What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Don Hall Accident Probable Cause

Jamie

Well Known Member
Don Hall was a good friend of mine, just as he was to many RV folks around the Atlanta area and beyond. His tragic accident last September took him and his friend and left a couple of wives and more than a handful of children without their fathers.

Don's accident shook me hard. I hope that we can all discuss his accident with the usual VAF courtesy, reverence, and respect for these beautiful families.

Now, having said that, the NTSB released this week their probable cause report. I have been waiting for this report to come out for quite some time. I know that often times the moderators (rightly so) shut down threads speculating on accidents, but those accidents are seldom ever brought back up to be discussed again. So I think that in the spirit of safety we should open a frank and earnest discussion on Don's accident. Don's tail number was N517DG. Here's the probable cause report in it's entirety.

NTSB investigators either traveled in support of this investigation or conducted a significant amount of investigative work without any travel, and used data obtained from various sources to prepare this aircraft accident report.

Witnesses observed the airplane maneuvering in the area before the accident. One witness reported seeing the airplane spin “wing to wing down” before it disappeared behind trees. The airplane impacted terrain in an upright, nose-low attitude in a residential area. Examination of the wreckage revealed no evidence of a preexisting mechanical malfunction or failure of the airframe or engine that would have precluded normal operation. Recorded engine and performance parameters revealed that the pilot reduced power and slowed the airplane while above 3,000 feet indicated altitude. Subsequently, the airplane began an uncontrolled descent. Engine performance indications were increasing during the seconds before impact. It is likely that when the pilot reduced engine power, the airspeed dropped below maneuvering speed, which resulted in an aerodynamic stall.

The National Transportation Safety Board determines the probable cause(s) of this accident to be:

The pilot’s failure to maintain airspeed while maneuvering, resulting in an aerodynamic stall and collision with terrain.

I am dissatisfied with the report. I'm not saying that the final conclusion from the report was wrong, but that the information made available to the public was shallow and useless -- the entire point of these reports is to improve safety.

[edit: I did not realize that the NTSB's docket system with all of the pertinent data existed]

Don's RV-7 had an AFS 3500EE installed with data logging turned on. This is the most detailed information that we get from that unit? My lowly Dynon stack in my RV-7A recorded far more.

There is no mention of heading data or attitude in the factual report. The only two parameters it mentions are MP and altitude. Does the AFS unit not record these parameters?

I want to know *what* maneuver Don was performing when he presumably stalled and spun into the ground.

I think we get more information from Don's APRS track. He was running APRS when he crashed but packet delivery was sporadic.

[ed. Thank you for posting this, Jamie. Looking forward to any discussion that can make my/our flying safer. v/r, dr]
 
Last edited:
Reading through the various dozens of pages of data, they did in fact have a whole bunch to look at. I believe in one of the documents it says they had about 51 parmeters (at least that's what is in this one particular document) HERE.

Most of the time I find the narratives almost entirely useless, but all of the data behind the reports is usually much more informative to me as a pilot. Of course rarely do you get an outright explanation of what happened, but sometimes you can formulate a decent extrapolation.

I also noted the radar plots HERE and HERE appear pretty close to the APRS for what it's worth. There are many other public documents and photos out there, but some are just not appropriate for posting I'm afraid.

Anyway, just my 2 cents as usual.

Stein
 
Last edited:
more APRS data

I won't speculate on the way the pilot flew the last few moments of the tragic flight, but here is an APRS composite image that may shed some light on the accident:

N517DG.jpg


There was one other beacon but it apparently occurred earlier in the flight and is most likely plotted due to a server error. I have inserted the data bubbles for what I believe are the last two beacons. The aircraft impacted the ground near the Presbyterian Church which would be a direct bearing from the last two beacons. The speeds shown are ground speed (53kts and 35kts), I don't know what winds aloft were at the time of the accident.
 
Last edited:
Wow....Stein thanks for the links. Those graphs tell a story that show at one point this aircraft experienced some sort of electrical anomaly. Something appears to have pulled the voltage down so low that the GPS providing position to the AFS system must have rebooted. I don't know where his shunt and hall effect sensors were in the system but notice how the currents trended went in opposite directions. I can't believe the narrative makes no mention of this.
 
Last edited:
Thank guys for the links.

For those engineering minded types like myself that like to look at raw numbers I've posted the data dump from the report here (CSV format).

I did not previously realize that this docket site existed. Thanks, Stein!
 
Something appears to have pulled the voltage down so low that the GPS providing position to the AFS system must have rebooted. I don't know where his shunt and hall effect sensors were in the system but notice how the currents trended went in opposite directions. I can't believe the narrative makes no mention of this.

This is mentioned on page 7 of the document in Stein's first link. There are several anomalous readings at this single point in time so it could be attributable to any number of things and I wouldn't focus too much in the voltage drop. It could have simply been a loose connection somewhere.

Having said that, I do find the anomaly very intriguing as well.
 
All this data and still a head scratcher.
Perhaps practicing stalls?
The 1min 30 seconds of slowing down with low MP, good FP, and generally maintaining altitude.... Head scratcher
 
Last edited:
Reading through the various dozens of pages of data, they did in fact have a whole bunch to look at. I believe in one of the documents it says they had about 51 parmeters (at least that's what is in this one particular document) HERE.

Most of the time I find the narratives almost entirely useless, but all of the data behind the reports is usually much more informative to me as a pilot. Of course rarely do you get an outright explanation of what happened, but sometimes you can formulate a decent extrapolation.

I also noted the radar plots HERE and HERE appear pretty close to the APRS for what it's worth. There are many other public documents and photos out there, but some are just not appropriate for posting I'm afraid.

Anyway, just my 2 cents as usual.

Stein

The data indicates what happened but why it happened may never be known.

A voice record up to the departure might answer that question but we do not normally record our conversations in these airplanes. It does help answer investigation questions and is required by law on commercial flights.
 
All this data and still a head scratcher.
Perhaps practicing stalls?
The 1min 30 seconds of slowing down with low MP, good FP, and generally maintaining altitude.... Head scratcher

My thoughts exactly, Mike. That or some sort of incapacitation.

The factual report states that he tested negative for carbon monoxide, but the toxicology report indicates that the specimens were not suitable for testing for carbon monoxide. I'm not inferring that this was the cause of course, just an apparent discrepancy I noticed in the reports.
 
Last edited:
One thing that is not clear to me is who was flying the plane. Were both sticks installed?
I don't read much into the electrical glitch, as the EFIS was not affected. I tend towards Kahuna's stall practice theory. Looks like the plane was being slowed deliberately.
The reason I ask about who was piloting is because recently I took another pilot up who used to fly a Kitfox. I had him slow the plane to demonstrate the stall. When it stalled, he jammed the stick forward. We went negative g and I had a face-full of mother earth. I took control and recovered, but it sure surprised the heck out of me! He explained that you had to do that in the Kitfox to get airspeed. I was very glad I had an excess of altitude.
Couple something like that with cockpit confusion and a secondary stall/spin...
Thoroughly brief any non-RV pilots on stall recovery technique and who is PIC if things go "pear shaped"!
 
Sobering Accident. Looking at the data stream in the last minutes of activity
, it appears a throttle back, slow flight, straight ahead, then power on and still pitch up, like a power on stall, all pretty normal. Engine making power through out data collection. It looks like something a normal (even a) student pilot does in training. Just plain ordinary.

This 7 has the tall rudder.

Is there any W&B information or fuel quantity information that would guide us to caution? I did not see that in the text, maybe missed it.
 
Sobering Accident. Looking at the data stream in the last minutes of activity
, it appears a throttle back, slow flight, straight ahead, then power on and still pitch up, like a power on stall, all pretty normal. Engine making power through out data collection. It looks like something a normal (even a) student pilot does in training. Just plain ordinary.

This 7 has the tall rudder.

Is there any W&B information or fuel quantity information that would guide us to caution? I did not see that in the text, maybe missed it.

The question I have is would someone normally practice slow flight and stalls at 2000agl. I prefer at least 3500 feet agl.
The actual NTSB investigation is typical. Their case load is large and they simply don't spend much time or effort on this type of accident.
George
 
I cannot so easily discount all the data that points to an electrical anomaly and at the same time a GPS reboot. Too much of a coincidence for me. Something significant happened and shortly after the accident.
 
This is mentioned on page 7 of the document in Stein's first link. There are several anomalous readings at this single point in time so it could be attributable to any number of things and I wouldn't focus too much in the voltage drop. It could have simply been a loose connection somewhere.

Having said that, I do find the anomaly very intriguing as well.

Saw that...I was referring to the short probable cause report. Sorry for the confusion.
 
Trying to put myself in the situation.

Everything is fine. I'm flying along in a plane that I built and have many hours of experience in. I'm flying with a friend and we're talking about flying or football or whatever. Its a beautiful VFR evening and nothing is wrong. Just as I am crossing the shoreline of Lake Lanier, airspace that I am intimately familiar with, something happens. It seems to be a power blip. After a few seconds, everything seems to be fine except my GPS...my really expensive GPS. In a flash of thought I'm wondering if my insurance is going to cover a power blip...what a pain in the butt its going to be getting that thing out and off to repair. Then I notice that the black screen is coming back on. It seems to be rebooting; I'll check it out later. The flight continues.

A little more than 5-minutes later my power starts to fade. Not suddenly, but slowly over the course of about 20-seconds, or so, it then stabilizes for 30-seconds and then slowly starts to fade even further.

At this point I am wondering if the engine doesn't stop all together. At 19:12:40, or so, look at the fuel flow. It is very nearly zero. There is even one tick mark that is actually zero (or so it appears zoomed in on my monitor). During this time HP is also zero. At 19:13:55 Fuel flow is restored...but not to it's full amount. Manifold pressure is max, oil pressure is max, rpm is max but fuel flow is less than half of the amount recorded during the parts of the flight where we assume all was fine. It seem incongruous. Refresh rate of the fuel flow sensor? I dont know.
 
Couple of questions:

  • Do we know what type of ignition he had? If it was all electronic, would it fail if there were a system power failure (even momentary)?
  • What happens in an RV if it enters a spin? How easy to recover? Could a pilot recover who has not got a spin endorsement in any aircraft?
 
Sure looks like a normal stall was performed, followed by an application of power and a dive to build airspeed, then perhaps into a loop where an accelerated stall occured then a removal of power to recover. Or he could have been reducing power on the back side of the loop to keep the speed from accelerating too quickly.
 
Last edited:
Sticking to the data analysis . .

Trying to put myself in the situation.

A little more than 5-minutes later my power starts to fade. . . .

At this point I am wondering if the engine doesn't stop all together.

Looking at the tabular data . . . .

What data is accurate? Not sure, but related engine data seems to make sense all taken together. It seems that the throttle was pulled back, but not prop speed. Manifold pressure drops, egt and cht lag, but drop too. Probably not fuel flow problem, pressure is still up, consistently except for high hall amp event. NTSB photos appear to show electronic and mag ignition. Was the throttle pull back because the engine was running rough? Maybe, but the egt does not indicate a step change in the data stream at that time, and egt can change pretty quick. No conclusion except that data (fuel pressure, fuel flow, EGT, CHT, Man P, RPM) is consistent with a running engine.

What might cause the 43 amp jump in the hall effect sensor? A sensor wire issue, or real high amp short. The latter would seem plausible but the battery voltage did not seem to drop (odd that it is always at 8.5v) so did all the amps come from the alternator? Regardless, the event seems to be real or the GPS would not have reset. Electrically minded analysts should consider. 8.5 batt voltage maybe should be thrown out.
 
Couple of questions:

  • Do we know what type of ignition he had? If it was all electronic, would it fail if there were a system power failure (even momentary)?

The data indicates that EGT was unchanged after the electrical event which points to the engine still making power.
 
Bill,

Maybe that steady volt reading is the backup battery and not the main bus???

Since there are two volt readings this may be the case. I am just looking at the chart not the raw data (on iPhone).
 
Last edited:
It's easy to get distracted by things like a lat/long dropout...both here and in the air. Always better to concentrate on what is required to fly the airplane.

Take the raw data Jamie supplied and throw out everything that isn't relevant, or does not change through the whole timeline. Then re-arrange the rest into a sensible format, and start thinking about what would generate those values.

25i17aw.jpg
 
Last edited:
Thats just it. What was the cause and the effect of the electrical glitch? It is real and it happened. The logging system caught it. Everything is normal electrically and BAM something happened that caused what appears to be main battery discharge amps to go thru the roof. Main bus voltage dropped drastically. The GPS that was being logged apparently rebooted.

Few minutes later this accident happened...

Was this pilot demonstrating stalls and or spins and it got away from him or was he distracted by something else going on and ended up in a spin and it took him by total surprise?

Unfortunate for me, I did not know this pilot other than his postings here but from looking at his log book it seems he was experienced. Looks like the amount of flying he was doing had slowed a little. What would cause a pilot with this experience level in a RV end up in such a spin?
 
Last edited:
I don't know if you guys have ever tried to recover from a power on spin in an RV but I would suggest not doing it for fun. I tried one day and gave up because of how much buffet/shake I was getting on the tail. Plus the roll rate was outside my comfort range.

His throttle was still working from the reports I have read leading me to think of 2 situations. Ground coming quickly and trying to force it out of a spin or passenger was fighting him for control. Sad either way.
 
I am still confused by the low altitude slow speed flight. It just does not make sense to me. I also am curious if anyone knows if the aircraft had AOA or some form of stall warning. I often hear discussion about a good pilot not needed any stall warning in a RV but if there was in fact a aircraft malfunction this is a classic reason why AOA or some form of stall warning is a great safety improvement. Accidents are never just one thing. They are always a chain and usually a long chain of events. There are so many different distractions that can occur and take someone out of their comfort zone into a area they do not normally fly in. The results are often not good.

George
 
Be careful flying near sunset. Our eyes take approximately 20?30 minutes to fully adapt from bright sunlight to complete darkness.

Possibly one very important factor...Sunset was 7:46 PM that day. No moon illumination at that time. Over water. What a perfect time for taking a picture off the left wing. Was there an electrical/electronics distraction? Was there an EFIS power switch switched off accidentally just prior? Was the sunlight or eye adaptation interfering with instrument readability? How was the cockpit lighting? Were sunglasses being worn? Who was flying? Was the correct pedal pushed to counter the spin? I know for sure I don't practice stalls in these conditions. How would I perform with a surprise stall/spin? I thought a lot about this accident and the family for several months after. I saved the flying picture posted of him and his wife. It is one of many constant safety reminders for me.
 
Caution: Speculation Ahead

Looking at the data I initially see a (edit: removed the word deliberate) stall - power off, speed slowing, holding the nose up but still getting a moderate rate of descent.

The aircraft stalls and the wing drops somewhat. Perhaps my normal stall recovery response in the RV is just add full power and fly out of the stall (I could do that in the RV-4). However I am surprised by the amount of wing drop and I try to correct by opposite aileron (a pro spin action). Now I am not only stalled, but am well on my way to a full blown spin. Then I add full power (another pro spin action at this stage). If I then push forward on the stick because I need airspeed this would likely flatten the spin.

It is very possible the RV will not recover from such a flat spin, even if I understood what I had inadvertently done and applied every possible anti spin action.

The book, Stall/Spin Awareness, by Rich Stowell, is good reading for anyone interested in this subject. IMHO every pilot should have an interest in this subject. I am not related in any way to the author other than I took some Emergency Manuever Training from him.
 
Last edited:
After review this AM I've made two corrections to the data table in post 22.

The previous line 19 note said airspeed had been dropping steadily for 30 seconds. Bad observation, and misleading; 30 seconds was just the time period for decreasing airspeed shown in this abbreviated data table. Review of the original data shows that power reduction and the beginning of airspeed reduction was actually 2:40 prior to stall and 1170 feet higher. Although possible, it is not a typical profile for an intentional stall demonstration.

The previous line 32 suffered from a clerical error; I shifted data one column when arranging the presentation, then assumed the odd numbers were an artifact of impact. Not so. The corrected numbers now make sense, and bring to light a new fact...the pilot had recognized the power-on error and had pulled throttle just before the data point was recorded.
 
Rate of descent too high for spin?

I have spun my RV6 from a normal entry and a cross controlled entry for training. I always reduced power to idle as the first action for recovery - I am sure this is what is required in all a/c as the first step to proper spin recovery.
I understand that there is a big difference between a planned spin entry and one the pilot isn't prepared for.
My question is with the rate of descent as high as it was, could this really have been a spin? Would the plane descend that fast if the wings are stalled?
My first guess would be a steep spiral.
C. Brenden
Commercial, Instrument
VAF dues current
 
I have spun my RV6 from a normal entry and a cross controlled entry for training. I always reduced power to idle as the first action for recovery - I am sure this is what is required in all a/c as the first step to proper spin recovery.
I understand that there is a big difference between a planned spin entry and one the pilot isn't prepared for.
My question is with the rate of descent as high as it was, could this really have been a spin? Would the plane descend that fast if the wings are stalled?
My first guess would be a steep spiral.
C. Brenden
Commercial, Instrument
VAF dues current

I think the 2.4 g's on the G meter and 90 degree bank supports this as well.

This thread describes what was very likely a spiral descent (not a spin) resulting from a botched stall:

http://www.vansairforce.com/community/showthread.php?t=86829
 
Last edited:
My question is with the rate of descent as high as it was, could this really have been a spin? Would the plane descend that fast if the wings are stalled? My first guess would be a steep spiral.

Let's explore the possibilities. Here is the data without any notes or conclusions:

5le81u.jpg


And two notes from the text report:

The wreckage was found upright, on a heading of 360 degrees....Other than the area of initial ground impact, there was no linear ground scar.

It would really help if we knew more about the AFS system. For example, does the "Pitch" data refer to nose degrees below the horizon regardless of roll position, or a rotation around the pitch axis of the airplane? And is the "Roll" data reliable given a possible high rotation rate?
 
Spins

Disclaimer: I have zero experience in spinning RV's.
One of the best kept secrets in aerobatics is that a flat spin can be stopped very quickly with full power in all of the high performance aerobatic aircraft I have flown. This includes Pitts S1S, S1T and S2B, Sukhoi SU26 and SU29 and several others.
A fully developed upright flat spin in the Pitts can be stopped in less than 1/4 turn, probably closer to 1/8 turn. The Sukhoi stops marginally faster.
I timed the late Bob Herendeen a couple of times performing a 33 turn inverted flat spin in the Pitts S1S. 58 seconds starting from the time the spin flattened out. This is much a much slower rotation rate than some of the "expert" aerobatic books claim.
The rate of descent in an upright or inverted flat spin in a Pitts or Sukhoi does not come close to correlating to the rate of descent in this accident.
There is only one spin scenario that POSSIBLY correlates and that is a full or high power setting accelerated spin. The issue I have with this posibility is that in the Pitts, Sukhoi and other high performance aerobatic aircraft, the pitch angle is much closer to vertical. I have never tried to time the rate of rotation of a sustained full power accelerated spin, but I can say for certain that the rate of rotation and the rate of descent is very high.
I worked my way up to seven or eight turns in the Sukhoi. I would not do more than three turns in the Pitts with full power because of the streses on the airplane.
The accelerated spin for training purposes in the Pitts goes like this:
normal upright power off spin entry, spinning to the right is preferred. After the normal spin is established, move the stick forward to slightly forward of neutral. Full forward stick is advocated by some but i never saw much difference with this technique. The rotation immediately increases dramatically. The indicated airspeed, which has been very low, increases to around 90-100.
Add full power and the rate of rotation increases further.
Yurgis Kairys did 17 1/2 turn accelerated spins in the SU26.The entry was a forced entry slightly above normal stall speed, and the airplane accelerated vertical down to near Vne which is around 270 statute. I tried to time him by counting, at around 7-8 turns I would start falling behind in the count.
That led me to believe that the rate of rotation as the speed increased was in the range of 720 degrees per second.
I no longer do aerobatics but I am a mostly self taught aerobatic pilot. I taught myself flat spins in the Pitts S1S. I did not do accelerated spins until much later. I don't know why the accelerated spin surprised me because i had been doing up to 7 turn snap rolls starting in horizontal flight and losing 1-200 feet as the airspeed decayed. The horizontal snap roll is essentially an accelerated spin on a horizontal axis. The difference in the accelerated spin is that the airspeed initially increases and then stabilizes with power at idle, with full power the airspeed continues to increase.
This is the only possible scenario I can think of for this accident, if the pilot was seeing this scenario for the first time he had virtually no chance.
As a benchmark I entered my early flat spins from 11-12,000' AGL. With a trainee in the Pitts I liked to be at 5,000' AGL, higher was better.
The accident pilot was in my opinion much too low to be performing an intentional stall.
 
I agree with JRS... unintentional spin (flat or accelerated) and a failure to reduce power to idle to get out of it. The first time I flat spun a Pitts S2C was after a botched hammerhead and I could not get out of the spin until my instructor had me reduce power to idle.
 
Spins

Regarding the statements by C. Brendon and Skylor that this was a spiral, I have several issues with that scenario. The airspeed would have continued to increase in a spiral. A spiral should be easily recognizable and easily recoverable. The low airspeeds do not support the spiral theory.
The low airspeeds also do not quite correlate to a sustained accelerated spin. There are several explanations for that, the most likely being that the pilot recognized that moving the stick forward made the situation dramatically worse so he reverted to stick full back which slowed the airspeed slightly but not to the normal spin rotation rate. Also the airspeed indications in an RV7 in flat and accelerated spins are likely a complete unknown. This is an area that is simply not explored in aircraft that are primarily transportation airplanes with basic aerobatic capabilities. The dividing line to me is snap rolls. Most RV pilots who do aerobatics are loop and roll types who don't do snap rolls. This to me is a very wise choice. If you have an overwhelming desire to do snap rolls get a Pitts.
 
the roll changes with pitch oscillations, were those aileron rolls? on the first page of data in the ntsb reports....looks like he may have been doing those over the water (common place for aerobatics)
 
He was at both a low airspeed and low altitude to be doing even basic rolls. I wish the NTSB had looked more into this accident. It would be nice to have more information on the background of the pilot and the aerobatic training he may have received. Sometimes a accident turns out to be something unexpected. Perhaps he wanted to show a simple stall and the right seater got scared and without knowing it stepped on a rudder. Perhaps someone was not strapped in properly and with a little negative G large inadvertent control inputs were made. It does appear that the aircraft should have been recoverable based on the airspeed and G available near the end. That assumes however you don't have unexpected control inputs.

George
 
We have the probable cause from the NTSB and more data than we usually see, and we're still speculating. So I'm going to add to the circle of speculation.

Here's a thought I've noticed only one other poster mention in passing. Did the NTSB calculate a CG for the flight situation? Could he have been demonstrating slow flight at a more rearward CG than he'd flown since his 25/40 hour flyoff period?

I haven't tried to dig into the detailed NTSB files so it may very well be there. I've seen them calculate W&B in other investigations and think it should be included in this one. If it's there, someone please point me to it.
 
Like Don Hull, I will throw in my interpretation of the data.

I see the flight beginning with some bank and yank fun, a little straight an level with smooth turns and then..... a smoothly decending practice approach of some sort starting at 4200 ft that ended in a full stall that turned sour (referencing the earlier spin commentary by JRS).

Sure, one could say "that was done too low". But, I am not 100% convinced another 1000 feet would have made the difference at the high rate of decent and rotation if there was lack of recognition of the solution or interferance with the controls.

Since I have not done spins in over 35 years (ppl), it makes me wonder what I would have done in a similar situation (meaning an "unusual" spin situation out of a stall recovery).
It gives me the "willies" thinking about it. Maybe I will accelerate that basic aero training I have on the list (my intent has been to spin, spin, spin).
 
What indications would you get on your gages if you ran a tank dry then switched to the other tank?
I assume you mean in the data.
Perhaps nothing given the sample rates here.
But you would expect to see everything unchanged except for the engine data FP, EGT, RPM and FF etc.
You will note in this data set that the fuel quantity read high on both tanks.
 
Also if the slip angle is correct?

The angle of sideslip is a definitive discriminator between a spiral and a spin.

The spiral is more or less coordinated. The slip would be small, as shown here in these data (assuming the units are right, 0.2 degrees is very small).

A spin has usually 10-20 degrees of sideslip. Thats why it takes full rudder to hold "most airplanes" in the spin, and recovery follows quickly with release of the rudder or application of opposite rudder. And airplanes that are slow or difficult to recover often suffer from a fair degree of fin-rudder blanking by separated flow from the fuselage, caused by the large sideslip angle, as well as from the horizontal tail, caused by the large angle of attack.

Looking at the data here, if we understand the measurement basis and accuracy of those numbers, we may be able to definitively answer the question.
Units matter too. If the slip angle in this data sheet is actually in Radians (0.2 radians = 11.5 degrees) , then this would be a serious slip and a sure indication of a spin. If it is correct in degrees, then this is an insignificant slip angle, and I would say spiral.

So, if someone could dig into the measurement basis? (what I mean is, what sensors, located where, based on calculations from other measurements, measurement units, Stuff like that)

Let's explore the possibilities. Here is the data without any notes or conclusions:

5le81u.jpg


And two notes from the text report:

The wreckage was found upright, on a heading of 360 degrees....Other than the area of initial ground impact, there was no linear ground scar.

It would really help if we knew more about the AFS system. For example, does the "Pitch" data refer to nose degrees below the horizon regardless of roll position, or a rotation around the pitch axis of the airplane? And is the "Roll" data reliable given a possible high rotation rate?
 
Last edited:
The data tells a story.....

.....in 5 second intervals.

It took 36 seconds using Dan's spread sheet to slow the aircraft to stall speed, actually it was longer I am sure if previous data were shown.

At 19:13:53, or slightly earlier, the nose pitched down 8.3 degrees and the aircraft rolled some 30 degrees. The airspeed had been holding steady at 58-59 knots for the previous 10 seconds. Pitch attitude was increased some 3 degrees sometime after 19:13:43 which finally induced the stall. It would appear this part of the flight was planned as heading, altitude, pitch and roll were fairly constant, about what you see when practicing or demonstrating a stall.

When the nose dropped 8.3 degrees, the pilot began to advance power, rolled the aircraft back to near wings level (32 degrees to -8.3 degrees) but induced a secondary stall by applying excessive back pressure pulling 2.4 G's at 93 knots. (perhaps the passenger panicked and hauled back on the stick - we will never know) When the aircraft stalled the second time, the nose dropped to near 60 degrees and the aircraft rolled some 90 degrees. It all happened in less than 5 seconds. By 19:14:03, airspeed was 119 knots, some 10 seconds after the initial departure. The aircraft was in a spiral, held there with aft stick pressure, and crashed.

If the excessive back pressure had not been applied, the aircraft could easily have been recovered from the initial stall, it had flying speed and altitude to spare, they were still some 1200' AGL when the secondary stall occurred. It all happened very quickly.

I believe this is the story. I do not believe there was any intention to spin the aircraft.
 
Reading the data I would have to agree with you David. I still don't understand doing stalls at that altitude if it was intentional. I also think there may have been some unwanted control inputs.

George
 
a few observations:
the default recording interval in the afs is currently preset at 5s. i suggest to set that to 1s when right away when setting up your efis.
it's much easier to take any meaningful data from the records as many things like runups, leaning, stalls, takeoffs, landings, rolls and even loops can completely or partly happen within very few such 5s slots. add some (actually very nice) filtering on to the sensor data and the log covers even a longer timeslot. data storage volume is no factor and we're running 1s interval since 200hrs, dumping the log of the last flight to the sdcard as SOP in the checklist at shutdown. i think the 5s default is for longer term recording / less write cycles / back from the days when sdcards cost a fortune and processor speed/capability was more limited.

as to the gps lat/lon freeze:
it is not quite clear from the report, which gps source the data was coming from. the afs certainly appears to have correctly logged whatever it received/did not receive.
in case of the afs gps (garmin gps18) in the default config, it loses track quite quickly when going inverted/covering the antenna with something like a kneeboard/doing maneuvers it isn't able to track/follow. also, i've had it take quite a long time to reacquire after a loss. also, knowing whether the gps was wired through the afs efis expansion plug or some other 5v source could rule out some failure modes/options/effects by the electrical system.
at this point, i wouldn't read tooo much into this gps anomaly IMHO, at least it would appear there is very little connection to the actual crash later on.

regards,
bernie
 
It took 36 seconds using Dan's spread sheet to slow the aircraft to stall speed, actually it was longer I am sure if previous data were shown.

Corrrect. The throttle was retarded 2 minutes, 40 seconds prior to stall. Over that period of time the aircraft steadily slowed from 151 knots and descended 1170 feet.

Allowing the altitude loss doesn't seem to suggest an intentional stall.

If the slip angle in this data sheet is actually in Radians (0.2 radians = 11.5 degrees) , then this would be a serious slip and a sure indication of a spin. If it is correct in degrees, then this is an insignificant slip angle, and I would say spiral.
So, if someone could dig into the measurement basis? (what I mean is, what sensors, located where, based on calculations from other measurements, measurement units, Stuff like that)

Yes please...pitch, roll and slip.

Steve, I agree with your slip data observation....surely not in degrees. A review of all slip data prior to departure from controlled flight finds only one time point where the slip indication climbs into the tenths, 0.11, half the value seen in the final 15 seconds. Almost all of it is 0.00 to 0.02. There's not much point in measuring slip in hundredths of a degree, and anyway, what pilot can fly that precisely?
 
Last edited:
The aircraft stalls and the wing drops somewhat. Perhaps my normal stall recovery response in the RV is just add full power and fly out of the stall (I could do that in the RV-4). However I am surprised by the amount of wing drop and I try to correct by opposite aileron (a pro spin action). Now I am not only stalled, but am well on my way to a full blown spin. Then I add full power (another pro spin action at this stage). If I then push forward on the stick because I need airspeed this would likely flatten the spin.

Pushing forward will never help produce a flat spin unless you're spinning inverted. In an upright spin, pushing forward will always accelerate the spin and cause it to spin more nose down - the opposite of what happens when a spin flattens. Power will have a flattening effect during a spin, but only if spinning left upright. But the forward elevator will overpower the flattening effect of power, and what you'll end up with when power is applied and forward stick is held during an upright spin is essentially a vertical down snap roll. I guarantee you in this configuration, a quick opposite rudder input would stop the rotation nearly instantly.

I don't remember trying power-on spins in my RV-3, but by adding forward stick and accelerating the spin, it would stop nearly instantly once opposite rudder is applied - much faster than the typical stick-back spin. I've done a number of spins in an RV-6 with the 7 tail, and it spins and recovers normally.
 
Last edited:
A long, long time ago I owned a Cessna 150 with some other guys and we set up a one airplane flight school just for kicks.

This particular airplane was purchased out of a wreck after repair and had a nasty habit of dropping the left wing when doing simple straight ahead stalls. If the wing drop was not caught immediately, the aircraft would begin to spin. I always commented, good training for a student. Not very funny actually. We should have had the rigging adjusted to correct it but the rigging supposedly was set properly. The wing may have been slightly twisted. It was a surprise for those not knowing it was coming.

The same thing can be induced by rudder misuse at the stall departure and may well have been a factor in this accident, no question, a wing dropped when it stalled. That would have been a surprise if there was no intention to go into a spin.

2:40 at idle power before the stall - I wonder what that was all about? It may have simply been a sloppy stall entry loosing over a 1000' slowing the airplane. The RV will fly forever in a slight descent right at stall speed.
 
One thing I think you guys are missing is that its impossible to pull 2.4G at -60 degrees of pitch, unless inverted. Assuming that the AHRS will correctly note negative pitch when inverted. This is why I have the impression he was doing a loop.

In a loop if one stalls on the upside or downside in an RV it will spiral but all you have to do is let go of the stick and the stall will break.
 
Huh?

One thing I think you guys are missing is that its impossible to pull 2.4G at -60 degrees of pitch, unless inverted. Assuming that the AHRS will correctly note negative pitch when inverted. This is why I have the impression he was doing a loop.

In a loop if one stalls on the upside or downside in an RV it will spiral but all you have to do is let go of the stick and the stall will break.

At -60 degrees pitch the aircraft is a .5g until you pull on the stick. Then the g load will increase as you pull the stick until an accelerated stall. What do you mean you can't pull 2.4gs at -60 degrees?
 
Back
Top