Accident Database / Post Crash Fires / How many flying?

[GraemeSmith]

The "Fatal in Florida" that I noted earlier in the week included a post-crash fire. This sorta jogged my brain a bit as I was under the impression that there have been a few fires caused by the under glareshield fuel tanks moving after a crash. Indeed, in my legacy Sonex I have modified the fuel line from the tank to the firewall to make it flexible so that it might avoid a rigid line tearing the tank fitting out the tank if the firewall deformed in a crash.

And it just so happened that the FAA WINGS Safety Manager at the Boston FSDO mentioned the Florida crash and also commented on the post-crash fire. He was under the same impression as me.

So - how prevalent are they?

Well the new NTSB database it much easier to search and sort - so I started there. I also then pulled the FAA database of currently registered and DE-registered aircraft (presumably accident aircraft) to get a sense of just how many Sonex Aircraft (all types) are out there flying and what is our accident rate REALLY? As a percentage of all aircraft completed and flying. So in summary:

Current Registry
================
502 Aircraft in Total

"AeroVee" - 1 (that's what is says the aircraft type is!)
Onex - 82
Sonex - 289
Sonex B - 7
Sub-Sonex - 15
Waiex - 79
Waiex B - 5
Xenos - 24

Deregistered
============
87 Aircraft in Total

Exports
-------

Australia - 3 Waiex, Waiex, Waiex
Canada - 4 Sonex, Sonex, Sonex, Waiex
France - 1 Sonex
Thailand - 1 Sonex

Deregistered still in USA
-------------------------

Onex - 4
Sonex - 62
Waiex - 10
Xenos - 2


Source of above - FAA Releasable Aircraft Data from Registration Database 10 Oct 2021

NOTE 1 - This does not necessarily represent aircraft flying. Aircraft that are close to flying but have not yet flown, but are registered appear in the figures above.
NOTE 2 - Aircraft that have been in recent accidents and that were destroyed - may still appear to be registered. The Sub-Sonex that recently went in a lake is an example of this. It only appears in a preliminary NTSB report and is still in the registry - though destroyed.
NOTE 3 - None the less there is a broad correlation between the 78 deregistered aircraft and the 60 reports in the NTSB database.


In the NTSB Database of Investigations
--------------------------------------

60 Reports (see attached summary in a spreadsheet - in attached ZIP file)


Investigating post-crash fires
------------------------------
Reading all 60 reports and dockets that were available in the NTSB Database (not all are) there were 3 post-crash fires that consumed the aircraft. 2 were almost certainly NOT survivable as the fire was after a blunt force stall spin impact. One was after a level forced landing but there was not sufficient information to determine if the pilot might have survived if there had not been a fire.

2 fires were in flight. One electrical in nature and the second was a fuel fire caused by improper assembly of the fuel lines from the tank to the firewall. The pilot survived that fire.

NOTE 1 - The Sonex crash and burn in Florida 8 NOV 2021 is an additional post-crash fire that has not yet appeared in the NTSB database.

NOTE 2 - While I didn't tabulate the 60 crashes - a very large number had an accident chain that ran: Distraction caused by engine problem/failure followed by loss of control/stall spin or loss of control/while attempting to land on whatever was available. Long and the short - don't get distracted - keep the plane flying. AVIATE first!

NOTE 3 - One investigation of a fatal crash had absolutely no probable cause even speculated as "aircraft was not recovered from the swamp".

--

So if you take 502 registered and 87 deregistered. There are 589 US completions or NEAR completions of Sonex Aircraft (all models). Based on a "guestimate" that at least 50 of the registered aircraft are not actually flying. (This a rough inspection of the list and recognizing folks I know are not flying yet and are still building). So let's call it 540 flying or used to fly. 60 made NTSB reports.

That's an accident rate of ~ 11.1% of aircraft built and flown
The fatal rate (20 people in 18 accidents) - 3.4% of aircraft built and flown

[pilotyoung]

Thanks for doing the research and posting about it. That is a valuable to all of us Sonex owners.

[kmacht]

Sobering numbers. If accurate it would say you have about a 1 in 10 chance of getting into an accident in your Sonex and a 1 in 30 chance of dying in it. Those are horrible odds.

[pfhoeycfi]

Great information. One of my early mentors told me that perhaps the best step I could take to protect against loss of control scenarios and prepare for loss of power situations was to get my glider rating and some XC experience. I have no idea if there is any hard data that would suggest he is right...but he did have 40k GA hours. Personally, the training and experience that I have had in gliders is by far the most valuable that I have had.

[radfordc]

Hard to say if the Sonex fares worst than other similar Experimental planes? Here is a good article discussing the overall fatal accident picture: https://www.avweb.com/flight-safety/ris ... l-factors/

Key points
- Pilot miscontrol—errors made in the control or guidance of the aircraft—is the major cause of homebuilt accidents regardless of injury severity.
- Pilot judgment issues. Nearly a quarter of the cases involve a deliberate decision by the pilot—mostly unnecessary low flying or continued VFR into IFR.
- Fatal Accident Rate-about a quarter of homebuilt accidents result in fatalities. The overall GA rate is about 18%
- What drives whether an accident results in a fatality? Two things: aircraft’s speed at impact. and ability of the structure to protect the occupants


Another article addressing fire: https://www.kitplanes.com/homebuilt-accidents-fire/

[WesRagle]

Hi Guys,

Yes it is sobering. I haven't learned to use the new data base interface yet but I did go through this a while back with the old interface, just scribbling down numbers, and came up with a similar result.

It begs the question: Why?

1) Pilot proficiency/training?
2) Airframe flight characteristics?
3) Firewall Forward?
4) Attitude/mindset of Sonex pilots?

I don't know, but I'm very suspicious of number 3. The recent accident in Burleson (south Ft. Worth) struck close to home (https://sonexbuilders.net/viewtopic.php?f=25&t=6019&p=45324&hilit=burleson#p45324).

If you compare the Sonex models with the RV-9A, I believe the RV-9 fatality rates are about 1/3 that of the Sonex models.

It's troubling.

Wes

[Scott Todd]

But I don't think we should let the 1in10 or 1in30 number scare us out of one. There are a couple things to look at. First its a small sample size. If the size were two, the 1in2 chance looks REALLY bad. As the sample grows, 1in3, 1in4, etc, it still looks really bad. Granted 1in10 IS bad as kmacht points out but it should decline as the sample size grows. There is also an infant mortality thing at play here which kind of leads to the second thing to look at.

The Sonex line of airplanes are not trainers. My apologies to the company but it really isn't. Its quick on the controls, has a moderate wing loading, many are built as tail draggers, and they use non certified engines. Any one of these factors contributes to accident causes. I think stall/spin is the leading cause which comes back to moderate wing loading and quick controls. Its just NOT a Cessna or Piper.

I don't want to discuss it here but non-certified engines statistically quit more that certified ones. They are usually assembled and maintained by non certified mechanics. This doesn't mean some of them are not done well but statistically speaking...And when they quit, the leading cause of accidents/injury comes into play, stall spin.

Our beloved little airplanes are promoted as affordable with descent performance, which sells airplanes but our accident rate is high. Insurance companies are figuring this out and hence the difficulty we are having. But notice the trend of highly experienced pilots with lots of Sonex time still getting affordable insurance. We need to continue to promote safety and training. Thank you Graeme for putting this together.

[radfordc]

One of my early mentors told me that perhaps the best step I could take to protect against loss of control scenarios and prepare for loss of power situations was to get my glider rating

I believe that having actual practice landing without an engine is a big advantage when the rubber band breaks.

I started in ultralights and engine failures were not uncommon (5 or 6 for me over 10 years). For practice I would routinely shut down the engine in flight and land deadstick...just like gliders do every flight. I didn't ever shut down a Sonex on purpose but did once have an engine stoppage on a test flight and landed with no problem. In a Sonex, pulling the throttle to idle and executing the landing to a full stop touchdown is good practice.

[pilotyoung]

I have two points to add. First, listen and read Dan Gryder's information. He recommends a speed, DMMS, Defined Minimum Maneuvering Speed. It is 1.404 times the clean stall speed. He recommends putting it on your airspeed indicator with red tape. I have it on my Onex, it is 65 mph. Then he says no matter what happens, don't go below this speed until you are on final and therefore have no more turns. This practice will keep you from doing a stall/spin. There are times when I am turning from downwind to base or base to final and it see I am heading below 65, DMMS, and I immediately correct. With it on the airspeed indicator in red, you can't forget it.

Second, my insurance broker told me that the main reason it is so expensive to insure a Sonex product is that there is no factory training program available. In reality, there is no training program at all. If Sonex would restart its training program, it would help us all buy insurance at a more reasonable rate and it might save some lives. Or alternatively is Sonex would partner with someone to start and run a training program it would be a big benefit to Sonex owners. And it would only take one airplane, a Sonex or Waiex tailwheel. With the exception of a Sub-Sonex, I think the insurance companies would accept dual instruction in a Sonex or Waiex tailwheel for the nosewheel airplanes and the Onex.

[GraemeSmith]

But I don't think we should let the 1in10 or 1in30 number scare us out of one. There are a couple things to look at. First its a small sample size. If the size were two, the 1in2 chance looks REALLY bad. As the sample grows, 1in3, 1in4, etc, it still looks really bad. Granted 1in10 IS bad as kmacht points out but it should decline as the sample size grows. There is also an infant mortality thing at play here which kind of leads to the second thing to look at.

Well one thing folks should NOT read into this is that you have a 1 in 10 chance of having an accident every time you fly. This is about how well the fleet fares.

Infant mortality is a factor. But you could only get a comparison if you could compare to the same number of (different type of) aircraft in about the same time period to see who fared worse. To take it to it's extreme:

The longer you leave a fleet in being - the worse the fleet figure will get. I mean they built about 23,000 C150/2's starting 50 years ago. Keep it up for long enough and their figures will look horrible as a percentage of losses against aircraft built. I think there are about 8,000 left on US registry. And of course many just rotted away or were exported. That said the current accident attrition rate on C150/2's in the USA is about 2 a week.

--

If you read the narratives of probable cause in the spreadsheet - people are just not staying in control of the planes once distracted.

There are also a few - first flight and fatal. Let's face it - jumping into a plane it took you 4 years to build and short on any recent currency because you have been busy building is not a good thing.

--

Interesting point about the insurance broker. I got some traction with offering to do the Foundation transition training syllabus as part of my first 5 hours with a CFI (down from 10 with no syllabus). Once we had that out the way - the CFI and I spent a lot of time doing improbable turns at maximum gross at ever lower altitudes over a very quiet and large airport. Minimum Maneuvering Speed established at 60knots Indicated at 45-degree bank.