd he had to be on.

CIT did not convince Lloyd of anything. Lloyd denies being on the bridge
with his cab even though the photos prove it.

SGT Lagasse He said he was at a certain pump, but changed his testimony when Craige challenged him with the Citgo video. He was leaning in his car talking to his K-9 when the plane passed with his back to the area he claims to have seen a plane, but in his words, he doesn't have "eyes in the back of his head", so he could not have seen it. Discredited.

Lagasse is on video stating he jumped in the car after the plane flew NoC
to call for help. He did not say he was in the car talking to his K-9.
His comment about "eyes in the back of his head" was his response to
Craig when asked if Lagasse was mistaken about the path of the aircraft.
Lagasse swears on his life that the plane flew NoC and could not have
seen the plane on the south side as "he does not have eyes in the back of his head.

This is on video, are you sure you watched it?

Robert Turcios - He too was at a different pump than he said he was. He also could not see the location he said the plane was from where he was. Not only that, he said he ran up the hill for a better look, but in reality he ran into the station scared witless. Discredited.

Can you provide a source for your claims John? The video I saw shows
Robert stating he saw the plane fly North of the north side of the canopy,
and Robert also stated he saw the aircraft "pick up" over the street sign.

SGT Brooks - He claimed the plane was a United plane. In his earlier recorded statement he said he saw the poles hit. Discredited.

I'd be interested in hearing this interview. Where can I find it?

Ed Paik - He said the plane hit the VSP tower and bent an antenna which was replaced the next day. He went into detail about how it was bent over 90 degrees and saw it. His brother backs up his account. Consider them both discredited.

Please provide a source for this interview. A few pages back I posted a
link to this on the CIT site and what you say is not accurate at all.

Terry Morin - He claims he was between the annex wings, but in 2001 in his own words he says he had just walked out from between them and turned towards the guard shack. Discredited.

Morin's recorded audio states he was inside the wings, and he also states
the craft was not moving 460 knots.

Boger He said the plane went NoC. But in 2001 he said he saw the plane enter the building and heard the sound of metal crushing. Discredited.

Sean saw the plane banking over the Annex and NoC. This contradicts
the official story and the FDR data. There are also several witnesses that
support his NoC account.

You should really watch the video testimony again and refresh your memory.

Now back tot he FDR, what system on the aircraft might have produced
the white smoke? Do you see a reflection of this smoke production in
the data? If so, where?

What are your measurements for the wings of AA77 hitting the camera mast,
and pole #2?



[edit on 27-11-2009 by turbofan]