Cessna C208B down near Seattle, 4 deceased

[atcclears]

The crash was just East of Harvey Field (S43) in Snohomish, WA. The Cessna Caravan may have broken up before the crash.

There are comments on YouTube that the Caravan was involved with Raisbeck Engineering.

N2069B

[Pilot DAR]

It is very sad for the crew, the last part of that ride was terrifying!

The ADS-B track can be found here:

https://globe.adsbexchange.com/?icao=a1abe2&lat=47.899&lon=-122.047&zoom=14.6&showTrace=2022-11-18&timestamp=1668795559

I note, with alarm, a "groundspeed" of 176 kts (one knot faster than Vmo), and a rate of descent of 9408 FPM as the last report. Also noteworthy is the apparent increase from just faster than 50 knots, to more than 175 in just a few seconds, as seen on the Youtube video report. I know that when I did a spin test program on a modified Grand Caravan, I calculated a momentary peak rate of descent of 9600 FPM, at Vmo, and a 2.5G pull to recover the dive resulting from the spin recovery. I installed a G meter for the testing, as spinning was required for a demonstration of design compliance, and I anticipated getting into the corner of the envelope, and wanting to know just where I was there. I have dive tested several Caravans to 193 KIAS, and they are a smooth as glass, but I sure avoided rough air, and rough handling! I will await the results of the NTSB investigation with great interest!

[1a sound asleep]

Aircraft was unmodified and conducting baseline performance figures tests. So what makes a wing on a Caravan come off?


https://raisbeck.com/raisbeck-responds-to-snohomish-county-aircraft-accident/SNOHOMISH, Wash. — Four people are confirmed dead after a small plane crashed into a field and caught fire near Snohomish, according to the Snohomish County Sheriff’s Office.

“With assistance from the Snohomish County Medical Examiner’s Office, investigators confirmed 4 fatalities,” spokesperson Courtney O’Keefe said in an email Saturday.

Authorities initially reported that two people died in the crash on Friday morning.

The crash involved four crew members working for Raisbeck Engineering, a Seattle-based company that designs and develops modifications to aircrafts, according to a statement from Hal Chrisman, the company’s president.

“At the time of the crash, (the aircraft) was was under the command of two highly-experienced test pilots, both with over 10,000 flight hours, collecting baseline aircraft performance data,” said Chrisman. “The entire crew of four also included a flight test director and an instrumentation engineer.”

According to Chrisman, the aircraft had not yet been modified at the time of the crash and Friday’s test flight was meant to help the company measure the plane’s baseline performance before any modifications would be made.

He said the company will fully cooperate with authorities as their investigation into the crash continues.

The single-engine Textron 208B crashed at approximately 9:35 a.m. on Friday, according to the Federal Aviation Administration.

The aircraft went down in the middle of an unworked field in the 13600 block of U.S. Highway 2, about a half-mile from Harvey Field.

The location is difficult to access due its rough terrain with vegetation and irrigation canals, officials with Snohomish County Fire District 4 said.

Those who first arrived at the crash scene made several attempts to reach the plane with hand-held fire extinguishers, but crews were unable to get close enough to the plane to control the fire because the flames were too large and intense.

[punkalouver]

Video initial analysis here for those interested.....

(72) Cessna Caravan Structural Failure 18 Nov 2022 Renton WA. - YouTube

[megan]

Think he jumped in a bit quick with his analysis, four fatal not two, aircraft was unmodified, streamlining of the pod had not been carried out..

[punkalouver]

Think he jumped in a bit quick with his analysis, four fatal not two, aircraft was unmodified, streamlining of the pod had not been carried out..

If true, a good example how these analysts cannot be trusted to be giving us accurate reports. Some of these analysts have a poor reputation, others better.

He says that he has a lot of information that will clear up confusion at the beginning of the tape. He talks about testing and what he thinks are the maneuvers that were performed but it is just guesses. He was told that the aircraft is being leased to Raisbeck for testing. Therefore, the video is made in such a way that the assumption is that the mod has already been made.

Lesson to be taken(even if this video turns out to be accurate)......do not put your faith in these analysts to be giving correct information.

[fdr]

The crash was just East of Harvey Field (S43) in Snohomish, WA. The Cessna Caravan may have broken up before the crash.

There are comments on YouTube that the Caravan was involved with Raisbeck Engineering.

N2069B

Sad news, James' program is competent, he was always a positive presence in development of capable and effective enhancements to aircraft. While testing is always an elevated risk, this does not yet appear to be test related, the data will be telling as will the wreckage disposition and component evaluations.

R&D adds additional risks to what always has an underlying risk.

Once the shock has dissipated, even though the pain will remain, the team needs to get back onto the horse and continue doing what they have done so well for so long.

[Pilot DAR]

While testing is always an elevated risk, this does not yet appear to be test related

Gathering baseline data is test related. Ideally, following the installation of the test article on the airplane, the test flying and comparison to baseline data, shows no change, or desired improvement. But, it should be that no maneuver is being flown with the test article installed, which was not first flown baseline. When I flew a Caravan spin program, I very certainly spun the basic airplane first, just to get used to handling and recovery techniques. That done, any difference with the test article installed was a delta to the base airplane, rather than a whole new unknown. And, if an airplane is edging on non compliant in the basic configuration, it's useful to know that before installing the test article, lest the test article be incorrectly thought to be the reason for an observed non compliance. I have certainly flown baseline, then modified, found something unexpected, and had the mod removed to fly more baseline, just to compare.

[fdr]

If true, a good example how these analysts cannot be trusted to be giving us accurate reports. Some of these analysts have a poor reputation, others better.

He says that he has a lot of information that will clear up confusion at the beginning of the tape. He talks about testing and what he thinks are the maneuvers that were performed but it is just guesses. He was told that the aircraft is being leased to Raisbeck for testing. Therefore, the video is made in such a way that the assumption is that the mod has already been made.

Lesson to be taken(even if this video turns out to be accurate)......do not put your faith in these analysts to be giving correct information.

Yes, & No,

He does usually do a fair job at getting info out. This one is not one of his best efforts. The info on the AFM limits should be read carefully, there is a big difference between a TAS and a CAS limit. The g limit that is being discussed, yes, it is a limit, and no, it doesn't mean much if the aircraft was lighter or heavier than the weight that it is determined at, which is MTOM. Underweight, the structural margin will be better, (or should be). He is correct with the assumptions on this being an inflight breakup, but the sequence will determine what failed.

[punkalouver]

Yes, & No,

He does usually do a fair job at getting info out. This one is not one of his best efforts. .

Which proves me right. You can't trust even the guys who are said to usually do a good job. Bad information happens sometimes and that reason is likely due to a combination of rushing to get things out and a willingness to make assumptions. One could wait until the facts are known. I suspect the NTSB won't make this type of error.

[fitliker]

What were the upper winds doing at the time ? Being that close to mountains ,the inflows and outflows can also create turbulence that can cause upset very easily . The full investigation will include the possibility of mechanical turbulence from winds that close to mountains.

[tdracer]

What were the upper winds doing at the time ? Being that close to mountains ,the inflows and outflows can also create turbulence that can cause upset very easily . The full investigation will include the possibility of mechanical turbulence from winds that close to mountains.
Winds were relatively calm when the crash happened. Oh, and I don't really consider Harvey Field to be "close to the mountains" - Snohomish is on fairly flat terrain - the mountains are several miles away.

[atcclears]

The NTSB believes the right wing separated in midair from the C208B Caravan before it crashed.

[PJ2]

Any thoughts on how the 'g' loadings in the turns might be calculated?
There appears to be no CAS data in FR24, only GS, (and no bank-angle, of course).
Has any information been provided that there have been some form of recording equipment on board specifically related to the test flight?

[grizzled]

Has any information been provided that there have been some form of recording equipment on board specifically related to the test flight?

Hi PJ2, (good to see you here)

The following is from a Seattle news outlet: "According to a statement from Hal Chrisman, president of Raisbeck Engineering, the aircraft was confirmed under lease to Raisbeck, and “… at the time of the crash was under the command of two highly-experienced test pilots, both with over 10,000 flight hours. The purpose of the flight was to collect baseline aircraft performance data. The entire crew of four also included a flight test director and an instrumentation engineer.”

According to Raisbeck, the flight was conducted prior to the installation of a Raisbeck modification, which is “standard industry practice that allows aviation engineering firms to establish baseline aircraft performance under a highly structured flight profile to later measure and compare the change in performance after any proposed modifications are installed. The aircraft was in this initial testing phase and had not yet been modified in any way.”

The above seems to indicate there was testing / recording equipment on board. If the memoriy modules of that equipment surivived the crash, they will likely be of significant help to the investigators.


Grizz

[Pilot DAR]

There appears to be no CAS data in FR24, only GS,

I think that CAS speeds could not be conveyed by ADS-B, as CAS is unique to each airplane type, and configuration. Though they were reportedly doing stall testing, where CAS is a factor worthy of consideration, at faster speeds the IAS to CAS difference for the Grand Caravan is zero, so not a factor relative to IAS at this point in the flight. The ADS-B data shows 176 knot groundspeed at roughly the same time as a 9400 FPM dive. So it may have been flying at an IAS faster than the 175Vmo. If G is added to that (like pulling out of a dive), the airplane is getting into a tight corner of the envelope.

As the last reported data point is still at considerable altitude, I infer that the ADS-B output stopped. Perhaps if the airplane broke up in flight, the ADS-B transmission stopped at that point in time, rather than when the airplane struck the surface, for which there is no data point. The last reported position point, and the location of the fuselage on the ground are 1850 feet apart, and the wing 650 feet to the right of that track.

[henra]

Any thoughts on how the 'g' loadings in the turns might be calculated?
There appears to be no CAS data in FR24, only GS, (and no bank-angle, of course).

Problem is that the critical part of the the ADS-B track seems inexact and cut short.
The circle before, leading up to the speed going down to around stall speed is rather unsuspicious. It has a radius of ~700m at a (Ground)speed of 100-130kts with a slight descent. This would equal to a g- load of ~1.2g.
Towards the end of the circle it climbs ~300ft and (ground) speed decays from 130kts to 55kts which energywise (trading kinetic energy for potential energy) almost perfectly matches the altitude gain. This would indicate a pull up at low power/idle. Up to that point the trace looks reasonable and credible for a test flight.
After that I wouldn't put too much trust in the remaining part of the track. It does look like a stall being recovered and a subsequent high speed dive possibly with an initiated pull- up..

And now for some musings:
If we would take the figures as they come (with the usual ADS-B caveat):
The acceleration is on the high but possible side for a steep dive out of a stall: G/S increases within 6s from ~40kts to 140kts, Descent rate reaches 14kfpm. Along the trajectory this would equal to an average longitudinal acceleration of ~1g (speed ~200kts). Descent rate then decreases within ~3s from 14kfpm at G/S of 140kts (would correspond to a dive angle of 45°) to around 9kfpm at a G/S of 175kts (corresponding to dive angle of 27°). This pull- up would produce an average g load vertical to the flight path (I assumed 35° angle as reference point for normal gravity) of ~2.3g.That doesn't say anything about peak loads, though. So besides general inaccuracies and having only G/S instead of airspeed this neither contradicts nor confirms a possible overload and/or overspeed in the dive or pull- up. Speed and descent rate both appear high, though.

[Pilot DAR]

around 9kfpm at a G/S of 175kts (corresponding to dive angle of 27°). This pull- up would produce an average g load vertical to the flight path (I assumed 35° angle as reference point for normal gravity) of ~2.3g.

These values correspond well to values I saw during repeated spin testing I did in a Caravan many years ago. The dive following recovery got me to 175 - 180 KIAS, and pulling 2.5 G (I installed a G meter for this testing) kept me from a faster speed that that. Had I not pulled the 2.5 G, I would have substantially oversped the plane.

This is the recorded G following my last spin of that session, which had been a near Vmo recovery

https://cimg0.ibsrv.net/gimg/pprune.org-vbulletin/800x600/img_2249_3c8dffd4f7cf8627084a7a275f1bbc01b6091c91.jpg



Here's a video of one of the spins

https://www.youtube.com/watch?v=zjB_q7AIvDo

[henra]

These values correspond well to values I saw during repeated spin testing I did in a Caravan many years ago. The dive following recovery got me to 175 - 180 KIAS, and pulling 2.5 G (I installed a G meter for this testing) kept me from a faster speed that that. Had I not pulled the 2.5 G, I would have substantially oversped the plane.

This is the recorded G following my last spin of that session, which had been a near Vmo recovery


Here's a video of one of the spins

https://www.youtube.com/watch?v=zjB_q7AIvDo

Very interesting!

When I look at the timing it was roughly 5s between stall/spin entry until begin of pull- up in your case. That matches the timing in the ADS-B trace rather well. There it seemed to be around 6s.
This additional 0,5 - 1s might also fit to a potentially somewhat higher airspeed in this case. Your recovery took about 6s to level horizon. Fits also quite OK to the 3s from 45° to 27° derived from the ADS-B data. Astonishingly in this case the ADS-B data might not be that far off reality even in the upset situation.

I really appreciate this invaluable real world data!
Thanks a lot!

[PJ2]

Hi Grizz...tx, nice to see you here too - it's good to be anywhere these days '-) !
The failure of a major part of the aircraft that doesn't have a history of such failures, is certainly concerning. Payload calculations will obviously be a focus. Terribly tragic, including the loss of experience and knowledge. Remarks by FDR and others tell us that this was a good and very competent company doing the work and the testing.

[atcclears]

Update from a local news article:


The crew of the Cessna 208B had already done three days of test flights, but the day before the crash they ended early because one of the crew members felt ill. The crew went back up the following day and was testing the Cessna’s aft center-of-gravity stall characteristics when the plane crashed, the agency said.

Witnesses said the airplane broke up in flight and descended in a near-vertical corkscrew to the ground and several witnesses reported seeing a white plume of smoke as the airplane broke into pieces, the NTSB report said. The agency has previously said a wing broke away from the plane during the crash.

[Pilot DAR]

and was testing the Cessna’s aft center-of-gravity stall characteristics

Two thoughts come to my mind with that information: If there had been a ballasting error, and the C of G was actually further aft than intended, recovery could have been an extra challenge. If the stall was allowed to develop into a spin (which may be inferred from the final very tight turn in the ADS-B trace), recovery would be demanding. If both factors were combined, it would be a great challenge for the pilots to recover.

I am responsible for not detecting a couple of ballasting errors during flight testing over the years, which has made me extra vigilant in that regard. I errantly accepted a "behind the aft limit" Cessna 185 for spin testing once, and it was immediately apparent that spin recovery was noticeably more difficult behind the aft C of G limit. Lesson learned!

[Pilot DAR]

Summary article in Flying Magazine here:

https://www.flyingmag.com/ntsb-releases-details-of-caravan-crash-that-killed-4/

[212man]

It is very sad for the crew, the last part of that ride was terrifying!

I’m sure, although possibly less so than one might imagine. I was involved in an airprox between two helicopters, whilst acting as an examiner/observer on an IF training session in one of them. I saw the other aircraft suddenly appear from behind the coaming in our 1 o’clock and slightly low. We were in his 7 o’clock, and converging, so totally out of view as he was single pilot. I started saying loudly “turn left, turn left” but there was no reaction. RHS pilot was fixed on his instruments and couldn’t understand why I was so upset about a 3 degree ADF divergence, LHS pilot I don’t know. In the space of about 5-6 seconds I went from calm, to surprise, to shock, to fear then to total relaxed acceptance of imminent death. Then the LHS pilot turned us left and we missed - my report said 20ft vertical and 50ft horizontal separation. The subsequent physiological reaction was intense! Point being, when you know you are actually about to die, it’s not as terrifying as you would expect!

[spornrad]

https://globe.adsbexchange.com/?icao=a1abe2&lat=47.899&lon=-122.047&zoom=14.6&showTrace=2022-11-18&timestamp=1668795559[/url]

I note, with alarm, a "groundspeed" of 176 kts (one knot faster than Vmo),.. !
In that dive the ADS-B GS reaches 197 kt in the dive. So the airspeed was probably faster and beyond Vne.

[Pilot DAR]

In that dive the ADS-B GS reaches 197 kt in the dive. So the airspeed was probably faster and beyond Vne.

And this is the part of the equation which is alluding me a little... The airplane moved forward around 3000 feet while losing around 7000 feet of altitude, so more down than forward. If the "forward" is the base of the triangle, described as "groundspeed", and peaked at 197 kts, and the "down" is the altitude lost, the motion along the hypotenuse of the triangle would represent the actual airpseed?

It is very sad for the crew, the last part of that ride was terrifying!
I’m sure, although possibly less so than one might imagine.

Yes terrifying. Very sadly, there will be no memories of the crew. That said, having experienced my own terrifying experience when a training flight went wrong, I can say that I have very little recollection of it at all. Only enough, to know that yes, it did happen to us. That, and another extremely traumatic experience a couple of years back, which I know I witnessed in its full horrifying duration but do not recall at all, have me convinced that the mind simply turns off observation when things get really really bad. I hope this for the crew, as I would for any person experiencing sheer terror, with zero opportunity to mitigate it. The movies show us an increasingly terrified pilot hurtling toward earth and a certain death. That's the movies, I think (hope, anyways) that our actual life experience has some kind of built in protection for our emotions.

[megan]

I can say that I have very little recollection of it at all. Only enough, to know that yes, it did happen to usTake off from an oil rig, 125'ASL at rotation, massive explosion and glass screen of engine instrumentation turned seemingly all red, memory ceased at that point, memory regained on climb out when other chap asked "do you want me to cancel the audio tone", not aware of it until he asked, nor aware that he had put out a mayday during the dive for the water to gain airspeed.

[RatherBeFlying]

Using two significant figures and 3,600 seconds per hour:

3000' ≈ .49 nm
7000' ≈ 1.2 nm
GS ≈ 200 kt

Time in seconds over horizontal 3000' ≈ 0.49 / 200 x 3600 ≈ 8.8

Speed over vertical 7000' ≈ 1.2 / 8.8 x 3600 ≈ 490

Speed down slope ≈ (200^2 + 490^2)^½ ≈ 530 kt

The structural failure may have happened before achieving this speed.
​​​​​​

[Pilot DAR]

Thanks RBF, I did my modest math, and got very similar results. I figured I must have made an error along the way, and set my results aside. Like a few other planes I've dived, the Caravan will build up speed very quickly when pointed down (as opposed to a few floatplanes I've dived, when once pointed down, I had to add power to achieve the required speed).

The two photos in the preliminary NTBS report do show the airplane descending vertically, probably already broken up, and they can probably validate the speed with those photos.

[First_Principal]

Not quite the same but in the Pacific GA forum TWT has posted this report (https://www.atsb.gov.au/media/news-items/2023/weather-training-and-procedures-among-considerations-going-investigation-cessna-210s-flight-break) which discusses what appears to be structural failure of a 210 wing in flight...