AA77 FDR Data, Explained

[wstutt]

Hi apathoid,

Thanks for your post #4078. Your reasoning is clear to me. I asked Rob what he thought of your post and he replied asking whether you had reviewed his earlier post. I know you are now saying that proving lag is unnecessary. I'll be telling him that.

Thanks again,
Warren.

 

[stewieg]

And what makes this turd *really* shine is that, without taking a breath, he states that the data he bases his above statement on is invalid because

Quote:
"...even if we obtain the Maintenance Manual from American, we must first determine the data is from an American Airlines aircraft, N644AA. This is why i have repeatedly stated this data is not "proof" of anything.



Proof positive that a box of hammers is smarter than Cap't Bob Balsamo.

That can't be to hard for him to find out can it? He has an AA mechanic on his membership list. Or is he a figment of FO Balsamos imagination? If he's real maybe Rob could move in with him!

 

[wstutt]

I now have version 1.5 of my AAL77 FDR Decoder available on my web site along with new output files.

I have added an option to replace blank (empty) values with the word null.

I have also added some new parameters.

You can read further notes on the new parameters here.

Warren.

 

[alexi_drago]

I've just been having a look through the boeing manual from warrens site and noticed one or two things, maybe they've already been mentioned.

I seem to remember Cap'n Bob trying to justify the cockpit door parameter in the data frame by saying the data frame was customised for the airline because UA had a different layout.

Section 3.xx in the manual shows that the DFDAU detects what plane it's in and selects the appropriate data frame layout automatically. So even with different frame layouts it could be the same aquisition and recording system in both UA and AA. This also applies to the aircraft model.

The next thing, I'm not so sure about.

from 1.5 in the manual:

"shall" is used to express a requirement that is binding

from 10.1.1 in the manual:

All unused data bits shall be padded zero in the data frames

But it doesn't specify whether that only applies to the fields that are labeled as 'unused' or if it also applies to labeled fields that are unused.

 

[apathoid]

Hi apathoid,

Thanks for your post #4078. Your reasoning is clear to me. I asked Rob what he thought of your post and he replied asking whether you had reviewed his earlier post. I know you are now saying that proving lag is unnecessary. I'll be telling him that.

Thanks again,
Warren.

Hi Warren,

Thanks for finding the Total Pressure parameter. Indeed, it's quite significant and I doubt any 757 before or since 9/11 has ever seen 42.5 in/Hg impact pressure.

I can see nothing in Rob's earlier argument that refutes anything in my post 4078. If he wants to refute the facts, he'll have to start providing evidence to support his claims, as I have. That means no more suppositions, no more rhetoric, no more speculation, technobabble, or outright lies. I'm not interested in Rob Balsamo's amateur opinion on this matter.

If he wants to claim the air data system is accurate 41% past Vmo, far in excess of the 757 operating envelope(so extreme, that damage could be occurring to the Air Data Computer), he'll have to support that claim with data that relates to the 757's air data certification criteria.

I've showed the calibration range, which doesn't include high-speed/low altitude flight. If he wants to claim the 757 air data system is accurate in such profiles, even within the 757's flight envelope, he'll need to show that.

I've shown that the allowed calibration error doesn't include the aerodynamic effects of flight, rendering his "the error can only be (x) feet at 0' msl" argument completely void. AA77 was in flight, not sitting on the ground undergoing a pitot-static system test. If he wants to claim that the error pertains to an aircraft in flight; he'll need to show that.

 

[apathoid]

I've just been having a look through the boeing manual from warrens site and noticed one or two things, maybe they've already been mentioned.

I seem to remember Cap'n Bob trying to justify the cockpit door parameter in the data frame by saying the data frame was customised for the airline because UA had a different layout.

Section 3.xx in the manual shows that the DFDAU detects what plane it's in and selects the appropriate data frame layout automatically. So even with different frame layouts it could be the same aquisition and recording system in both UA and AA. This also applies to the aircraft model.

Yup, I asked Warren to include B117 and B118 in his readout as to confirm what DFL needed to be used. What's funny about Rob's argument that the 757-3B_1 layout was specific to American's fleet, is that it included all Pratt & Whitney specific engine parameters(AA77 was equipped with Rolls Royce), which proves the layout has parameters which dont apply to N644AA and needed to be discarded. That little finding destroyed his notion of "if its in the custom DFL, it's a working parameter".

But it doesn't specify whether that only applies to the fields that are labeled as 'unused' or if it also applies to labeled fields that are unused.

That's the thing about these generic DFL's. They kind of assume that if a certain Flight Data Acquisition Unit/FDR combination is installed - the airplane will be capable of (x) number of parameters. It doesn't do any operation to figure out which ones are there and which ones aren't. This is undoubtedly done manually by NTSB technicians, as evidenced by the "group of parameters not working or unconfirmed" in the AA77 FDR Investigation.

Since the "FLT DECK DOOR" parameter piggy backed on a frame which did actually have other parameters working, the software couldn't discriminate against that particular frame and lump it into an "unused" category.

 

[alexi_drago]

....
That's the thing about these generic DFL's. They kind of assume that if a certain Flight Data Acquisition Unit/FDR combination is installed - the airplane will be capable of (x) number of parameters. It doesn't do any operation to figure out which ones are there and which ones aren't. This is undoubtedly done manually by NTSB technicians, as evidenced by the "group of parameters not working or unconfirmed" in the AA77 FDR Investigation.
....

This seems to be something Rob can't quite get his head around and suggests that if that was the case then there would have to be some form of guesswork involved in knowing which of the parameters are valid if they are recorded as 0 for the entire flight.

 

[tfk]

Hey Warren,

Would you pass a message over to Robbie for me. (Since he is too terrified to allow me to post on his web site. LoL.)

Tell him that I had a great laugh reading about his, uh, "unique" interpretation of what happens to aneroids if you over pressure them.

Tell him his comments on "RUAC" made his technobabble all the sweeter.

I really enjoyed this paragraph:

This is the difference of those who change parts for a living (crapathoid), and those who actually have to know the system because their ass depends on it. This is why pilots train on partial panel work for their instrument rating (what happens if the pitot tube is clogged? Static port clogged? Both? etc..). And then further learn the systems more thoroughly when they become an Instrument Instructor (as I am trained and trained others).

The strutting peacock is especially ironic since just about everything he says about this system "that he has to know because his life depends on it" is utterly wrong.



Tell him he hasn't a clue about failure modes in aneroid capsules. And that his entire diatribe about the consequences of exceeding the max delta P on the airspeed capsule was a load of BS.

Sounded great for his puppies, of course.

I guess his life didn't depend on his knowledge quite as much as he'd like to bluster...

To make the point, ask him:

1. What gas do they use to check for leaks in the capsule welds?
2. What gas do they use to test for the burst pressure of the capsule?
3. What is the typical burst pressure of a capsule?

Let me know when/if he posts an answer, wouldja?

Thanks,

Tom

 

[apathoid]

Warren, any chance you can add the "STATIC PRESSURE" parameter? It might be helpful to know what the raw, uncorrected altitude is before the ADC does its conditioning. Thanks in advance.

 

[tfk]

Hey Ap,

Nice post. Here are my thoughts.

The excessive delta P condition (Pt - Ps) that you've noted has to apply to the airspeed diaphragm, not the PA one. The PA only gets the static pressure, not the Pitot pressure. And the PA aneroid is designed to see over 31" Hg. So it is not out of its absolute range of operation.

You have discovered source for error in the calibrated air speed measuremet. And possible damage to its aneroid.

And therein lies - indirectly - the source of the error in the PA reading: What we've been saying all along. It appears that the ADC takes the measured static pressure & makes empirical (flight test derived) corrections to that measurement in order to calcuate the recorded PA reading. The corrections are based upon airspeed & AoA.

Diagrammatically, you'd have: Static Pressure Reading + AoA corrections + Airspeed correction => PA Reading.

But here comes the complication: the corrections themselves are going to be different at different altitudes.

As you've shown with some of your data, when you characterize a plane's performance, you can't afford to test performance at every possible flight configuration. So you pick a group of typical configurations, plus you make sure that you get at least one at the outside edges of the performance envelope. This gives you what is called a "sparse data set". A table with only a few points in it (like your altitude & air speed table).

But you don't want your corrections to be applied as sudden "step functions" as you move between flight configurations. So you create "interpolating functions" that change smoothly between your calibration points.

I've got three points to make here:

Point 1.
When your corrections INTERpolate between known test points, they are typically quite accurate. When you EXTRApolate outside of your known test points, these interpolation functions typically blow up & give you large errors. This is one of the specific reasons why you try to get data points on the edge of your flight envelope. So that your instrument corrections are interpolations, not extrapolations.

Here's a good example, from other 9/11 data.

This is the graph that NIST provided for the speed of the fall of WTC7 vs time. Note that we're discussing the math of interpolating functions, so what the data represents doesn't matter.

On this graph, we've got two interpolating functions:

a) the linear one (red line) that is calculated based upon only the points within Stage 2
b) the exponential one (black line) that is calculated based upon the points within Stages 1 & 2.

You can see immediately that, within the zones for which they were calculated, both functions are pretty good. But outside of those zones, both functions diverge dramatically from the experimental data.

As you've been saying, the airspeed of AA77 at sea level was clearly an extrapolation beyond the intended performance envelope of the plane. Therefore the CORRECTION that the ADC is going to make to the calculated (i.e., recorded) PA is going to have sizable errors to it. Errors that are well beyond anything that any regulations or manuals require.

Point 2.
The airspeed corrections (the interpolating functions) are DIFFERENT at different altitudes. You cannot simply say that, since 0.7M was within the planes performance, therefore the airspeed corrections to PA will be accurate. There is an additional parameter for the correction that you'd be ignoring: the altitude at which the airspeed corrections were calculated.

Point 3.
Balsamo demonstrates an inordinate lack of knowledge (and judgment) when he comments that:

In order for those who blindly support the govt story to hold onto their theory, they must prove the PA was 120 feet in error, they must prove .70M-.72M is above Mcrit for the 757 and as such, the compressibility issues cause an artificially high PA reading, they must prove the "+/- error" is in their favor, they must prove the Pitot-Static system was "damaged" to their favor, and finally, they must prove the data came from N644AA.

1. "We" don't have to prove anything about the PA. The plane flew into the Pentagon. Nothing flew over the building. End of story.

2. Compressibility does not start at Mcrit. It begins to have effects starting around 0.3M and increases with increasing speed. Refer to any fundamental aerodynamics text.

3. His postulate "... as such, the compressibility issues ..." is simply wrong. Compressibility is not the only source for errors in PA readings. Compressibility is not calculated or measured as a correction parameter. Rather, it is embedded in the airspeed corrections.

4. There is no requirement to do any damage to the pitot-static system in order for the calculated PA to be erroneous. Since the plane was operating well outside of the MULTIDIMENSIONAL (not one dimensional, i.e. airspeed) flight envelope.


Tom

 

[tfk]

Warren, any chance you can add the "STATIC PRESSURE" parameter? It might be helpful to know what the raw, uncorrected altitude is before the ADC does its conditioning. Thanks in advance.

Bingo!

Excellent idea, if it is available. It'll put hard numbers to all sorts of effects that we've been talking about qualitatively.

tom

 

[tfk]

Warren,

No need for you to carry any message over to Robbie & P4T for me. Sorry to ask that of you. Not very cricket. I won't do that again.

Regardless, it was unnecessary. Robbie apparently monitors JREF pretty closely. After I'd posted mine, his reply appeared in no time.

I'd love to hear if your data has raw, uncorrected static pressure, as apathoid requested, tho.

Regards,


Tom


PS: Here was Rob's post to me, and my reply...

I figure it'll stay up about 10 seconds beyond the time that he sees it over there.

LoL, such a little boy...
___

Robbie,

I took a stroll into Randiland before calling it a night. Found that "tfk/tom" still doesnt realize he can post here. We have told him many times he is not banned and can still sign on to post in this debate forum (he has been restricted to this section due to his many immature personal attacks you see here and elsewhere), but he can still post here.

You're such a fibbing little boy, Rob. And a coward, to boot.

Your history betrays you, Rob.

You ban everyone that disagrees with you. You know it. I know it. EVERYONE knows it.

I don't even need to pull up the posts that detailed your "suspension for lying" (great euphemism for "someone who disagrees with you") within about 3 days of joining, because I brought up the fact that an Aussie pilot put someone into a 757 simulator & was able to perform the maneuvers at the WTC & Pentagon that you claimed were "impossible".

And then your acknowledged tracking of my movements around your board. Paranoid much?

And finally, your fevered banning of my account when we disagreed previously about aneroids & procedures that need to be followed for meeting 91.312.

You admitted that you were wrong about the aneroids.
When I PROVED that you were wrong on all your babbling regarding 91.312 (http://www.abovetopsecret.com/forum/thread512723/pg46 my post @ 12:08 pm), you simply slunk away. How typically cowardly, Rob.

With that said, Tom wanted Warren to pass on these questions to me regarding my post above....

"1. What gas do they use to check for leaks in the capsule welds?
2. What gas do they use to test for the burst pressure of the capsule?
3. What is the typical burst pressure of a capsule?"
​

My reply...

1. Irrelevant
2. Irrelevant
3. Somewhat relevant. Perhaps you may want to inform crapathiod that its not exactly at 10 inHg differential?

Yeah, didn't think you knew any of that. Didn't stop you blowing smoke about aneroid failures tho.

1. Helium. Just like they do with all vacuum systems. Know why they use helium, Rob?

2. Hydrogen hydroxide. At room temp. Do you know why they use this deadly compound (tens of thousands of people die from it every year) to do this test, Rob?

3. The burst pressure is 10 - 50 times greater than the pressure it takes to cause the capsule to fail. That's because you don't know the failure mode of these capsules, Rob.

Care to take another guess?

We know "tom" (who also refuses to give his real name which can perhaps be verified to name quite a few skeletons), likes to also throw crap at the wall to see if it sticks.

Sure thing, Robbie. You're a real "respectable" kinda guy.

One that speculates on "quite a few skeletons" in the closet of someone you don't know from Adam. The ONLY thing that you know about me is that I disagree with you about 9/11. And that I'm an engineer, and laugh at your attempts to do math.

Now Tom, if you have the gonads to log on here (and yes you can post here), please tell us:

Oh, I might have the gonads...

1. What type of damage is done to the "Pitot-Static System" as referred to in the documents provided by crapathiod, if differential pressure between Pt and Ps is too great?

This IS the failure mode of a traditional aneroid capsules from overpressure: you'll irreversibly deform one of the convolutions. Put enough pressure on it & you can invert one or more of the convolutions. Just like when you freeze a plastic bottle of milk. They have the circular indents on the side so that when the milk expands, it pushes out the indent, and doesn't burst the container.

Too much pressure & you'll also likely damage some of the diaphragm follower mechanisms, too. In fact, the aneroids are pretty rugged. This damage to the follower mechanism may well happen before you damage the aneroids themselves.

2. What indications would you expect to see with in-flight differential pressures 2 inHg above the Bench test limits?

Nothing will change on any of the static port systems, because this pressure isn't sufficient to burst a good diaphragm. The airspeed will go off the scale (way too high), tho.

3. What do you think crapathoid saw when he busted the ADC due to over-pressure at Pt?

I haven't seen his description of what he did before it broke.

What do YOU think that he saw?

Sign on anytime to answer as i rarely check your cesspool.

Sure thing.

Unfortunately, you history proves that you won't be able to tolerate your sycophants seeing anyone disagree with you. And that you'll undoubtably ban me again.

You've become way too predictable, Robby.

You've become the farcical caricature of the snake-oil salesman, constantly afraid of exposure.

Always nice chatting with you.

Toodles,

Tom

 

[stewieg]

Warren,

Regardless, it was unnecessary. Robbie apparently monitors JREF pretty closely. After I'd posted mine, his reply appeared in no time.

He's probably reading JREF right now. When I mentioned last month that it had been three days since anyone posted anything on the PfT Flight Deck Door thread long behold he posted on it within 24 hours.

Hey Rob, it's been four days since anyone posted on your flight deck door thread. You might want to refresh it a bit. Maybe your groupes are getting bored with it. One thing is for sure, they're a bunch of cheap bastards. $275 out of $5000 for 1qt2010 research and operations, cmon!

 

[tfk]

Just noticed something kinda funny...

At PfffT, Robbie's has set up his server to automatically change the word "JREF" into "the government loyalists site".

Perhaps we should respond in kind...?

I'm open to suggestions.

How about:
"the Union of Terrorist Apologizers"?
Or "The Traitors Collective"?
Or "The Collection of Once Honorable Military Veterans who Used To Support & Defend the United States of America, but Have Decided to Turn Their Backs on Their Country"?

Naaahhhhh...

That wouldn't be very mature.

Just because they have decided to accuse their own government of mass murder & treason. And their country's law enforcement agencies of colluding with & covering up mass murder. And their country's military officers & enlisted men & women of colluding with & covering up mass murder & treason ...

Nah, it'd be way too adolescent & venial to do something THAT vindictive.

We should leave that behavior to Balsamo.

Sad.

Tom

 

[stewieg]

How about "PfT, which is really just Rob Balsamo looking for a handout from all the chumps out there". You're right though, lets leave this sort of stuff to Rob.

 

[wstutt]

Hi apathoid and tfk,

Warren, any chance you can add the "STATIC PRESSURE" parameter? It might be helpful to know what the raw, uncorrected altitude is before the ADC does its conditioning. Thanks in advance.

Unfortunately, I tried decoding the STATIC PRESSURE parameter before and did not get sensible values so I removed it from the program.

Warren.

 

[DGM]

Hi apathoid and tfk,



Unfortunately, I tried decoding the STATIC PRESSURE parameter before and did not get sensible values so I removed it from the program.

Warren.

[truther] COVER-UP............INSIDE JOB.............[/truther]



Sorry couldn't resist.

 

[DavidJames]

How about "PfT, which is really just Rob Balsamo looking for a handout from all the chumps out there". You're right though, lets leave this sort of stuff to Rob.

How about ignoring them and watch them fade into oblivion rather then providing them a popular forum to feed their ego's. At least until they are able to produce an intelligent argument.

 

[apathoid]

Hi apathoid and tfk,



Unfortunately, I tried decoding the STATIC PRESSURE parameter before and did not get sensible values so I removed it from the program.

Warren.

That's kinda what I figured, but it couldn't hurt to ask. Thanks.

@Tom:

Pretty predictable response(s) from Rob. Apparently we struck a nerve, I can practically picture him frothing at the mouth as he makes his keystrokes. Poor Robby doesn't like that we've supported our arguments that the air data system wasn't calibrated to be accurate in that environment. He babbles aimlessly while we provide hard evidence; I'd be pissed too.

.....(good stuff)....

The excessive delta P condition (Pt - Ps) that you've noted has to apply to the airspeed diaphragm, not the PA one. The PA only gets the static pressure, not the Pitot pressure. And the PA aneroid is designed to see over 31" Hg. So it is not out of its absolute range of operation.

You have discovered source for error in the calibrated air speed measuremet. And possible damage to its aneroid.

And therein lies - indirectly - the source of the error in the PA reading: What we've been saying all along. It appears that the ADC takes the measured static pressure & makes empirical (flight test derived) corrections to that measurement in order to calcuate the recorded PA reading. The corrections are based upon airspeed & AoA.

...(good stuff)......

Right, that's really the point but I didn't articulate it properly...the ADC is looking at airspeed to refine the altitude data. So failure in the Pt sensor can(will?) lead to an error in computed altitude. However, my ultimate point was that since the air data system was beyond damage threshold, it was most certainly beyond its calibration range.

As you've shown with some of your data, when you characterize a plane's performance, you can't afford to test performance at every possible flight configuration. So you pick a group of typical configurations, plus you make sure that you get at least one at the outside edges of the performance envelope. This gives you what is called a "sparse data set". A table with only a few points in it (like your altitude & air speed table).

Great point. The final pressure station(40,000, Mach 0.85) does exceed VMO, albeit by a single knot. For those who may have missed it.. AA77 was exceeding VMO by 141 kts!

Dont get me wrong, there are many Avionics Techs out there who are excellent. We have quite a few who are in our organization. Crapthoid isnt worth his weight in salt and i'm surprised he still has a job at Delta after damaging an ADC/instruments due to his incompetence.

(on the ADC I messed up)I haven't seen his description of what he did before it broke.

What do YOU think that he saw?

It was many years ago, when I was still wet behind the ears. We were in the middle of a pitot-static full range leak check on a 727 when another mechanic approached me about losing all the lights inside the plane(ones plugged in from a ground receptacle we call a "doghouse"). Without thinking, I thought 'I'll just reset the breaker', but my air data cart was plugged into the same doghouse. I reset the breaker; the pumps in the air data cart shut off and 30,000' or so and 300+ plus knots of pressure suddenly vented to ambient. Luckily, I was only testing one side, so the offside ADC, standby gauges and elevator feel computers all survived. The ADC side under test had its bellows blown, so the subsequent leak test failed miserably. Lesson learned. And no, I didn't get into any trouble. If you ain't breaking stuff from time to time, you ain't working! I had another incident with a APU battery charger that I'd rather not re-live though.

btw, that's a heck of an endorsement coming from Cap'n "what is ACARS?" Bob.

 

[tfk]

Hey Ap,

It was many years ago, when I was still wet behind the ears. We were in the middle of a pitot-static full range leak check on a 727 when another mechanic approached me about losing all the lights inside the plane(ones plugged in from a ground receptacle we call a "doghouse"). Without thinking, I thought 'I'll just reset the breaker', but my air data cart was plugged into the same doghouse. I reset the breaker; the pumps in the air data cart shut off and 30,000' or so and 300+ plus knots of pressure suddenly vented to ambient. Luckily, I was only testing one side, so the offside ADC, standby gauges and elevator feel computers all survived. The ADC side under test had its bellows blown, so the subsequent leak test failed miserably. Lesson learned. And no, I didn't get into any trouble. If you ain't breaking stuff from time to time, you ain't working! I had another incident with a APU battery charger that I'd rather not re-live though.

This is an interesting failure...

I believe completely the sequence of events that you describe. But the failure mode isn't immediately obvious.

The aneroid in the ADC is a static pressure only aneroid. Zero psia internal and 14.7 psia absolute at 0' elevation, reducing to roughly 1/3rd atmosphere at 30,000 feet (about 5 psia). So you have your maximum differential when it is at 30,000' and as you release the vacuum, the delta p increases. But not outside of its normal operating range. Which shouldn't put any sort of excessive delta p on the unit. Even accidentally pressurizing the static lines puts a positive pressure (outside vs. inside) which won't pop an aneroid.

The same thing happens with the air speed aneroid. Pitot pressure (positive gage pressure) to the inside of the diaphragm and static pressure (negative gage pressure) to the outside. At your test condition, the delta P was at its maximum. As the airspeed (alone) drops the delta p drops. As the "elevation drops" suddenly to zero, the static pressure increases (vacuum decreases), and again, the delta p decreases. Still no excessive delta P as a result of the change, or any combination of rates of change.

There is one aneroid that I can see being over-stressed by this action: the vertical speed aneroid. It gets pure static pressure (a vacuum) to the inside of the capsule, and static pressure (again, a vacuum) thru a bleed pin-hole to the outside of the capsule. Working in absolute pressures, at the start of your sudden shut down, you've got about 5 psia on both the inside & outside. When you release the static pressure, you suddenly get 14.7 psia internal & 5 psia external. This give an excessive (for the VSI) delta P of about 10 psi. Let's say that it takes 1 second for the pressure to equilibrate. That's equivalent to (30,000'/sec * 60 sec/min =) 1.8 million fpm vertical airspeed. Going out on a limb here, I'd wager that was beyond the expected limits of that device.

The strange thing is that you shouldn't record a leak in the static pressure line with a burst VSI capsule, since you're just connecting static pressure to static pressure. You'd actually increase the responsiveness of the whole system, since the internal pressure in the VSI case can now bleed out thru both the capsule and the bleed hole, instead of just thru the bleed hole (as it normally does).

I don't believe that there is a VSI or air speed aneroid in the ADCs that I've seen for the 757s tho.

Bit of a mystery. We may never know.


Tom