Re-visiting the 1999 Hawk 200 Crash

[Fonsini]

I'd like to hear the thoughts of experienced pilots, especially Hawk qualified pilots, on the chain of events that led to Graham Wardell's tragic death while displaying a Hawk 200 in 1999.

A video of the event is linked here:

SIAD 1999 - Hawk 200 (+ sunday's fatal crash) - YouTube

And the summary of the crash investigation findings determined that:

========================================

"On 6 June 1999, a BAE Hawk 200 aircraft crashed during SIAD '99 air show at M. R. Štefánik Airport, Bratislava (BTS/LZIB: co-ordinates 48°10′12″N, 017°12′46″E). Top British test pilot Graham Wardell was killed in Slovakia after his Aerospace Hawk 200 military jet came down during aerobatics.

The plane exploded when its wing hit the ground after failing to pull out of a low turn at a show at the Milan Rastislav Stefanik airport in Bratislava; the test pilot entered a barrel roll too low, the roll was too tight, and the exit speed was too fast, and at the wrong angle.

ZJ201 suffered a high speed stall, struck the ground and bounced clipping a building, killing a woman spectator. It bounced again and finally struck the ground and was destroyed. The pilot did not eject and was also killed".

========================================

The strange "pitch" wobble just prior to the crash never looked quite right to me - was there ever a suggestion of control failure, "pilot error" is always the easiest, but not always the most just verdict.

[Sun Who]

What prompts your query?

Sun.

[Fonsini]

Just a desire to gain a better understanding of what happened, why it happened, at what point the maneuver went wrong, and how the resultant crash could have been avoided.

[Sun Who]

Fair enough, but I doubt you'll get anything meaningful here (in terms of useful insights) - beyond what was said in the official report.

Regards,

Sun.

[Fox3WheresMyBanana]

The only other Hawk (T1) LL barrel roll crash I am aware of was NMacC's in January 1988.It may be of use.

Aircraft by type (http://www.ejection-history.org.uk/aerobatic/Red_Arrows/RED_ARROWS.htm)

There was quite a bit of detail about how the manoevre was flown in the original accident report, if you can get hold of it, but I cannot remember that detail now.

[Craven Moorhed]

I was stood next to my jet watching this live in Bratislava. It was horrendous to watch such a nice chap lose his life having been chatting to him just beforehand.
I have a few Hawk hours and sadly have to agree with the high speed stall theory - the wing rock you see (and the rumble i heard while watching) is typical of the Hawk at high AoA in heavy buffet. On the day the wind was much stronger than the practise the day before; not sure if this may have affected his sight picture at the top of any manoeuvres by being in a different position over the runway??

[EGNH Flyer]

Graham Wardell had survived a previous accident back in 1979, on the 18th July. 14 Sqn Jaguar XX960 hit a TV mast in poor weather removing the entire wing assembly, Graham ejected and was very lucky to survive. I believe the location was Iserlohn, Germany.

[Courtney Mil]

Only my opinion, but it looked like he was too low (possibly a missed "gate") at the top of the last manoeuvre and had insufficient altitude to pull out. It looks like he was pulling very hard at the end there, possibly having realised he was too low/too nose down. Possibly not recoverable.

I agree with Craven, it looks like he then pulled too hard, causing the stall symptoms, but I don't think that was the cause of the crash, he was already too low/fast/steep. Step back to the start of that last manoeuvre and not pulling up enough and there's the point it started to go wrong. It may have been possible to abort the manoeuvre up to the point where he commits nose low. After that it was all history.

Hadn't really thought about this one for a long time. Just my thoughts as you asked for them. So very sad to lose a good man.

[Fox3WheresMyBanana]

CL's crash (Phantom at Abingdon 1988) was IIRC influenced by having to extend upwind at the top of a LL display loop. NMacC's crash was (again IIRC) possibly recoverable in his previous aircraft (Lightning) but not in the Hawk.

Both of these may be relevant after Craven's notes.

[Lima Juliet]

I was also standing next to my jet watching this on that day. It looks like he flies a "loaded roll" as he rolled over the top which also left him critically low on his practice (which I also watched the day before). I remember that the GR1 display team had a good chat with him over a beer afterwards and they discussed the requirement to unload during the rolling sequence to keep his gate height. Sadly, the rest is history...:sad:

What was even sadder was that an airport worker had smuggled his other half onto the airport and she was sitting on top of a small building on the display side. The wreckage of the Hawk landed on her - killing her. Absolutely tragic and a lesson re-learned why the display side is kept free of spectators.

So in my opinion, what might have killed Graham?

1. Not unloading during the roll over the top and then burying the nose when below gate height. You could hear him pulling through the light buffet to the heavy buffet as he realised he had run out of room to pull out.

2. I seem to recall that Graham was not a display pilot but was the company roll-demo pilot/test pilot. I suspect that he may not gone through the same rigorous work-up process that a display team normally does for a pre-season display authorisation (I might be wrong on this though).

3. The Flying Control Committee (or equivalent in Slovakia) should not have let him display after his pre-airshow display performance (again, in my personal opinion). But that is hindsight.

4. Finally, as Slovakia had just started to think about joining NATO and the company that Graham was working for was on the 'hard sell' for Gripen and Hawk. Could it have been that he was rushed into doing this display by the company to increase their sales potential?

Again, this is all my thoughts on the matter and I have no hard evidence to prove any of it. Therefore, my comments are without prejudice.

LJ

[500N]

A question if I may ask.

Why wouldn't he have eject at the last minute ?

I assume he knew he it was going in.

[Fox3WheresMyBanana]

This is why I mentioned the other LL barrel roll. It is possible that the situation may have been recoverable in a previous aircraft with which the pilot was more familiar. It also may well have been the case that, after passing the inverted position, the aircraft was continuously thereafter outside ejection seat limits. They aren't a magic wand, and the zero height - zero speed seat depends on zero sink rate and zero bank angle, neither of which are present in the second half of a barrel roll.

[500N]

Thank you, great explanation.

[Lima Juliet]

PS

The strange "pitch" wobble just prior to the crash never looked quite right to me - was there ever a suggestion of control failure, "pilot error" is always the easiest, but not always the most just verdict.

This is him pulling light buffet to heavy buffet to light buffet to heavy buffet. The poor bloke probably realised his fate after the first pull on the stick. At his second pull he would never had made a successful ejection if he'd tried as his sink rate was too high. Watching from the ground at the time this went in slow motion through my mind as I was pretty sure as he rolled in the vertical that he was not going to make it. Sadly my thinking was correct as it played out in slow motion before my eyes.

It is one day (and evening) that I will remember for the rest of my days. It taught me a lot about people's reactions in the aftermath of a crash and just how hot an aircraft crash fireball is even when you're standing 500-700m away. It also saved my life a couple of years later when a student started doing a loaded roll during fighter affil at about 1,500ft in the low flying system. I can remember yelling 'unload' to this very day as we subsequently dished out at a couple of hundred feet above the ground! The video debrief was interesting...:eek:

[Lima Juliet]

Fox3

Concur - you are faster at typing than me!!!

[Fox3WheresMyBanana]

I note your comments on the practice LJ. A crew from our Squadron were at Ramstein in 1988 with a static. The JP told the experienced nav he was going to watch the Frecce display from crowd centre. The nav said "No you aren't. I saw their practice yesterday. They are going to crash. We'll watch it from way over here."

[Fonsini]

Amazing insights from all of you, I knew the knowledge would be here.

I would however like to apologise, especially to those of you there on that day, if this thread brought back some bad memories.

Now I have to go away and research the difference between a loaded roll and an unloaded roll.

[Lima Juliet]

Fox 3

I think I would be buying that Nav a beer every 28th August for the rest of my life!!! :D

LJ

[Fox3WheresMyBanana]

The Lightning and the Jaguar (I believe Graham flew the Jag in the RAF) are classic swept wing supersonic fast jets where, when the angle of attack goes past what would be the stall in a straight wing aircraft, drag increases massively but lift remains, and may even increase. If sufficient thrust is available in a clean aircraft (e.g. reheat), recovery may be possible. The Hawk, IIRC, does lose lift in the heavy buffet, and of course does not have reheat.

There is a tendency under severe stress for the human brain to do what it knows best, rather than what is correct. Thus pilots can "revert to type" and fly the way they did in the aircraft they know best, rather than the way they should in the one they are in.
I am not saying this is what happened, but it is something investigators consider.


LJ - Beers - I believe he did!

[tartare]

Well you learn something every day.
I'd previously thought zero/zero seats were able to get you out in any situation other than being inverted below a certain height.
Is the zero sink rate restriction due to not wanting to overload the canopy on opening?
Don't understand the zero bank angle restriction either...

[Fox3WheresMyBanana]

They are not restrictions, but limits on effective performance. It's just the physics of it. To get technical, the ejection seat gun/rocket pack give the seat approximately a constant impulse. In practical terms, this means a set change in the velocity of the seat (say, 100mph change). If the seat is designed to go from zero to 100 mph with zero sink rate, then if there is already an aircraft sink rate of -20 mph (minus meaning downwards) an upright seat will go from -20 to +80 mph. Still a 100 mph change, but the first 20mph change was used to overcome the sink rate. Since the seat is moving more slowly upwards than designed after the gun/rockets fire, it won't get as high. If the aircraft is banked, some of the velocity change will be sideways, which is no use in an ejection above a flat surface. For example, with 30 degrees of bank, our seat will only gain 100*cos(30) = 86.6 mph upwards, so will therefore also not reach the desired maximum height for safe parachute opening.

The effects are cumulative. If the aircraft were sinking and banked as above, the vertical speed would go from -20 + 86.6 = + 66.6 mph. - a lot less than the + 100 mph with zero sink rate and bank (and you really don't want a number with 3 sixes in it! ;)).

[DITYIWAHP]

A very sad event indeed.

Ninja'd by Fox3...

All modern zero-zero seats are just rockets that accelerate the pilot/seat combination to a speed whereby the chute will successfully open if ejection is initiated on the ground; watch the testing videos and you'll see that man-seat separation starts with an upward vector ie at max speed after rocket firing ceases.

A small bank angle on the ground (eg due to a gear collapse) will reduce the upward throw but will probably not prevent the chute from deploying successfully. The problem with high rates of descent is that the time from handle pull to the chute opening is something like 1.5 to 2.0 seconds (generic) and then the chute has to slow the pilot down enough so that a soft ground 'impact' can be achieved. During that time you can fall whilst decelerating quite a long way. The conservative rule of thumb is to eject when your height is above 0.1 x rate of descent (in feet per minute). So at 300 KIAS in the vertical (approx 30000 ft per min) the min safe ejection height will be 3000 ft AGL. The aircraft document set will contain specific details about rates of descent v heights for a successful ejection.

[Fox3WheresMyBanana]

DITYIWAHP is right about the time delays being very important too. In fact, on the Harrier, in the hover, the delay due to human reaction time is so critical that the advice used to be to eject immediately if there was any red warning caption. It may have been a harmless OXY caption, but in the around-a-second it took to look in and read an OIL caption or similar major engine problem, the aircraft would have built up enough sink rate to kill you, ejection or not.

In the Air Defence world, the sink rate calculation means that if are doing a 'post-hole' manoevre in air combat, going straight down supersonic, you are pretty much stuffed if you don't start recovery by 10,000 ft, as even the seat can't save you. Definitely not zero-zero.

[tartare]

Wow - sobering reading.
Thank you for the explanation - makes sense now.

[Lord Spandex Masher]

F3, didn't the Harrier have a bigger rocket because of that, or better than zero zero?

[John Farley]

LJ

Thank you for your post 10.

The first para says it all and entirely accords with my understanding of the events surrounding the accident.

JF

[Tashengurt]

Never worked on Harriers but I think later ones had a rocket assisted drogue?


Posted from Pprune.org App for Android

[John Farley]

I wrote an internal company memo after the first Hawk 200 accident way back in Feb 87. I would not change a word today. Sorry about the formatting.

Our Hawk 200 Accident

I have obtained a copy of the official accident report on the F20 display rehearsal accident, that happened in Canada en route to the Paris Air Show 1985. The report shows at least 10 remarkable similarities with our Hawk 200 accident.

1. Both pilots intended to pull up abruptly at high G in order to gain height, and then to obtain displacement from the runway by rolling as required, before turning back onto the runway heading.

2. Both pilots pulled the stick back when descending inverted, instead of when the aircraft was correctly banked.

3. Both reports discussed the effects of motion on the pilot as a likely cause of disorientation.

4. No fault was found with either aircraft.

5. Neither pilot had been given formal instruction on combating the effects of G, even though it is now available.

6. Both pilots were well practised and current in the routine.

7. The weather was ideal on both occasions.

8. No transmissions were made by either pilot associated with the final stages of the flight.

9. Neither pilot tried to eject although both could have done after the display sequence had first gone wrong.

10. The manoeuvres of both aircraft could be established by film or flight recorder.

I have copied the report widely in the hope that it will help senior management associated with this year's Paris participation appreciate the many different things that can be done to assist our display pilots to fly effective but safe displays. I am taking this opportunity to list below some points for discussion. I am sure others will be able to think of many more.

• Ensure that the display is properly developed, have it agreed and then fixed. Keep the pilot current.

• Obtain the best available advice on the effects of high G and rapid rolling because these are potentially greater with the latest breed of aircraft. If the best aeromedical advice referred to above recommends training on a centrifuge, do it.

• A ban on general test flying and navigation equipment, not needed for the display, being taken into the cockpit by the pilot.

• Formally establish a team to assist the display pilot, consisting of:

a. Another pilot.

b. A Flight Test engineer to acquire, process and log data on all
flights and make the records available as soon as possible.
(Video,Film, ADR, Voice etc)

C. The crew chief on the aircraft.

d. A marketing representative with the responsibility for all liaison with customers/press/company VIPs and to assist as the point of contact with the display team for all admin aspects of the rehearsals and actual displays. This person will enable the other members of the team to concentrate on executing the displays.

Only in exceptional circumstances should a rehearsal or display be flown without the full team being present.

[Lima Juliet]

JF

Wise words mate. It always really saddens me when highly respected and competent aviators 'buy the farm' when we all know they are more than capable of flying the manoeuvre in a less pressured environment. After Graham's accident, the next I was witness a couple of years later was Ted Girdler's sad demise off of the beach at Eastbourne. At the time I was probably at the peak of aviation skills as both an A2 OCU instructor and also on the BBMF. To see such magnificent individuals waste their lives in this way was sobering indeed.

We should all use this to be more wise after their untimely demise to try and stop it happening next time. I always think that financial pressure and aviation is a far greater risk than the human factor alone. Skill cannot always save us and luck is definately on the side of those that have astonished audiences at airshows/displays in the past. Sadly for Graham and Ted their luck ran out.

LJ

[Shaft109]

Some of the previous posters have suggested that had the type been a Lightning or Jaguar or similar then the accident MIGHT have been avoidable, so without complicating my questions due to the gate positioning -

1. So is it correct that a more classically swept or delta would be able to hold a higher alpha over the straighter Hawk wing, with a proportional increase in induced drag?

2. At this specific point in the routine I understand the extra thrust of those types would mean that they had excess thrust available to overcome or balance the drag of high alpha but the Hawk couldn't due to having less thrust - they could have 'powered through' the drag?

3. I've seen the phrase mentioned on here that the 'Barrel roll' has claimed lots of pilots and is an underestimated manouver, - is there any specific reason why proximity to the ground being an obvious one?

Thanks in advance.

[Pittsextra]

Just a desire to gain a better understanding of what happened, why it happened, at what point the maneuver went wrong, and how the resultant crash could have been avoided.


3. I've seen the phrase mentioned on here that the 'Barrel roll' has claimed lots of pilots and is an underestimated manouver, - is there any specific reason why proximity to the ground being an obvious one?

Its quite disappointing reading all the complex chat in this thread because the accident is simply summed up. The pilot crashed because fundementally he didn't understand a barrel roll.

If you are not inverted at the top a barrel roll you will loose height, the problem that causes depends upon how much altitude you have to play with. The solution is simple, you just roll wings level.

As I said its a surprise that this accident happened and perhaps reflects military pilot training.

[pontifex]

I was just about to post Pittsextra when you said it all for me!

[Bob Viking]

Really? You want to go there?! Let's see where this goes.
BV:rolleyes:

[Pittsextra]

Really? You want to go there?! Let's see where this goes

Really what? Really that's not how a barrel roll works out, or really because I said it?

I mean see where what goes? It can't go anywhere because how I've explained a barrel roll is how it is. The end. The picture at the "top" needs to be fully inverted. If it doesn't then if height loss could be an issue then stop pulling and rolling wings level is your friend.

You see it differently??

[Bob Viking]

I don't doubt your understanding of the barrel roll. I am calling into question your apparent disdain for military pilot training. I am also questioning your need to criticise a guy who made a fatal error. I'm sure if you'd been there you could have taught him to do it better and we wouldn't be having this conversation.
I'm guessing that you are well versed in aerobatic manoeuvres from your handle. Are you equally well versed in fast jet handling?
Do I really need to continue?
BV:ugh:

[Pittsextra]

Sorry Bob but you misunderstand me. There is no disdain for military pilots and far from it. That's said there is little value to be gained from making some false god from military fast jet pilots, RAF and/or associates. In the end there by the grace of God go us all. We can all err, make mistakes, whatever, we could just as easily be talking spinning accidents with T67's couldn't we??


As for criticising the pilot in question I'm sorry but in the end (and sad to say) if he understood a barrel roll he wouldn't be dead that's the only conclusion. If by saying that it causes deep offence or upset then it doesn't change the fact. I'm not sure what my experience of a fast jet or otherwise changes that element.

[Lima Juliet]

Pitts

I would think it safe to say that he understood a barrel roll - the pilot in question was a graduate of the Empire Test Pilots' School (ETPS), the very first Brit to fly the F117 Stealth Fighter (well before it was publically acknowledged), he flew Hunter, Hawk, Tornado, Jet Provost, Jaguar and a myriad of others including small aerobatic aircraft. As I understand it he was flying a display routine at low level in front of a crowd with tight display restrictions that he hadn't practiced very much.

Did he understand the barrel roll? - absolutely, as an ETPS graduate he could probably write you 10,000 words on the subject. Could he fly it safely? On a good day, yes, however he was probably having a bad day. He had been tasked at short notice to put something together - a pseudo role-demo come display without the normal full work up process. He paid for it with his life (and sadly someone else's).

Overall, Graham Wardell was a exceptionally skilled pilot who had a very bad day at the office...

I think that is why you have got BV's back up! :ok:

LJ

[Pittsextra]

Hello Leon. I wasn't trying to pick a fight, the OP asked why it happened and at what point it went wrong and I was just explaining that in simple terms.

I’m sure Graham Wardell was a legend behind the stick. I’m sure he was a lovely guy and I mean and meant no offence to his memory, to those guys who knew him, worked with him and so on, but the thing that killed him wasn’t type experience, fast jet experience, crowd line, display restriction or role demo experience.

The thing that killed him is that at some point in that display the picture was wrong, he didn’t recognise it and as you say sadly he and someone else paid for it with their lives.

You say he understood a barrel roll, could write 10000 words on the subject as an ETPS graduate. Given all of that to some it may seem therefore that the humble barrel roll would hold no challenge, however whilst we wish to pay due respect to that experience and ability you have to reflect on what happened on the day.

I’m not sure what his process was for the display but in that part of the sequence the picture should have looked wrong and there would be a plan? That's normal. We can dress it how we like but in the end the mistake was simply not recognising that the picture was wrong and that pulling wasn’t going to save the situation.

I guess that is what I meant by saying he didn’t understand a barrel roll.

Does anyone have the official report into the accident as I’ve not seen it and I’d be interested to see how it was framed officially.
:ok:

[Lima Juliet]

Pitts

Yeah, but, no, but...If he had flown the routine a few more times at 5,000ft base height then he may have made the same mistake a couple of times (having dished out at 4,500ft) and corrected it. When flying a fast jet the margins for error are far less - having flown Pitts, Extras, Chippies and a few aerobatic gliders in my time, compared to a fast jet it feels like a knife fight in a telephone box! If you bury the nose to 90 nose down at 250kts in a Hawk you will need at least 4,000ft to pull out without smacking into the ground or having to change your g-pants for clean ones. In a Tornado make that closer to 6-8,000ft! :eek: Compare that to a Pitts then at 90 nose down at 180kts (no idea of Vne!) I would guess you need about 1,000-1,500ft? Have you ever flown low level aeros in your Pitts, the ground can seem awfully close as you max perform to meet your minimum stipulated CAP403 height?

As Fox 3 pointed out better than I, he was in 'coffin corner' as he let the nose drop below gate height. Even Martin Baker's finest couldn't save him once he started pulling hard to the heavy buffet.

I agree that he 'porked it' on the practice day and even more so on the day of his demise. He fully understood the discussion about the 'loaded roll' with the other Display Team that he discussed it with. Indeed they were absolutely gutted when he crashed and felt dreadful that he did it again. Sadly, he did it again - but how many civvy aero pilots debrief their sequence and then go on to make the same mistake again as they refly a compex series of manoeuvres according to their Aresti notes? You're spot on me old with 'to err is to be human'. :ok:

BTW, never seen the accident report. One wonders if it was 'sat on' due to the commercial involvement and as it was mil registered jet in a foreign country then it wouldn't have been AAIB material in 1999 either.

My other lasting memory of the event was I had an overwhelming urge to run towards the fireball; goodness knows why, as there was nothing I could do. The lass from the local embassy, who was standing next to me, fainted - that gave me something else to distract me and stop me being a silly-arse in trying to run to the fireball!

LJ

[Lima Juliet]

PS. Pitts - here is a Typhoon missing the ground after a max performance pull-out - there was nothing left but the ~30ft and disaster.

Eurofighter Typhoon near miss at RIAT - YouTube

[Bob Viking]

I didn't know the guy involved but from what I understand he was a pretty competent aviator (deliberate understatement). The bit that I'm struggling to comprehend though is that you seem to infer that your understanding of the humble barrel roll is greater than his.
I can assure you that us lowly FJ pilots are taught, and fully understand, the rigours of various aerobatic manoeuvres, the barrel roll being one of them. Throughout training I remember being reminded on numerous occasions how it was one of the most dangerous manoeuvres to fly.
No matter how awesome we may be (or think we are) though it can never stop us from making a mistake. Trying to claim that, had he possessed your zen like level of barrel roll acumen, he'd be alive today just seems a little obtuse to me.
Still, it's possible we are just failing to understand each other repeatedly so I'll give it a rest.
BV:cool:

[glad rag]

A very sad event indeed.

Ninja'd by Fox3...

All modern zero-zero seats are just rockets that accelerate the pilot/seat combination to a speed whereby the chute will successfully open if ejection is initiated on the ground; watch the testing videos and you'll see that man-seat separation starts with an upward vector ie at max speed after rocket firing ceases.

A small bank angle on the ground (eg due to a gear collapse) will reduce the upward throw but will probably not prevent the chute from deploying successfully. The problem with high rates of descent is that the time from handle pull to the chute opening is something like 1.5 to 2.0 seconds (generic) and then the chute has to slow the pilot down enough so that a soft ground 'impact' can be achieved. During that time you can fall whilst decelerating quite a long way. The conservative rule of thumb is to eject when your height is above 0.1 x rate of descent (in feet per minute). So at 300 KIAS in the vertical (approx 30000 ft per min) the min safe ejection height will be 3000 ft AGL. The aircraft document set will contain specific details about rates of descent v heights for a successful ejection.


Pretty comprehensive instructional video...US biased but relevant even today I would say...


Oxi_rqAW9vM

[Pittsextra]

Bob - Can you tell me how flying a barrel roll in a fast jet differs from flying one in a Pitts Special, Extra 300, Chipmunk, Moth, Stampe, Yak, Sukhoi or Spitfire?

Fundementally at the top of the manoeuvre how does the picture look? Does it look differently?

Actually it's kind of relevant in 2014 because the current Typhoon demo has a barrel roll as figure 6, so what conversations are had to prevent accidents here? I might offer the suggestion that the pilot says if the picture doesn't look like this I'll roll wings level. What is your suggestion?

Leon I hear you re sink rates but the point I was making is it should never have got to that position and actually i dont think you suggest that it might.

Ultimately his DA was signed off and he would have no doubt made the sequence he flew.

Bob it's got nothing to do with me thinking I know better than him, I'm better than him, I'm being smug or anything of the sort. Things go wrong and when you look at the Hawk crash he was not fully inverted at the top. That's the beginning of the end.

Just like how some World class aerobatic pilots have sadly died by making poor weather decisions or RAF instructors getting killed spinning T67's. Don't shoot the messenger.:ok:

[Bob Viking]

Pitts.
I tell you what. Once you have read the accident report, read the students study guides from EFT, BFJT and AFT (which will all, if memory serves me correctly, tell you that the wings should be level with the horizon at the inverted position) you should find that he did indeed know that he should have been fully inverted at the top.

As has been mentioned several times already, he made a mistake. For whatever reason he didn't realise in time. It would appear that commercial pressure played its part also.
Maybe you could submit an amendment that could be incorporated into all of our training manuals to ensure future generations have a better understanding of barrel rolls. Sarcastic I know but I'm just not getting your point to be honest. I think you've already had the answers you need so I'm not really sure why you're persisting.

Personally I never enjoyed aerobatics as a student. My aerobatic sequences always involved the bare minimum of manoeuvres and were unspectacular to say the least. Occasionally now I fly some if I have some spare fuel to burn. To be honest I don't fly barrel rolls very often since I don't like them. Slow loops, slow rolls and vertical aileron rolls are usually enough to sate my appetite.

BV
Broke my own resolve to give it a rest.

[Bob Viking]

As an aside. Which Firefly incidents are you talking about? I recall one crash but no deaths.
BV

[DITYIWAHP]

Pitts,

https://dl.dropboxusercontent.com/u/184657779/misc/hawk%20crash%20top%20of%20roll.png

As you can see from this first fame of the video where the display manoeuvre is shown, only a person with super-human perception can determine the aircraft's angle of bank when it was at the top of the barrel roll (which you seem to think is important). Unless you have a different source of evidence, it appears that you are making your conclusions solely based only on your own preconceptions and limited experiences. Your technique of assessing whether the wings are level when inverted will ensure your display looks pretty but it is not a useful metric for ensuring you maintain safe recovery options when nose down in a fast jet. Frankly your 'expert' advice makes me question your expertise.

Angle of bank at the top of a barrel roll can be adjusted very quickly and, consequently, it is not a direct measure of whether or not any sort of rolling vertical manoeuvre will be successful - a quick adjustment of angle of bank can easily be made at any time later, so angle of bank when inverted is almost a meaningless check of whether the rest of the manoeuvre will be safely flown in a jet. You have to ensure that the nose is not lowered too much below the horizon for your available height, whether doing aerobatics or fighting in air combat.

In fast jets the critical factors for nose-low recoveries is the available g that the aircraft can achieve. Many aerobatic manoeuvres in jets like the hawk are flown near to (but not at) the CL Max, so there is some margin for increasing g - but not much (in the hawk at least). Having flown the Extra and Yak52 as well I can vouch that these slower types of aerobatic aircraft have the potential to increase their g quite substantially above that required for typical rolling manoeuvres, because of their light wing loading; thus the margins you might be used to are greater than those afforded to an average jet pilot who commits his nose below the horizon.

When I taught barrel rolls to students on the hawk I taught that it was important to always stay cognisant of the remaining height below the jet in combination with the jet's pitch attitude and bank angle / roll rate. There are rules of thumb available as a guide to how nose low you can commit yourself for any given height and these are padded for rolling vertical manoeuvres like barrel rolls. Similar techniques apply when flying a rolling scissors (I have also taught many of these...) - you always have to be aware of your pitch attitude as well as your angle of bank. The jet in the video attempts a 'floor save' by ceasing the display and rolling wings level, but the lack of height for pitch attitude was identified far too late, sadly.

The bottom line: when you say "when you look at the Hawk crash he was not fully inverted at the top. That's the beginning of the end" you're actually saying "I have no idea what I'm talking about".

[John Farley]

Chaps

I look at these two Hawk accidents from a slightly different perspective to some posters.

It seems to me that the common factor with both was that the pilot pulled while at the wrong bank angle.

At the point in the first accident routine where it went wrong the pilot rolled to 180 deg bank and pulled hard where normally he only rolled to about 110-120 deg bank before pulling hard round in a descending turn. Why he did this concerned me because apart from my test flying and display background I was the manager of Dunsfold at the time of the accident. Remember it was an instrumented aircraft and the records survived so we knew exactly what happened.

I also witnessed the event and saw the aircraft roll a little one way and then a little the other (not wing rock) while it was going close to vertically down. I speculate that at that point the pilot realised all he could see was ground and not the horizon and was trying to find the nearest horizon.

I had seen the display countless times before and it never gave me the slightest concern.

All of which leaves me wondering why the pilot did not realise he had the wrong bank angle at the moment of starting the pull. As we know one’s eyes are normally the most reliable of our sensors when it comes to a violent manoeuvre. However they can give you a false picture under some conditions, in particular if a bout of nystagmus is stimulated by the manoeuvre. I have a number of possibilities in my mind that could account for the first accident pilot’s disorientation but they are pure guess work so I am not going into them here.

I have been watching and later participating in airshows since 1950. After I saw the 10th aircraft fly into the ground I stopped counting. Of those first 10 only one had a technical issue. On three of the occasions I realised the pilot had lost it and told those who were with me that he was going to crash. On another in South America it was so obvious what was going to happen that I told those I was with that I was not going to watch and would go back into the chalet until it was all over.

[Flap62]

Pitts,

As the previous two posters have stated, you may have rather over-simplified the "picture at the top" argument. Being purely inverted at the top is not the important criteria, it is the relationship between pitch and roll at that stage that is important. Too much pitch with too little roll will level you in big trouble, correctly inverted or not!

[glad rag]

Pretty sobering stuff there JF.

:sad:

[BEagle]

Flap62 wrote: Too much pitch with too little roll will level you in big trouble, correctly inverted or not!

Well, personally I'd say too little pitch attitude (above the horizon) at the inverted point is the real killer.... Because what happens next, unless it's a 'spot roll out' recovery manoeuvre is a nose-buried descent, with a dangerously steep pitch attitude and high descent rate once wings level is achieved.

Unless there's a lot more blue than green at the half-way stage, throw it away.

Was sun dazzle ever considered in this Hawk accident? Pull up, pitch and roll, roll and pitch and...!!!!, where's the horizon gone :confused:

[Flap62]

Each to their own Beag

I personally tend to say to myself "pitch" til heals go through horizon then introduce increasing roll, backing off on the roll as heals go through horizon on way down with increasing pitch. That said, my aeros are shocking!

It's like most of a good aeros sequence, balanced co-ordinate inputs.

[EAP86]

The BOI was convened as a D Flying 'PE fleet' investigation (a/c was operated under Def Stan 05-122 'COMA' arrangements) but the report may not have been published widely. ISTR (vaguely, it was a long time ago) that the view was that the barrel roll became steeper than intended and in effect became a loop from which there was insufficient height to recover.

[Pittsextra]

My apologies Bob re: T67's I was thinking of this and you're right thankfully no deaths. However I guess the point was that its an odd thing to happen given the experience and the task at hand. We digress.

http://www.aaib.gov.uk/cms_resources.cfm?file=/Slingsby%20T67M%20MKII%20Firefly,%20G-BUUH%2010-95.pdf

Flap and "used to fly harriers" - the cross referencing of pitch to roll was of course taken as given because how else are we barrel rolling..!??

However re my over simplification of the picture at the top, lets not take my word for it... Here are the words of someone to please everyone... RAF and aerobatic champion Neil Williams....

a very common fault with the barrel roll is to roll too slowly to start with so that the nose comes down through the horizon before the wings are level inverted. The result of this is a high speed spiral dive with considerable height loss.

Funny old thing that.

[Two's in]

Regarding the original question from the OP, the "chain of events" started well before the manoeuvre. The discussion here dissects the components of a correctly executed barrel roll, but a pilot of his experience would have been well aware of those. Display flying is littered with examples of errors of judgement driven by the very nature of public performance. The temptation to either make the display more "sporting" than the original DA, or to ignore airborne cues that the safe or authorized operating envelope is being exceeded is always there. That is why the practice and review are critical. This was a commercial display as part of a sales drive for the aircraft. Whatever the airmanship errors were in the cockpit, the environment for increased risk and display pressures were already present. No one will ever know if these were a significant factor in the accident, but it's probably safe to assume this accident was not caused by the pilot not knowing how to perform a barrel roll.

[LOMCEVAK]

Although Graham's accident was now many years ago I studied the report thoroughly at the time (and had access to the ADR traces) because he and I had been at UAS together and later were together on Experimental Flying Squadron at Farnborough. What I write below is from memory but it has always stuck with me.

Let's first consider some basics for a barrel roll. It is a manoeuvre that involves a simultaneous roll rate (generally quite low and significantly less than the maximumm achievable, espcially in an aircraft like a Hawk) and nose up pitch rate (ie load factor will be greater than +1g throughout). To fly one neatly in a display you try to to maintain a constant roll rate from commencing the roll until back at wings level erect. Therefore, any adjustments required for positioning the manoeuvre are achieved largely by adjusting the pitch rate/g although roll rate can be varied, especially for safety. The concept of a gate height as is used foor looping manoeuvres is not valid for a barrel roll because the minimum top height from which the manoeuvre can be safely completed will be a function of the roll rate and pitch rate during the second half of the manoeuvre. Also, if you are attempting to fly a barrel roll in a display positioned symmetrically about crowd centre the maximum nose low attitude on the way down will be dictated by the maximum pitch attitude on the way up.

In this thread so far there has, surprisingly, been no discussion regarding the orientation of the manoeuvre with respect to the crowd line (although this is not shown on the embedded video clip). Graham entered the manoeuvre from crowd left angled towards the crowd line at approximately 45 degrees and pitched up wings level at around +6g to approximately 60 degrees (numbers all from memory). He then commenced rolling left which turned the aircraft away from the crowd and he maintained at least 4g throughout. This generated a good angle away from the crowd but the roll rate that he used combined with such a high pitch rate resulted in a nose down attitude/altitude combination during the third quarter of the roll which was irrecoverable. I recall that the g values on the final manoeuvre were even higher than those on the practise that he had flown. However, the height achieved at the top would have been affected significantly by the nose up pitch attitude at which he started to roll and I cannot recall the difference between the practise and the accident sorties.

This set-up is inherently difficult to fly and requires large variations in pitch rate to achieve the required positioing. When you start to roll you need a large pitch rate to achieve the change in flightpath from towards the crowd to away from it. However, once the bank angle is about 120 degrees you need to slacken the pull or you become too nose low at the wings level inverted. During the second half of the roll you have to adjust pitch rate/g and possibly roll rate to have a final flightpath, which will always be angled towards the crowd line, that is clear of the end of the actual crowd area. Therefore, in the second half of the roll the pilot is concentrating on making control inputs to achieve the exit line (and thus prevent a crowd infringement) and could lose awareness of altitude and attitude. Overall, it can be quite an asymmetric manoeuvre when viewed from the cockpit and from the end of the display line although it looks OK from crowd centre.

This was Graham's first public display. He had worked the display up in an approved and supervised manner. He was still learning the art of display flying and had included an extremely tight barrel roll from a difficult set-up. I discussed informally my views on how this accident could have been avoided with the pilot member of the Board of Inquiry (and I did disagreed with at least one of their recommendations) but I will not expand on those here.

A barrel roll is a difficult manoeuvre to position well in a display. There are ways to do so safely and there are orientations, such as the one used by Graham, that involve a higher risk of CFIT. I am not being critical of the decision for him to fly this because I used the same during my first display season which was in a jet much heavier and less manoeuvrable than the Hawk.

Finally, reference has been made to NMacC's acccident. This was during a rollback which, although a simultaneous rolling and pitching manoeuvre, is not a barrel roll per se. It has a much smaller radius and faster roll rate and results in a relatively small lateral displacement. During this manoeuvre NMacC's aircraft departed from controlled flight which resulted in ground impact. Another tragic accident but one in which the considerations were totally different to the above discussion on barrel rolls.

Rgds

L