A Critical Analysis of JDX's math (split)

[qarnos]

I think an accelerometer will be reading zero in normal gravity. Of course, it is really in 1G, but this will be its reference. So an accelerometer will show the differential G value.

My problem with this is that, in order to cancel out gravitational acceleration, an accelerometer would need to know its orientation. This would require gyroscopes to be built in to the accelerometer and assumptions to be made.

For example, if your accelerometer is assuming 1g of downward force is gravity, then rotating it 90 degrees clockwise would result in -1g downward and +1g to the right.

There is a technical reason for this. The accelerometer has to be set to zero, somehow. The transducer types used for this are not capable of holding an initial setting indefinitely, so you would either have to take them for frequent recalibrations or you set them to zero when there is no differential force.

Well, I am no expert in these matters. I am doing this through thought experiment.

What the reading in an FDR is, I don't know. It could have one G added, but I somehow doubt that. It wouldn't be of any use, since what we are interested in is the differential G.

Hans

The problem is, differential from what?? Acceleration is relative. There is no way you can construct an experiment to determine if you are being accelerated by say, a rocket, or if you are stationary in a gravitational field.

I see no way around this.

 

[Anti-sophist]

Right, if you were take a 3-axis accelerometer and hold it in your hand, on the ground, and read the actual readings, you'd see they all read 0. When they paint on the dial, they literally paint on a '1' for the z-axis to put it in terms of G-Force. This is done because 'G-Force' is really a measure of how much NORMAL force you feel relative to your own weight, which is an estimation of "apparent relative gravity".

I've been very careful about using 'g' versus 'G'. If you need +1g of acceleration (heading up), you are said to be pulling a "2G" move, since it requires 2*mg force to overcome gravity, and accelerate upwards at 1g. Your chair will be pushing on you at a 2*mg force, and thus everything will feel twice as heavy as normal.

 

[MRC_Hans]

My problem with this is that, in order to cancel out gravitational acceleration, an accelerometer would need to know its orientation. This would require gyroscopes to be built in to the accelerometer and assumptions to be made.

For example, if your accelerometer is assuming 1g of downward force is gravity, then rotating it 90 degrees clockwise would result in -1g downward and +1g to the right.

No, that's not how it works. One way of building an accelerometer is to suspend a weight on springs (in three axes), and place strain-gauges on the springs. There are other types, but they are mostly for dynamic uses, and their lower frequency limit is too high for aircraft use.

Now, a strain-gauge is not that good at long-term stability, so you have to zero it now and then (like when you use your electronic bath-room scale, you first dap it with a foot, it turns on and sets itself to zero, then you can step on it. It uses strain-gauges, too). So when you set it to zero, that is your reference, period. In a plane, you will do that while parked, and yes, the accelerometer will assume that the plane is sitting on its wheels, so down is indded down. If you turn it on its side, you turn the plane with it (since it's bolted on somewhere), and it will correctly read -1G down and +1G to the right (G, not g, as A-S notes), which will also be what the passengers perceive.

Well, I am no expert in these matters. I am doing this through thought experiment.

Fair enough, and important for interpreting the FDR data.

The problem is, differential from what?? Acceleration is relative. There is no way you can construct an experiment to determine if you are being accelerated by say, a rocket, or if you are stationary in a gravitational field.

I see no way around this.

Well, I suppose that if you are sitting in a parked aircraft, and about to reset your accelerometer, you can safely assume that you are stationary in a gravity field, and that the gray stuff outside is in the direction you will, for the remainder of the trip, consider as down.

.... Excuse the irony . The problem is, we are mixing an entirely valid theoretical discussion with the, equally valid, pragmatic discussion about what the FDR from flight 77 reads out.

Hans

 

[CurtC]

Guys, just look at the vertical acceleration data in the CSV file (available at 911myths.com). The acceleration values are usually right around +1, but go as high as +1.7 and as low as +0.6. It's obviously calibrated such that sitting still in the hangar it's set to +1.00, not 0.00.

The fact that weedwacker didn't realize that G's are acceleration (not force) doesn't indicate that he's not a pilot, it just indicates that pilots are not physicists. Pilots often have an intuitive grasp of a lot of the concepts without the formal stuff.

We know the plane was close to level as it came across the Pentagon's lawn, because it hit the light poles hundreds of feet away. However, the light poles were slightly elevated above the Pentagon's lawn. I think it's reasonable to say that the plane droped 30 or so feet between the light poles and the wall, which at that speed is still a vert speed of -2000 fpm!

I'm also a little confused about why the FDR data shows Gs less than one in the final second of recording, when I would expect 1 or more. Maybe that data was not for the last second of flight, but was the second previous to that? I notice that the couple of seconds prior were about 1.7 Gs, which could indicate that Hanjour pulled out of a descent just before the intersection in order to clear it, then over the intersection was beginning to resume the dive, so had Gs less than one. Is that reasonable?

 

[Anti-sophist]

There are a few other explainations, as well, but yours sounds plausible. I'm not an expert on how to interpret this data, but simple things can effect G-forces.

For instance, we know during the last second the plane was rolling, and angular velocity will show up as an acceleration in an accelerometer, especially if the accelerometer is not located upon the axis of rotation. As such, if you were in a plane that was doing a roll, you'd feel outward (negative G) accelerations.

(This is a perfect example of why expert analysis is needed).

 

[R.Mackey]

Correction there. The accelerometer will read +1 in normal, straight & level flight. If it reads zero, then you are actually in freefall.

Here's a piccie of one:

Notice how it is reading 1 in 'normal' condition, and that marking is thicker than the other lines?

That's why the DFDR data will show '1' vertical acceleration a lot of the time.

And here we are in the harsh light of morning.

The accelerometers I work with are the ones used for vibration monitoring, not the cockpit display, so your picture and your explanation may well be correct.

However, if the cockpit display is reading 1 G when the aircraft is parked, it's because it's been calibrated that way.

You need to consider the total force on the aircraft. When the aircraft is parked, you have 1 G of gravity, true. But you also have 1 G of force applied through the structure.

Explained more simply, if the accelerometer is sitting on the ground, it is not accelerating. Therefore it should report ZERO.

But again, if the aircraft display is calibrated to read 1 G when stationary, then your answers are correct. I don't know whether they do this or not, and I'll bow to others' knowledge here.

No, that's not how it works. One way of building an accelerometer is to suspend a weight on springs (in three axes), and place strain-gauges on the springs. There are other types, but they are mostly for dynamic uses, and their lower frequency limit is too high for aircraft use.

Thank you Hans, that was the ingredient I was missing. The accelerometer measures force in a way that explicitly discounts the counter-force transmitted through structure, by using a calibrated spring. Not the same as a true measure of acceleration, but more relevant to the pilot's actual problem.

I stand corrected. In this case it does make sense for it to read 1 G when parked. It's just not an "accelerometer" in the Newtonian sense. It's a "G-Meter."

 

[rwguinn]

And here we are in the harsh light of morning.

The accelerometers I work with are the ones used for vibration monitoring, not the cockpit display, so your picture and your explanation may well be correct.

However, if the cockpit display is reading 1 G when the aircraft is parked, it's because it's been calibrated that way.

You need to consider the total force on the aircraft. When the aircraft is parked, you have 1 G of gravity, true. But you also have 1 G of force applied through the structure.

Explained more simply, if the accelerometer is sitting on the ground, it is not accelerating. Therefore it should report ZERO.

But again, if the aircraft display is calibrated to read 1 G when stationary, then your answers are correct. I don't know whether they do this or not, and I'll bow to others' knowledge here.


Thank you Hans, that was the ingredient I was missing. The accelerometer measures force in a way that explicitly discounts the counter-force transmitted through structure, by using a calibrated spring. Not the same as a true measure of acceleration, but more relevant to the pilot's actual problem.

I stand corrected. In this case it does make sense for it to read 1 G when parked. It's just not an "accelerometer" in the Newtonian sense. It's a "G-Meter."

I was gonna correct you, but, as the engineer does, you found the discrepancy andfixed it yourself...Good job.
It's mainly because we stress stuff in and of the airplane to + and - g's. The pilot need to know these things, as it has a lot to do with whether or not he can use the airplane again, as well as passenger comfort.
Some parts just don't work as well in compression as they do in tension.

 

[R.Mackey]

I was gonna correct you, but, as the engineer does, you found the discrepancy andfixed it yourself...Good job.
It's mainly because we stress stuff in and of the airplane to + and - g's. The pilot need to know these things, as it has a lot to do with whether or not he can use the airplane again, as well as passenger comfort.
Some parts just don't work as well in compression as they do in tension.

Y'all got that right.

Last summer I worked with a young Air Force officer, just going through flight school, used him as a "test pilot" to run a full-mockup F/A-18 flight simulator we were using to drive simulated air data sensors. The guy had an absolutely fanatical aversion to positive-G maneuvers. Kinda funny watching him roll inverted just to flatten out his climbout, but it does make sense.

The accelerometers I work with report true dynamic acceleration, and they do so because they filter out the DC component... so what shows up on the pilot's gauge is the part that I throw away. Strange world, huh.

Always helps to have more than one pair of eyes on a problem!

 

[tsig]

the aircraft shows a 3960 fpm dive at time stamp :44. The G's after that show no correction. It actually shows less than 1 G. Where is the correction to start pulling out of the dive?

The G's shown prior to the :44 second time stamp are pulling level from 5960fpm to slow the descent rate.

/QUOTE]

jdx even your computer skills seem to be going.

 

[tsig]

But I now also understand JDX's 480MSL based on local pressure correction. Wow.

I'm going to try to contact this JDX guy and see if he worked for FLYI.

Just look in the mirror

 

[DavidJames]

the aircraft shows a 3960 fpm dive at time stamp :44. The G's after that show no correction. It actually shows less than 1 G. Where is the correction to start pulling out of the dive?

The G's shown prior to the :44 second time stamp are pulling level from 5960fpm to slow the descent rate.

/QUOTE]

jdx even your computer skills seem to be going.

About the time JDX's analysis began to receive serious attention on this forum, he stopped posting in both his forum and the LC forum. Coincidence?

 

[Anti-sophist]

I stand corrected. In this case it does make sense for it to read 1 G when parked. It's just not an "accelerometer" in the Newtonian sense. It's a "G-Meter."

That is a critical point. A G-meter is, essentially, a orientated, offset, accelerometer.

So, in vector math, it would probably be...

G-force = a / (mag(g)) + 1*(normalized(g)),

where a, G-force, and g, are all vectors

Note that to use normalized g, you obviously need to know the direction g is pointing (hence the orientated bolding, above). As such, for a proper calculation, you will need to have this value, in addition to an accelerometer.

G-force represents the "contact force" (normalized with respect to mass, so it's an acceleration) on a body to keep it's current acceleration. Therefore, to not accelerate at all (acc. vector = {0,0,0}), you would require 1 G-force worth acceleration (which is equal to your weight, in the upward direction) in "contact" with you.

 

[DavidJames]

JDX is back at LC.

All Airdata systems go through rigorous testing to remove errors at various speed, angles of attack, and position error. The following are highlights to show how an airdata system is calibrated to remove errors when undergoing certification. Full report is linked below.

"The presence of the aircraft in the airstream causes input errors to the measuring
instruments — the aircraft disturbs the air that it flies through, thereby also disturbing the airdata
measurements. Figure 1 shows the airflow around an airplane wing."


"Accurate airdata are necessary for many purposes and applications. Obviously, the pilot
cannot safely fly the aircraft without knowing airspeed and pressure altitude. In civil aviation, the
small vertical separation between flight levels assigned by air traffic controllers is based on
accurate knowledge of pressure altitude."


"Static pressure can be measured with a pitot-static tube or a flush-mounted port on the
fuselage. Figure 3 shows a typical subsonic static pressure distribution on an aircraft fuselage
(ref. 2). The measured minus true static pressure, , normalized to compressible dynamic
pressure, , is plotted as a function of fuselage position. Zero static pressure error on the
fuselage exists at locations 2 through 5."

"Even with the selection of the best static port position, some pressure errors will remain, and
these errors must be determined in flight. The difference between the locally measured static
pressure and the ambient static pressure, which is dependent upon angle of attack, airspeed, and
aircraft configuration, is called
position error.
.
Three calibration types are generally used to determine position error: direct comparison,altimetry, and velocimetry."

"This subsection describes typical maneuvers and methods for most airdata calibrations.
Tower-flyby, trailing static or trailing cone, pacer aircraft, radar tracking, and dynamic maneuvers
are included."

"Lag and attenuation can be estimated or measured experimentally. Criteria can be set for how
quickly pressure can change in the pneumatic system without affecting the airdata. Such
calibration methods as the trailing cone may have very large pneumatic lags and may have to be
used in steady flight."

"Airdata quantities are needed for a multitude of tasks, including flight safety, control,
navigation, weapons delivery, flight test, and flight research. These quantities generally need to
be measured and then calibrated to remove errors. The techniques and procedures have been only
briefly described here; numerous references should be studied if airdata values are to be measured and calibrated."

http://dtrs.dfrc.nasa.gov/archive/00000221/01/104316.pdf

Conclusion - The Flight Data Recorder doesnt have any errors in terms of air data (altitude, airspeed). Those errors are removed during calibration testing for certification.

From Loose Change Forum->September 11, 2001 Sites->The Pentagon -> Airdata Calibration And Measurement

 

[R.Mackey]

Conclusion - The Flight Data Recorder doesnt have any errors in terms of air data (altitude, airspeed). Those errors are removed during calibration testing for certification.

His comments look pretty good, except for his conclusion. Probably because everything but his conclusion is cribbed from NASA Dryden.

The conclusion is all wrong. Calibration and certification are different things. And the Dryden report talks about why there are errors, so for him to conclude that it says there aren't any... well, connect the dots.

 

[CurtC]

I'm pretty shocked by his conclusion, even for JDX! I had already read that paper, and it specifically talks about the fact that there are errors, but the airplane designers try to orient the sensors for the normal range of operation for the airplane. The only conclusion I could draw from it is that the pressure altitude reading from a commercial jet operating at >500 mph less than 1000 feet off the ground, would very likely be wrong.

How he comes to the conclusion from that same paper that it wouldn't have "any errors" just floors me.

 

[Anti-sophist]

His comments look pretty good, except for his conclusion.

Maybe because he copied his entire comment up to that point from a scientific paper?

The entire thing is gibberish, anyway. That paper doesn't describe FDR avionics.. it describes general calibration for test aircraft. Commercial airliners don't do atmospheric modeling, they don't do radar calibration, they don't do tower-fly-by calibration, and they definitely do not do pacing aircraft calibration.

He is applying NASA test documents to FDR sensors. It's a complete and utter red herring, and factually false.

I don't understand why he doesn't just look up the referenced regulation from the FAA, and get the actual allowable error rates, and go from there. That should be easy.

 

[Anti-sophist]

http://www.airweb.faa.gov/Regulator...E?OpenDocument&Highlight=flight data recorder

I win?


Of specific note, according to FAA rules:
+/- 2 degrees error for roll and pitch
+/- .36G (1% for calculation, 5% for data, of maximum G force, which is 6) for G-force
+/- 100 to 700 feet for pressure altitude

 

[LashL]

I don't understand why he doesn't just look up the referenced regulation from the FAA, and get the actual allowable error rates, and go from there. That should be easy.

Because that would require him to admit that he is wrong.

 

[apathoid]

The entire thing is gibberish, anyway. That paper doesn't describe FDR avionics.. it describes general calibration for test aircraft. Commercial airliners don't do atmospheric modeling, they don't do radar calibration, they don't do tower-fly-by calibration, and they definitely do not do pacing aircraft calibration.

I've done many, many, many pitot-static/airdata checkouts on Boeings(the one where you connect hoses to the p/s probes as well as the flush static ports and simulate the pressures)and I can give you a basic idea of the ranges we check and the allowable errors(these are approximate)

Test condition    Error
(ft/kts)         (+/- ft/kts)
0/120              30/3 
5,000/150          45/4
10,000/250         60/4
20,000/280         100/5
25,000/300         120/5
40,000/350         200/6

As you can see, we dont test for high speed at low altitude. There is probably a reason for that.
I'll have a look at the testing criteria for the 767 Air Data Computer in the Component Maintenance Manuals this evening.....should be something pertinent to our discussion in there.

 

[DavidJames]

JDX hasn't acknowledged any of these comments yet. Does anyone still have the ability to post there, it would be nice to give him a nudge.

Also,
I wonder if we'll ever hear from weedwacker or skeptic4sure again?