0
pchapman

Philosophy of banning the Argus

Recommended Posts

Quote

CYPRES doesn't arm so early, and that's what invites the Vigil to fire when it should not.



I suspect there is more than just the arming altitude at play. If the plane had ascended above the CYPRES arming altitude, then descended back to the altitude at which the incident occurred, do you think the CYPRES would have fired? Or do you think it has data filtering and/or algorithms that would have prevented a firing due to a physically impossible event?
"It's amazing what you can learn while you're not talking." - Skydivesg

Share this post


Link to post
Share on other sites
Quote

Quote

Argus just seems more like a no brainer.



So in the case of the Vigil, you have an AAD that could misfire at any time - in the plane, in the door, etc. I don't trust them to not fire at an inopportune moment. I don't want to jump one, and I don't even like having them in the plane with me. Seems like a no brainer to me!




I gave a reason why I thought one might be banned over the other, but I actually tend to agree with both you and riggerpaul on this. I personally would not jump either one of them.
*I am not afraid of dying... I am afraid of missing life.*
----Disclaimer: I don't know shit about skydiving.----

Share this post


Link to post
Share on other sites
Quote

Quote

The Vigil, had 2 miss-fires in the incident that you mentioned, that seem to be due to a combination or rare set of circumstances. This has yet to be proven as a defined problem with the device. (not yet cut and dry)



Shortly after the incident in question, Vigil released graphs from the affected devices showing the recorded altitude and speed.

The graphs indicate that both units fired due to a very short spike in air pressure which represented an acceleration from zero to firing altitude in under a second - an acceleration several times that which is physically possible. It was very obviously not a situation requiring reserve activation.

There are also several incidents where they have fired on the ground (pressurized aircraft, slamming a car trunk lid, etc). Another physical impossibility - there is no way to be on the ground, and then in freefall at firing speed a fraction of a second later.

Any sudden, short pressure spike will cause the Vigil to fire, even if it is sitting on the ground. It is clear that the Vigil makes little or no effort to filter the data, nor to detect what should be non-firing situations. These are situations that are easy to deal with, yet the Vigil is no better than a mechanical AAD such as the FXC 12000 in that regard.

Not proven as a defined problem? The problem is clearly defined and it has been shown quite a few times that it exists. The manufacturer says so themselves every time they explain why one of these firings occurred.



There is a principle of keep it simple. Adding software filters changes algorithms and increases the amount of software. A risk assessment can show that it is lower risk to not implement something that may seem obvious to an outsider.

I would say that any software algorithm that is "life critical" is not trivial or easy to deal with.

I am sorry you have just caught me after a meeting explaining to someone why a "bloody obvious" software algorithm bug is hidden in 3000 lines of code. To quote a colleague "don't expect a shitty little 8 bit processor to do the same job as the 300 billion node neural computer between your ears does"
Experienced jumper - someone who has made mistakes more often than I have and lived.

Share this post


Link to post
Share on other sites
Quote

Yes, different reasons for the failures. But failures nonetheless. Either can kill.



Well, the argus is a clear failure.. It didnt cut the loop.

The vigil operated within it's parameters... You may disagree with the parameter (as do I), but the door opening created a pressure change that fooled the unit.

So one failed to fire, the other was tricked into firing... A CYPRES will 'misfire' if you pull low too. You may disagree with that parameter as well.

Don't get me wrong, I only recommend the CYPRES (and have gotten TONS of flack on here for that). But these two cases are the difference between a unit not firing (and thereby locking the container) and an operations parameter that is within the units programming that some people (me included) do not like.
"No free man shall ever be debarred the use of arms." -- Thomas Jefferson, Thomas Jefferson Papers, 334

Share this post


Link to post
Share on other sites
Quote

Quote

CYPRES doesn't arm so early, and that's what invites the Vigil to fire when it should not.



I suspect there is more than just the arming altitude at play. If the plane had ascended above the CYPRES arming altitude, then descended back to the altitude at which the incident occurred, do you think the CYPRES would have fired? Or do you think it has data filtering and/or algorithms that would have prevented a firing due to a physically impossible event?



Yes, that is what I understand (underlined text), that the CYPRES family will filter for impossible situations.

I seem to recall Airtec telling us that a tumbling skydiver will make it take longer for the device to fire. That also supports that they have data filters that try to detect spurious conditions.

Either way, such a descent back to firing altitude would constitute an unusual event that is easily recognized by the skydiver.

Having it happen shortly after takeoff when the door popped open is a completely different thing that no one could expect or plan for. It was only luck that it turned out as well as it did.

Even without other filtering, raising the arming altitude out of the firing range would have prevented the activations that happened at the particular moment that they did. Even that improvement would make me feel a lot better about being on a plane with them.

Share this post


Link to post
Share on other sites
Quote

There is a principle of keep it simple. Adding software filters changes algorithms and increases the amount of software. A risk assessment can show that it is lower risk to not implement something that may seem obvious to an outsider.

I would say that any software algorithm that is "life critical" is not trivial or easy to deal with.

I am sorry you have just caught me after a meeting explaining to someone why a "bloody obvious" software algorithm bug is hidden in 3000 lines of code. To quote a colleague "don't expect a shitty little 8 bit processor to do the same job as the 300 billion node neural computer between your ears does"



Of course unnecessary complexity is never good in a safety critical device. Neither is insufficient complexity to do the job properly.

What we are talking about here is something that improves safety. Yes, it will increase the chance of bugs; that is always the tradeoff. The question is whether the improvement in safety outweighs the risk of bugs - in this case I believe that it does.

BTW, I'm an engineer with a background in embedded systems, including safety critical systems, and including systems using pressure transducers to measure the altitude and speed of a skydiver.
"It's amazing what you can learn while you're not talking." - Skydivesg

Share this post


Link to post
Share on other sites
Quote

Either way, such a descent back to firing altitude would constitute an unusual event that is easily recognized by the skydiver.



On rereading my post, I think I wasn't clear enough. I meant that if the plane had ascended to arming altitude, descended again, then the door popped open. I was trying to come up with a scenario that would eliminate the arming altitude from the equation to allow a more equal comparison.

Sorry for not being more clear.

Quote

Even without other filtering, raising the arming altitude out of the firing range would have prevented the activations that happened at the particular moment that they did. Even that improvement would make me feel a lot better about being on a plane with them.



I agree, though I'd prefer to have both, and if I could only have one, I'd take the better algorithms over the increased arming altitude.
"It's amazing what you can learn while you're not talking." - Skydivesg

Share this post


Link to post
Share on other sites
Quote

Some opinions:

Quote

The Argus isn't much danger to others on the plane or the sky, as it isn't firing inadvertently.



I personally saw one fire at 11,000' while the jumper was in freefall. The only response/explanation from Aviacom was to send a new unit once the affected unit was returned to them. I wonder how many similar incidents have happened that are not widely known?


Quote

It would only be a problem if one smoked it low, had the loop not cut properly, and go up again without checking one's AAD.



So Argus users should open their reserve container before every jump to "check one's AAD?" My understanding is that you can also go through the Argus menu to see if it has "fired," but that as a procedure does not inspire confidence. Besides, neither option is practical to do before every jump.

Quote

The Argus also creates little increased risk to the jumper if they behave properly.



It could be easily argued that people have an AAD *precisely* for situations when they are unable to "behave properly" (i.e. unconscious).

Quote

An Argus can only impede a manual reserve pull if you pull your reserve a lot lower than you should



So let's make a bad situation even worse?

Quote

But if one's philosophy is that AAD's are an optional safety device, then Argus' are still useful. If you lose altitude awareness or are knocked silly in freefall, it quite easily still has a 90+% chance to save your life. That may be a good tradeoff against a chance of jamming up (depending on the rig) if you pull your reserve below 1000'.




Some would say that you're probably better off with an Argus installed than no AAD at all. I'd say the jury is still out on that one. But if you're going to spend the money on an AAD, why not just get proven technology like a Cypres?

Share this post


Link to post
Share on other sites
Quote

Quote

Either way, such a descent back to firing altitude would constitute an unusual event that is easily recognized by the skydiver.



On rereading my post, I think I wasn't clear enough. I meant that if the plane had ascended to arming altitude, descended again, then the door popped open. I was trying to come up with a scenario that would eliminate the arming altitude from the equation to allow a more equal comparison.

Sorry for not being more clear.

Quote

Even without other filtering, raising the arming altitude out of the firing range would have prevented the activations that happened at the particular moment that they did. Even that improvement would make me feel a lot better about being on a plane with them.



I agree, though I'd prefer to have both, and if I could only have one, I'd take the better algorithms over the increased arming altitude.



I understood what you meant.

What I meant is that if the airplane had reached the arming altitude, and then we decided to come down, I'd already know that I should take steps to be sure that nothing bad will happen. It might be turning off the AAD, which can be difficult. Or it could be as simple as putting my back to a wall of some sort, so that even if it fired, the pilot chute is trapped.

Are we on the same page now?

Share this post


Link to post
Share on other sites
I think it's funny that everyone wants to bash the Argus and Vigil but conviently forgets that the Cypres had a known problem that wasn't corrected until after someone died from it.

-Blind
"If you end up in an alligator's jaws, naked, you probably did something to deserve it."

Share this post


Link to post
Share on other sites
Quote


I agree, though I'd prefer to have both, and if I could only have one, I'd take the better algorithms over the increased arming altitude.



Per Nigel99's post, if they weren't comfortable with putting filtering software into the system in the first place, I'd likely feel better that they didn't try to add it later as an afterthought. So, I'd probably prefer just raising the arming altitude.

(I too have a background in embedded software systems, though not life critical or skydiver related like yours.)

Share this post


Link to post
Share on other sites
Quote

Some would say that you're probably better off with an Argus installed than no AAD at all. I'd say the jury is still out on that one. But if you're going to spend the money on an AAD, why not just get proven technology like a Cypres?



That doesn't help people (like me) who have an Argus already. I am still noodling over if I am better off without an AAD or using the one I already purchased. If I could climb into my wayback machine, I wouldn't have purchased a rig with an Argus in it.
For the same reason I jump off a perfectly good diving board.

Share this post


Link to post
Share on other sites
Quote

Are we on the same page now?



Yep, thanks!

Quote

if they weren't comfortable with putting filtering software into the system in the first place, I'd likely feel better that they didn't try to add it later as an afterthought.



Well, if it's just a quick fix without much engineering and QA effort, then I agree, and that would likely have prevented most or all of the undesired firings that have happened.
"It's amazing what you can learn while you're not talking." - Skydivesg

Share this post


Link to post
Share on other sites
Quote

I think it's funny that everyone wants to bash the Argus and Vigil but conviently forgets that the Cypres had a known problem that wasn't corrected until after someone died from it.



OK to repeat, though some of us remember it. Airtec took a long time (and one death) to finally admit that somebody some day might dive fast enough to set one off under canopy. Still, that was a change in one aspect of the sport of skydiving that caught up with what had been a perfectly reasonable design for years.

But the Argus is today's issue.

I'm kind of surprised nobody has a web page where they've taken the time to summarize all the AAD failures and rumours over the years, and how the companies responded. It would take some work! Things fade from memory and one doesn't always remember the details.

And there are always incidents one hasn't heard of, that seemingly haven't been investigated in detail, like in another post just made here:

Quote

I personally saw one fire at 11,000' while the jumper was in freefall. The only response/explanation from Aviacom was to send a new unit



(What year was that? Not one of the early prototypes presumably, gbstuar ?)

Share this post


Link to post
Share on other sites
Quote

Quote

CYPRES doesn't arm so early, and that's what invites the Vigil to fire when it should not.



I suspect there is more than just the arming altitude at play. If the plane had ascended above the CYPRES arming altitude, then descended back to the altitude at which the incident occurred, do you think the CYPRES would have fired? Or do you think it has data filtering and/or algorithms that would have prevented a firing due to a physically impossible event?



I know of two student cypreses that fired on the same load under these exact circumstances.

The pilot went above 1500 ft. Then descended to about 1000 ft. to quickly look for a cutaway canopy when he flew over the suspected location. He opened the cockpit window for a better view, and POP, two pilot chutes in the face. The pressure onset wasn't fast enough to trigger any other AAD, but it was for the student version. (Student versions trigger at a lower sensed airspeed and higher altitude)

So from this, I gather that there are aren't any cypres algorithms that perform in the way you are suggesting.

Share this post


Link to post
Share on other sites
Quote

I know of two student cypreses that fired on the same load under these exact circumstances.

The pilot went above 1500 ft. Then descended to about 1000 ft. to quickly look for a cutaway canopy when he flew over the suspected location. He opened the cockpit window for a better view, and POP, two pilot chutes in the face. The pressure onset wasn't fast enough to trigger any other AAD, but it was for the student version. (Student versions trigger at a lower sensed airspeed and higher altitude)

So from this, I gather that there are aren't any cypres algorithms that perform in the way you are suggesting.



Interesting! Thanks for posting about it.

When the window was opened and the AADs fired, was the plane still descending, or flying level? If it was level, how long had it been level?

Was there a response from Airtec?
"It's amazing what you can learn while you're not talking." - Skydivesg

Share this post


Link to post
Share on other sites
Quote

Quote

I know of two student cypreses that fired on the same load under these exact circumstances.

The pilot went above 1500 ft. Then descended to about 1000 ft. to quickly look for a cutaway canopy when he flew over the suspected location. He opened the cockpit window for a better view, and POP, two pilot chutes in the face. The pressure onset wasn't fast enough to trigger any other AAD, but it was for the student version. (Student versions trigger at a lower sensed airspeed and higher altitude)

So from this, I gather that there are aren't any cypres algorithms that perform in the way you are suggesting.



Interesting! Thanks for posting about it.

When the window was opened and the AADs fired, was the plane still descending, or flying level? If it was level, how long had it been level?

Was there a response from Airtec?



I don't really remember how long the plane was level, but I don't think aircraft decent was the cause.

The window that the pilot opened was a side window in the cockpit, but it is slightly angled forward. So when he opened it, it caused a slight ram air effect. (kinda like the nose of a canopy). This caused a sharp increase in cabin pressure.

The only thing an AAD senses is pressure. It records how fast that pressure changes to get decent rate. It has no way of knowing if it is in freefall or if a pilot opened a window. In this case, the pressure increase from opening the window was high enough to and fast enough to trigger the student cypreses, but not fast enough to trigger other AADs.

I believe that the consensus of dropzone people involved was that the student AADs performed as designed. The problem was with the pilot (a newer one if I recall) that should have known better than to open that window.

Because nobody blamed the AAD, I don't know if Airtec was notified or if they sent a response.

As far as filtering goes, the more data filtering you add to smooth out the pressure spikes, the more lag is added to the measurement, which slows down the unit's reaction time. If you add more logic so the measurement has to meet more criteria, there will probably be a skydiver out there who will eventually be in some situation that doesn't meet all the criteria, but really needs an AAD fire to save his ass. (In other words, better to use the KISS approach)

Or at least, that's the way I see it... :)

Share this post


Link to post
Share on other sites
An AAD should be expected to do TWO things:

1) Properly cut the reserve closing loop and

2) Do so at the appropriate time (high speed, low altitude).

The Argus has failed multiple times to do #1, and so we ban them until they can sort their shit out and perform under expectations 1 and 2.

The Vigil has failed multiple times to understand #2. If I owned an H/C company, the Vigil would be banned as well until it could perform under expectations 1 and 2.

If only there were an AAD manufacturer that did so. Maybe someone could recommend one to me. I'm preferably looking for one with a 20 year track record.
108 way head down world record!!!
http://www.simonbones.com
Hit me up on Facebook

Share this post


Link to post
Share on other sites
>An AAD should be expected to do TWO things:
>1) Properly cut the reserve closing loop and
>2) Do so at the appropriate time (high speed, low altitude).

Honestly to me those two things are secondary. By far the #1 thing in my list of AAD requirements is:

1) Under no circumstances interfere with the normal operation of the rig.

Then comes

2) Do not fire when the jumper is above about 1000 feet, and do not fire when the jumper is doing less than about 80 mph.

Those are the two really critical items. Everything else is secondary. I'd much rather have an AAD that fails to fire 10% of the time it's needed than an AAD that fires 10% of the time when it's not supposed to. (Or worse yet, prevents you from opening the reserve 10% of the time.)

Share this post


Link to post
Share on other sites
Quote

Quote

Yes, different reasons for the failures. But failures nonetheless. Either can kill.



Quote

Well, the argus is a clear failure.. It didnt cut the loop.



I wouldn't say that it is clear. That it happened it clear. What made it happen is not. It could be attributable to rigger error or other factors as yet undiscovered?

I think that the impact of this decision is going to be huge. Most of my rigs for the season are now out of commission and half my customers will be grounded. I am fortunate that I can afford to by replacements for my work gear but many DZs will not be so fortunate. I think this could have been resolved more effectively with an advisory requiring inspection after any low pull incident. I have also hear (not verified) that Aviacom have not been given access to the rig that has the mentioned incident. How are they to investigate it in this case?

Don't get me wrong I think Aviacom have not offered the best customer response in the past and this is probably at least in part why they are now in this position. But I think a response that would not have screwed a large part of the skydiving population would have been in order.

I hope for everyones sake that this will have a resolution soon. Right now I am sitting on $30Ks worth of bricks.
https://www.burblesoftware.com

Share this post


Link to post
Share on other sites
Quote

>An AAD should be expected to do TWO things:
>1) Properly cut the reserve closing loop and
>2) Do so at the appropriate time (high speed, low altitude).

Honestly to me those two things are secondary. By far the #1 thing in my list of AAD requirements is:

1) Under no circumstances interfere with the normal operation of the rig.

Then comes

2) Do not fire when the jumper is above about 1000 feet, and do not fire when the jumper is doing less than about 80 mph.

Those are the two really critical items. Everything else is secondary. I'd much rather have an AAD that fails to fire 10% of the time it's needed than an AAD that fires 10% of the time when it's not supposed to. (Or worse yet, prevents you from opening the reserve 10% of the time.)



Interesting. We agree. Except that I place not hurting someone else as the first priority.

For me -
1) pose no hazard to anyone else.
2) don't get in the way of my saving myself.
3) save me if I clearly need it and it can.

What do you think of locating single use, untestable cutters, where it could lock a rig shut?

And where do you stand on the whole matter that one is banned and the other is not?

Clearly, a partially cut loop that could fail later violates the #1 priority, so I have no problem with banning an AAD that can partially cut a loop.

Share this post


Link to post
Share on other sites
>What do you think of locating single use, untestable cutters, where it could
>lock a rig shut?

The technology is there to produce very reliable cutters. They are used in spacecraft and ejection seats with 99.99% reliability rates. You just have to spend the money.

That being said, it also makes the most sense (to me) to locate them where a failure does not cause a locked-closed rig. I.e. put them on the bottom of the pack tray on a conventional rig.

>And where do you stand on the whole matter that one is banned and the
>other is not?

I think it would be tough to come up with really hard and fast rules over which one to ban and which one not to ban. The Vigil came close during the "cutters are damaging closing loops" issue of a few years ago, but PIA (and rig manufacturers) never crossed the threshold of thinking they were too unsafe to use at all.

Share this post


Link to post
Share on other sites
Quote

I think it's funny that everyone wants to bash the Argus and Vigil but conviently forgets that the Cypres had a known problem that wasn't corrected until after someone died from it.



No one is forgetting the swooping accident. The thing is that was not an error *in the system*... It operated within it's parameters, period. It is not the fault of the CYPRES unit that the operational parameters were met AFTER the parachute was opened.

The Vigil did the same thing (fired when parameters were met). You may not LIKE the parameters, but that is different than a malfunction of this type.

So you could lump the Vigil and the CYPRES into the "I don't like the parameters that it works" And that would be fine.

But the Argus is creating a situation where it does not work properly. Worse, in some rigs it locks the reserve closed.

I didn't buy a Vigil based on the door issues. I didn't buy an Argus based on the groundings in Europe. I don't fly a 69sqft canopy anymore so I could easily live within the parameters of the CYPRES... so I bought a CYPRES.
"No free man shall ever be debarred the use of arms." -- Thomas Jefferson, Thomas Jefferson Papers, 334

Share this post


Link to post
Share on other sites
Quote

That doesn't help people (like me) who have an Argus already. I am still noodling over if I am better off without an AAD or using the one I already purchased.



Only you can answer that..... With a Dolphin (IIRC) the cutter is in the bottom of the container. So the chance of a partial fire locking the container is nil (If I remember the Dolphin correctly). But DAMN that would suck.... Your AAD locking your PC and not being able to get a launch.

So as long as you check the AAD to see if it has fired after you land from each jump..... You should be fine.

But like always you should expect the AAD not to work and plan accordingly.

Personally.... I have the money so I'd buy something else.

Edit:

The question is mute if this happens

http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=4084058;sb=post_latest_reply;so=ASC;forum_view=forum_view_collapsed;;page=unread#unread

And if I owned a HC company, I'd ground them till there is a fix as well.
"No free man shall ever be debarred the use of arms." -- Thomas Jefferson, Thomas Jefferson Papers, 334

Share this post


Link to post
Share on other sites
Quote

Quote

I think it's funny that everyone wants to bash the Argus and Vigil but conviently forgets that the Cypres had a known problem that wasn't corrected until after someone died from it.



No one is forgetting the swooping accident. The thing is that was not an error *in the system*... It operated within it's parameters, period. It is not the fault of the CYPRES unit that the operational parameters were met AFTER the parachute was opened.

The Vigil did the same thing (fired when parameters were met). You may not LIKE the parameters, but that is different than a malfunction of this type.

So you could lump the Vigil and the CYPRES into the "I don't like the parameters that it works" And that would be fine.

But the Argus is creating a situation where it does not work properly. Worse, in some rigs it locks the reserve closed.

I didn't buy a Vigil based on the door issues. I didn't buy an Argus based on the groundings in Europe. I don't fly a 69sqft canopy anymore so I could easily live within the parameters of the CYPRES... so I bought a CYPRES.




+1
*I am not afraid of dying... I am afraid of missing life.*
----Disclaimer: I don't know shit about skydiving.----

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

0