AAD Fatality Thought

[nigel99 143]

RiggerLee

They're trying to do it with just a static pressure sensor. Depending on how smart you tried to make it you could try to do some interesting things. When you tumble or change position you basically get discontinuities in the graph. With enough brains you could try to peace that togather to extrapolate where you actually are at. The people at cypress have at times implied that they are doing some thing like that. I can also see it getting fooled like when you set up on opening under a snivelly canopy. Every body know that it will fire high that way. What I've always wondered about is why they don't have a three axis accelerometer in there. They've gotten pretty cheep and it's not like you need a fancy super accurate one. Even just the real simple questions like are you on your belly or on your back would allow you to correct a lot of your burble error. I'd really like to see how much correlation you could get out of the data. It's fixed pretty well to the rig and your back. How would it relate as you did transitions? And you could actually look at the total acceleration relative to the rate of change in pressure. It might let you distinguish between a sudden roll and a sudden deceleration. It could show you the difference between the pressure increase from setting up under canopy and the increase from a continued fast fall rate. Basically I think you could peace the discontinuities in the pressure curve together with the acceleration data. First with a burble correction to the data it self from the angle. Second with a correction at the points of discontinuity based on the total acceleration.

Lee

From memory when I was reading one of the manufacturers patents I got the impression that they had an accelerometer in use (I think it was Cypres, but not 100% certain).

Ultimately though - the more you filter the data and ensure that you are not going to fire falsely, the more likely you are to delay the decision to fire.

To be honest this whole debate is why I would be much more comfortable with having the software change logs published. Cypres admit that during their 4 and 8 yr services they update the hardware and software. This implies that they are 'tweaking' things along the way. Without visibility as a user, you have no idea what level of test user you are. I think Vigil are a bit more transparent on the software side of things.

Experienced jumper - someone who has made mistakes more often than I have and lived.

[df8m1 19]

Last year there was a BLM fatality during a series of test jumps, for what I am not allowed to say, but they were using my Parachute Flight Data Recorder on each jumper and the free fall data of the fatal jump was collected. There was video from the plane and the ground, but where everything went wrong neither had any useful footage. To my knowledge, this is the first time that there was an incident with a "Black Box" to collect data.

We were contracted by the Government to analyse the data of the incident as well as the other jumps to try to provide the investigation team more detail as to how that particular jump compared to the others, and what the jumper was experiencing out of sight of the cameras.

This was enough of a challenge due to the how their equipment positioned the jumper in free fall, but we found out that the data recorders were not mounted properly in any respect. We were able do some hanging harness tests and found new calibration reference for the incorrect orientation so we were able to get a more realistic idea of the jumpers orientation during the jump which was of great interest by the investigation team.

This particular data recorder was designed to measure parachute flight performance, not a free falling body, as such, the instrumentation was not perfectly suited for collecting this type of data. The data recorders had a 3-Axis Accelerometer, Z-Axis Gyro, Baro pressure, Temp, Humidity, digital compass, and air speed (not used during the jumps in question).

It was quite a challenge, as of all the data we had from many different conventional drop tests really did not help us other than identifying the exit and canopy deployment. We used these points and went in both directions. We were able to come to a conclusion as to what happened, up there, but obviously not exactly why it happened. I was actually surprised to see how much detail we were able to see in the data once we knew what we were looking at.

I can't talk about the details, but after doing that I can definitely see the value of a free fall data recorder on every service jumper. I have added a 3-Axis Gyro and a 3-Axis magnetometer (configured different than a compass), and my Military AADs have these free fall data recorders built in and record every aspect of a jump.

From what I understand every AAD has an Accelerometer except for the Argus which used a Magnetometer. Although useing a single 3-Axis Accelerometer to identify orenation is easy to do in a static state, it is a whole different matter when the object is rotating on more than one axis or there is straight line energy being input. As I am sure you know this is getting into the 6 or 9 degrees of freedom arena, which requires a good amount of live processing power to utilize at 200 ft/sec.

No Barometrically referenced AAD will fire at constant altitude regardless orientation to the pressure front (relative wind). There will be a "window" in which the AAD will fire depending on the orientation of the jumper. Now as far as the AAD is concerned, it "usualy" has a set of fixed firing altitude, but if it is in the burble it will fire at one AGL and if it is exposed to the pressure front it will fire higher, if it sees an average, then it will fire in the middle (+-).

Adding some intelligence to the processing logic along with some additional sensors besides a baro sensor, allows the AAD to tighten up on the firing tolerance quite a bit as well as reducing or avoiding firing when it is really not called for. But that requires lots of processor involvement and memory, which reduces battery life or requires larger batteries.

From what I can remember, and they may have changed this since then, but I recall that the Cypres can get confused if it sees conditions that are outside its limits and will shut down. I recall some of them shutting down on a helicopter ride at the WFFC? If it still is unable to handle certain situations then it is certainly possible that the logic could have issues when encroaching this limit?

Even if it shuts down, I find it hard to believe that the error was not logged internally and the data up to the point where it shut down would not be there. I have seen quite a few "suicides" of Cypres owners that "could" be due to it shutting down. I am not sure what Vigil or M2 does, or if they even can "get confused".

If the Cypres code is such that it can get locked up, or confused by rapidly changing conditions, that would imply that it is trying to draw conclusions based on previous data and when the new data doesn't match what it expects to see, then the logic tree may hit a dead end, no pun intended..

[df8m1 19]

Divalent

Something I thought about 2 years ago after the Z-hills double fatality is how an AAD's triggering algorithm would handle unusual data.

Sure wish we had the data from the AADs in that incident.

One of the first things we plan on doing is trying to recreate that scenario as best we can with the data recorders in the reserve containers so we can get as realistic of an idea what those conditions are like.

The Fact that 2 different reserves in different containers “seemed” to be deployed late and at the same time would indicate that the common link, the AAD, had some issues for what ever reason.

This particular incident really is one to study for sure. Either the calibration was not right when they took off, or the conditions in free fall were such that they through the calibration off far enough to delay the indicated altitude reading?...

I think it would be good for the AAD manufacturers to release the data from the AADs involved in an incident to show that it worked properly, even if in reality it may not have, as, if the “zero” was off then the data would look normal except for a shorter time from activation to impact.

If the indicated altitude was skewed due to the two of them in close proximity, (nothing we do not do every jump, granted generally not that low), then I would expect to see the baro graph indicate high rates of change prior to activation and right after, and again with a shorter time from activation to impact.

Most of the data I have is owned by the Gov, so I am very anxious to get the free fall AAD beta unit done so we can start collecting some comparative data.

Is this kind of information of any interest to the jumping public or just a few of us?

[mark 102]

df8m1

The Fact that 2 different reserves in different containers “seemed” to be deployed late and at the same time would indicate that the common link, the AAD, had some issues for what ever reason.

This particular incident really is one to study for sure. Either the calibration was not right when they took off, or the conditions in free fall were such that they through the calibration off far enough to delay the indicated altitude reading?...

I think it would be good for the AAD manufacturers to release the data from the AADs involved . . .

The police returned the AADs to the families. The families declined to make them available to investigators.

Mark

[pchapman 261]

df8m1

Last year there was a BLM fatality

FWIW, for the rest of us, the public report on the Mark T. Urban test jump accident of 2013 can be found at:

http://www.wildfirelessons.net/communities/resources/viewincident/?DocumentKey=a7aa252b-85dc-4c03-9cce-b73a43301fa2

"Smokejumpers were performing extended drogue ride parachute jumps for research and development (R&D) purposes "

The main report is 111 pages.

As for confused AAD's, someone could dig up the threads but it involved a C-130 doing pressurization stuff while on descent, to an apparent altitude below ground level. It was World Team 2006. Both Vigil and Cypres made public statements afterwards, see below. C1's detected pressures outside the normal range and so concluded there was an error so the best action was to shut off. C2's were more advanced and decided it wasn't a permanent error so stayed on and did nothing. Vigils got jumpy and fired.

Cypres: http://www.cypres-2.com/cypres_news_letter_feb_2006_c.pdf

Vigil: http://www.vigil.aero/wp-content/uploads/VigilStatementAUG06_000.pdf

and
https://web.archive.org/web/20060714212226/http://www.vigil.aero/pdf/WorldTeamResponse_004.pdf
(They seem to have cleansed the original document from their site, hence using archive.org.)

Do electronic AAD's have any accelerometer at all? If so, it isn't something talked about by the companies. Supporting data?

Interesting research you've been doing.

[RiggerLee 61]

I think you are way into the "just a few of us" category. The typical jumper can not hook up a three ring. Hell a lot of them can not pack their own main. But every single one of them could benefit from what you might learn from this. And the just a few of us that want to know... we really want to see that data.

Lee

Lee
[email protected]
www.velocitysportswear.com

[Rstanley0312 0]

pchapman

***Last year there was a BLM fatality

FWIW, for the rest of us, the public report on the Mark T. Urban test jump accident of 2013 can be found at:

http://www.wildfirelessons.net/communities/resources/viewincident/?DocumentKey=a7aa252b-85dc-4c03-9cce-b73a43301fa2

"Smokejumpers were performing extended drogue ride parachute jumps for research and development (R&D) purposes "

The main report is 111 pages.

As for confused AAD's, someone could dig up the threads but it involved a C-130 doing pressurization stuff while on descent, to an apparent altitude below ground level. It was World Team 2006. Both Vigil and Cypres made public statements afterwards, see below. C1's detected pressures outside the normal range and so concluded there was an error so the best action was to shut off. C2's were more advanced and decided it wasn't a permanent error so stayed on and did nothing. Vigils got jumpy and fired.

Cypres: http://www.cypres-2.com/cypres_news_letter_feb_2006_c.pdf

Vigil: http://www.vigil.aero/wp-content/uploads/VigilStatementAUG06_000.pdf

and
https://web.archive.org/web/20060714212226/http://www.vigil.aero/pdf/WorldTeamResponse_004.pdf
(They seem to have cleansed the original document from their site, hence using archive.org.)

Do electronic AAD's have any accelerometer at all? If so, it isn't something talked about by the companies. Supporting data?

Interesting research you've been doing.

All the more reason to have an AAD checked periodically. Electronics deteriorate over time. If it is my last chance I want to be damn sure that the AAD is working properly. We will never know that with 100% certainty but periodic checks make me feel a lot better about it. Also, the most important thing an AAD actually does.... is decide not to fire. That is my opinion.

Life is all about ass....either you're kicking it, kissing it, working it off, or trying to get a piece of it.
Muff Brother #4382 Dudeist Skydiver #000
www.fundraiseadventure.com

[df8m1 19]

That is the first time I have seen the final report. We are supposed to meet with them at PIA and go over FFA’s report together.

That was a very emotionally taxing project that will never really be finished. Every time we look at the hundreds of graphs that were made, we see more patterns that indicate things, and we think of new ways to analyze the data after we have had time to think about it.

I hope that report is public information. If it is, you can see what the data looks like, and there is even some limited data from a Vigil.

As for AADs and Accelerometers.. I just looked at the Cypres 1 patent and it mentions storing acceleration values but I think they are referring to values based on processing barometric data. I would have sworn that Airtec had an Accelerometer mentioned in the patent, but I don’t see it jumping out at me. The Vigil Patent does mention an accelerometer, and I have talked to the designer if the Argus and we talked about him using a Magnetometer as apposed to an accelerometer, so I may be all wet in regards to the Cypres.

If the Cypers is only basing its decision on one channel of data, that is one complex algorithm.

[Divalent 89]

mark

The police returned the AADs to the families. The families declined to make them available to investigators.

Mark

I think it's unlikely the student owned the AAD. More likely it belonged to who ever rented the rig to him.

[pchapman 261]

As for the unfortunate BLM accident & the report, it does show how messy data is and how you and other had to work to tease things out of it. (I've got a couple Russian D-6 jumps and found droguefall easy to control but it was without bulky draggy gear and much more freefall experience than smokejumpers would have.)

Accelerometers:
My best guess is that any acceleration stuff that gets mentioned for Cypres or Vigil are just calculations based on pressure based altitudes.

I double checked the main Cypres patent 4858856 and Vigil patent 6,378,808 and they don't say anything about using any accelerometer - although Vigil does mention some unrelated patent that wanted a single accelerometer.

Indeed later patents by others try to claim using all sorts of multi axis MEMS acceleration sensors. (eg, 7,073,752, although it is about detecting low altitude military exits quickly)

Long ago I thought Cypres or Vigil boasted about how many parameters they used in order to make their calculations (even if based on just one sensor!). Can't recall details or find any record?! It was one of those marketing things that sounds nice but didn't really provide useful details.

Cheers.

[df8m1 19]

Data is a mess even at really high rates that are averaged down. As a body falls through the air there are frequencies generated to no end. And they are constantly changing as during the drop, so it is not as easy as filtering out a certain frequency range. Changes in frequencies indicate a change has taken place. Even just a change in amplitude means something changed to cause the frequency change. The trick is to look at the other instruments and see if they also picked up the change.

Many times certain points on a graph of a given instrument are identified, but not were not solely identified by that instrument. After a while you start to be able to read the squiggly lines like another language. The trick is to let the data dictate the results, not what you want it to be or think it should be. If the clamed can be backed up, then it is, if not, then there are varying levels of confidence in the possibility, and ones credibility as a data interpreter is dependent on the validity of the interpretation as a whole...

Thinking about it, the Cypres 1 design is over 20 years old, and the Cypres 2 is not to far behind it in reality. I would wager that the Cypres 2 code is largely based on the 20 year old code with some changes as they learned along the way, which is real life.

The Vigil patent that I have that says something about an accelerometer is the same one that you have and after reading it again, I see that they say the use of an accelerometer is “cumbersome and expensive”.. Implying to me now that they are against using one, where as before I was scanning the doc and miss read the section..

Now I am even more nervous about being on a plane with any AAD lol… Vigil thinks an accelerometer with its additional 3 channels at 8 Hz is cumbersome, I am dealing with 10 channels at 20 Hz, storing them to memory, processing and evaluating, and executing if required. Lol

[pchapman 261]

To give Vigil credit, their comments about accelerometers that was in a patent first filed in 2000. You've got some newer technology to play with!

More tidbits of info:

From an Airtec document that was online as of 2007, on archive.org:

Quote

What we did not expect at the start of the production but learned very quickly was that such a device would need many modifications during its life. There are technical reasons to modify things, there are environmental reasons and sometimes it is even the sport itself, which requires modifying the unit.

As an example: When we started in 1991 there were very few mobile/cell phones. When that situation changed significantly around 1993 – 1994, that required design changes. Another example is the change from the classic disciplines in the sport to freestyle techniques and further to head-down flying. These issues demand essential adjustment.
Within the production bracket of 1991 to 2003, we have carried out no less than forty modifications to the CYPRES 1.

Fortunately, we had installed the 4 years maintenance cycle. That gave us the possibility to incorporate the necessary modifications into the units. And that is exactly what we did.

So they do indeed mention how they have updated things over the years. They hint at the radio interference problem (although I thought it was more instructor radios than cell phones) and maybe even tweaks to the essential altitude & speed code.

Any admission of an improvement is also an admission things weren't perfect, but it is still an improvement.

Aha, here's the parameter thing, also from something in Airtec's site as of 2007:

Quote

CYPRES constantly checks 7 criteria. When they are all "yes" then CYPRES will activate.

No other details though. The criteria could include data derived from the pressure, or internal flags about the electronics status, so on its own the statement doesn't in the end help much.

It is also briefly mentioned in the Cypres Design & Test report, their PIA TS-120 report, that seems to have the most info anywhere I've seen about how the Cypres works. It stated:

Quote

For instance: altitude, vertical velocity. The other 5 activation criteria are still not disclosed.

[df8m1 19]

They could easily be processing the data from the baro sensor and have 7 checks, or conditions, so to speak that the processed data has to meet.

It sounds like the Vigil and the Cypres are trying to Dead Recon position strictly using a baro sensor. Position (past and present) and Velocity (descent rate) are no brainers, and it sounds like they take those values and do more computations to possibly predict (in the case of Vigil) the jumpers location, or answer conditional statements.

Simplistically complicated for sure. The problem with basing everything on one channel is if the base data, that all the calculations is based on, is wrong, then it’s a house of cards, and could explain why the Cypres 1 would lock up, not to say the Cypres 2 will not either, but if the base data results are whacked out, then the 4th and 5th level process results will be way out, I can see where it would say, hay, this is not right.

The technology has certainly advanced since the Cypres 1, heck, I am using instruments that have gone through 4 evolutions in capability since I started using them. The instruments can do more, but that just creates more data that has to be handled.

I am a simple guy and I do not like “complex algorithms”. Lol

I would think that Airtec would promote that they are constantly updating the units and how the update improved the unit, but then again German philosophy would determine that as a failure to get it fight in the first place which is not acceptable. That is what I have taken away from working with a few German companies, it is not a quality driven mentality, it is a fear of failure that drives the sometimes overly complex and sometimes quality, that the marketing department promotes lol..

[RiggerLee 61]

This was my understanding. That everything was based upon or extrapolated from the barometric data. My question, and based on the data you've gathered I think you might be able to answer it, is how well the accelerometer data can be correlated to the pressure. Your study of the tumbling incident represents an extreme and difficult case. Keeping it more simple.

How much change do you see between belly to earth and back flying?

The Accelerometer located in the standard location in the tray is fairly close to the CG. How much "noise" do you see from normal movements, center point turns, tracking, spins, rolls, vertical transitions, etc.? Can it tell you which way is down?

How well does that correlate to the burble? Can you predict and correct the error in the pressure to any degree from the accelerometer? Oh, look, I'm on my belly. Let's take 10% off the height. I'm on my back now. I'm 5% higher then it says.

So what I'm asking is can you filter the accelerometer enough to correlate it to the burble in any way? Is this some thing that could be done real time or just in retrospect.

Lee

Lee
[email protected]
www.velocitysportswear.com

[riggerrob 558]

.... As for confused AAD's, someone could dig up the threads but it involved a C-130 doing pressurization stuff ... to an apparent altitude below ground level. It was World Team 2006. Both Vigil and Cypres made public statements afterwards, see below. C1's detected pressures outside the normal range and so concluded there was an error so the best action was to shut off. C2's were more advanced and decided it wasn't a permanent error so stayed on and did nothing. Vigils got jumpy and fired.

..................................................................................

The story I heard (from a Cypres factory rep) was that as the engines started up - flight engineers closed all the doors and pressurized the cabin below sea level. Standard practice for military transport crews. The rapid rate of descent freaked out the Vigils and they fired.
Meanwhile, Cypres got confused by "below sea level pressures" and shut down.
The next day, a Cypres factory rep flew to Thailand with a suitcase full of Cypres 2 AADs.

[riggerrob 558]

Variable outside of pre-programmed parameters can really mess with an AAD's head. Fore example, when the Vigil 1 was introduced, they sent a couple to Strong Enterprises. Strong installed them in tandems and strapped 500 barrels to the front of the tandem-equipped dummies. 500 barrels are fairly common of military tandems re-supplying observation posts high in the Afghan mountains.
They tossed the dummies out of airplanes (10,000 feet) and watched. Nothing deployed and the dummies impacted at 200 miles per hour. It seems that the Vigil 1s got confused by descent rates faster than 200 miles per hour, so they shut down.
Vigil engineers never expected solo jumpers to exceed 200 miles per hour.
Now Vigil 2s understand descent rates exceeding 200 miles per hour when set to tandem mode.

[skydiverek 54]

Check this "Cypres did not fire" test jump:

http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=3881462#3881462

http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=3872546#3872546

[JohnSherman 1]

This is a great discussion. It's about time.
I believe it is rig/canopy specific and that it has to do with bag extraction effort vs. pilot chute drag capability. There are videos which show it happening. There have been about 20 of these instances recorded. All have been with the main closed.
There are some rigs on which this has never happened and there are rigs on which it has happened numerous times. The distribution is not aligned with rig sales distribution.

The problem is easy to define with a simple test.
During the FAA/Rigger Investigation with the same rig from an incident.
Assemble the rig with a fresh AAD charge and pack job.
Packed with both main and reserve.
Add enough weight to the harness reach AAD firing speed.
You could use bar bell weights in the leg straps.
You will need enough weight to achieve about 7 pounds per square feet of rig surface area, which I would guess at 3 sq ft. or, 25 pounds which will give you the weight of the rig plus the 25 pounds enough to reach the necessary speed.
Tape down all lose webs and stuff.
Toss it out at an altitude high enough to satisfy the AAD firing requirements.
We would expect it to fire at 750 and to deploy the reserve before it hits the ground, if it does it won't go far from less than 450 feet.
If it doesn't leave it there and call the manufacturer.
If it doesn't we will know the answer.

If you can't get the involved rig from the incident get one just like it with the same canopies.

The Deland test project seems to be stalled, however there are other testing activities occurring. We should have data by Symposium.

[df8m1 19]

RiggerLee

This was my understanding. That everything was based upon or extrapolated from the barometric data. My question, and based on the data you've gathered I think you might be able to answer it, is how well the accelerometer data can be correlated to the pressure. Your study of the tumbling incident represents an extreme and difficult case. Keeping it more simple.

For the sake of discussion lets assume that a tumbling situation is rare, but in reality if there is any chance a scenario can happen, then you have to factor that in to the logic. Students are required to have AADs and all they do at first is tumble especially at the bottom end because they get REALLY stiff lol..

RiggerLee

How much change do you see between belly to earth and back flying?

I would say a safe (general rule) number is between 125 to 200 Ft change in pressure altitude if you rolled over, definitely not any less but possibly more. It depends on a couple things, width of the jumper, speed, shape (arched or grabbing air etc.)

RiggerLee

The Accelerometer located in the standard location in the tray is fairly close to the CG. How much "noise" do you see from normal movements, center point turns, tracking, spins, rolls, vertical transitions, etc.? Can it tell you which way is down?

The noise values are generally much lower than any “discernable” rotation. I say discernable because it is in the center of mass so you would need to rotate at a good rate during a center turn to get an X or Z axis reading. A gyro would be much better as it would pick up rotation of less than 1 degree pre second.

As far as which way is down that is a different story. If you are falling stable, then no problem, you will get a X-Axis reading of about .9G because there is usually a little Z due to the angle of the rig, and that assumes a steady terminal fall rate.

If you were to stand on your head then there would be a free fall moment, say about .4G for a moment then it would get back to 1G, minus any pitch which would be seen in the other axis.

An accelerometer indicates orientation in a static state very nicely, but when it is in a dynamic environment, then you really need to know what is going on to have any chance of discerning which axis is down. With additional information from a Gyro you can tell what axis is rotating, and there for registering on the accelerometer’s complementary axis, so it can be factored in.

The best instrument to determine orientation, independent of movement, is a magnetometer as it is not influenced my movement, only magnetic fields.

Does that begin to answer your question? Or did I go off in a different direction?

RiggerLee

How well does that correlate to the burble? Can you predict and correct the error in the pressure to any degree from the accelerometer? Oh, look, I'm on my belly. Let's take 10% off the height. I'm on my back now. I'm 5% higher then it says.

So what I'm asking is can you filter the accelerometer enough to correlate it to the burble in any way? Is this some thing that could be done real time or just in retrospect.

Lee

The basic answer is yes, what you are talking about can be done with an accelerometer, but it the values would be based on assumptions, like, it will not fall faster than XXX Ft/Sec, and if you go outside of those assumed parameters, then the logic will fail.

IMOP, the best instrument for orientation measurement is a Magnetometer. What you are asking about regarding altitude calibration compensation for orientation, my AADs already can do. My AADs know if the jumper is in or has cleared the plane, or if they are being towed, or released from being towed, (in which case it will fire). I consider my AADs to be intelligent as they can determine the situation they are in, and make decisions based on those conditions. They will not fire in the aircraft at any altitude or descent rate. They will not get confused by going 500mph in free fall, or seeing below sea level pressures. All that has a price in that the processor is running at 100MHz and processing a lot of data, which is actually is handled by the data recorder.

The other AAD manufacturers have nothing to worry about in regards to me stealing how they are doing it lol..

[df8m1 19]

Yes there are definitely issues with some rigs and packing practices, but the incident that this thread is really focused on, had two rigs, one a student rig, and one was an instructors, and both reserves failed to open. I can't imagine the law of averages would predict that both reserves would be stuck in the two different sized, and probably aged rigs.

I think there was video of that jump from the instructor, and that would be helpful in determining their proximity and configuration at the time the AADs did fire. It they were on top of each other, then you have a tandem without a drogue, and the additional speed may have put the AADs behind the power curve.

Or the burble from the two close together may have killed the reserve pilot chutes pulling the bags out. Or the burble through off the AADs sense of altitude… What ever happened it happened to both rigs.

If they were both wearing identical team rigs then I could believe the tight reserve theory, but we are talking a student rig, (not saying that some one could have shoe horned a tandem reserve in there), and a instructor rig which I bet was in better shape and probably newer, ( but I an only guessing on that).

It will be interesting to see the results of some recreations.

[df8m1 19]

I have to apologize for the incorrect accelerometer axis (X should have said Y), in the above post regarding Lee’s question. It was 3AM when I posted it and after reading it to day I see some errors and it will not allow me to edit the post. So.. I am posting a corrective statement lol…

I also see that I did not address Lee’s question about being able to correlate the readings of an accelerometer to changes in sensed pressure by the AAD where he asked it. The answer is yes but…, if the jumper were to fly in crisp movements like an aerobatic plan will snap role 90 degrees, then the correlation is pretty straight forward. If however the pilot does a lanuchivec (sorry on the spelling) , like a tumbling jumper, then you will see above 0 readings on all axis’s at the same time and with only one accelerometer keeping track of the movements will take a super computer I think.

Having additional instruments to read independently so you have data to compare to makes identifying the combination of readings as a right turn or barrel role much easer. Using orientation to “tighten up” the firing “window” is very doable, but there will still be a “window” because of variables that are not considered in the logic.

I wonder if they have to try to “predict” where the jumper is going to be in an effort to make up for the latency caused by high levels of averaging a slow sample rate? I am surprised that they would do it in such a way that there is a “fixed speed range” of the jumper is part of the equation, and going faster will through the formula off into the weeds or cause a loss in AAD “actual” altitude awareness so to speak.

They are some vary smart people to have been able to get that to work IMOP.

[df8m1 19]

Would it make sense in regards to a testing standard for AADs that the AAD would have to function within the same parameters as a reserve is tested to, at least in regards to fall rate? The rationality is going beyond what the reserve is rated for in regards to fall rate is potentially disastrous, but then again the reserve has a loading variable.

I can imagine a test process that takes a AAD and pushes it till it fails to operate properly, and if increasing the descent rate will push the actual activation altitude lower, then the users should know what the limit is so they do not exceed it without knowing what will happen. Swoopers know there is a speed limit to respect if they do not want 2 out on final, so the same should apply for activation altitude accuracy. If a jumper was going to be doing some high speed dives, they could raise the firing altitude setting to compensate for the higher descent rate, but then again, there is the limits of the reserve and survivability of a high speed opening to consider with that thought as well.

We know the maximum “rated” limitations with the reserve and harness, but we have no idea what the limitations are for AADs. If speed is an issue then the manufacturers should put that in the owners manual what that speed limit is so we know to avoid descent rates above the limit, bellow a certain altitude so the AAD can catch up.

Or if having two jumpers near each other causes a lower pressure altitude than a solo jumper, then close proximity should be avoided near activation altitude. (this is a no brainier to me, but some instructors can not help but to chase a student bellow the hard deck instead of deploying as a signal, “hay.. you might want to think about pulling now” but that is for another thread.)

Am I over thinking things?

[RiggerLee 61]

Just for reference. The package we were dropping fell around 300 ft/sec. The cypresses were expired if that counts for any thing. The first batteries we used were old but new ones did not solve the problem. It was a small compartment just slightly bigger then the cypress so there was not a volume issue. We thought it might be an issue with the size or number of the static ports. They redrilled them and added a couple more and we do not believe there was an issue there. It was just firing late at inconsistent altitudes. Our conclusion was that 300 ft/sec exceeded the limits of the unit.

I repeat that should not be an issue for a skydiver under any normal circumstances.

I don't see an issue with AAD's firing at decent rates above the rated speed of the reserve. At least not at the lower hard deck. Some systems designed for bailout from air craft have move complex logic with mutable firing altitudes and time delays. But the bottom line is if you are burning through firing altitude with that high of a decent rate what other option do you have? Better broken lines on a reserve then a creator.

Lee

Lee
[email protected]
www.velocitysportswear.com

[df8m1 19]

300 ft/sec is just over 200 mph.. Any idea what the terminal speed for a no drogue tandem weighing = to the max rating? ( I'm not a Tandem instructor so I'm not sure what the limit is by the book)

___________________

John,

I know you are an advocate for a shorter reserve bridal. Would the effects of two jumpers tight together (creating a larger burble), benefit more or less from a shorter reserve bridal? I think it was my second AAF jump I remember pulling the main and looking back at the spring loaded pilot chute trailing above the three of us. Then I did a better "check" and gave it enough air to pull the main bag out.

Granted there should not be anyone around at reserve time, but in this case there was. Has anyone seen the video to have any idea how close the two were to each other when the AADs fired?

[RiggerLee 61]

Stable, I would guess a tandem pair should fall around 180 mph. I'm not sure what they would do head low on their back. There might be more then one reason why the opening altitude on the tandem unit is higher.

Lee

Lee
[email protected]
www.velocitysportswear.com