Monday, March 30, 2009

'Chute First ...


I've read with great interest some assertions about Cirrus aircraft, the pilots who fly them, and whether or not the airframe parachute makes the Cirrus pilot safer or just emboldens them to take risks. By now most of you know that on Sunday, March 15, 2009, a 64-year old instrument-rated private pilot flying a new Cirrus SR22 elected to deploy his aircraft's airframe parachute shortly after takeoff from Montgomery County Airport in Gaithersburg, Maryland. He departed runway 34 and his plane came to the ground about a half mile from the airport, no injuries were reported on the ground, the pilot walked away, and the aircraft was substantially damaged. It's dangerous to generalize, but I feel compelled to make some observations about this particular accident, the pilot's decision to launch into the weather, the efficacy of Cirrus door latches, and under what conditions a Cirrus pilot should consider deploying the 'chute.

The pilot involved in this particular accident was reported to have had 320 hours total time and I'm assuming, given his age, that he came to flying later in life. None of the reports I've read give any specific numbers, but given his total time and the fact that his last certificate was issued in June of 2007, it seems reasonable to assume that he didn't have much experience with solo flying in IMC. A low time pilot with a powerful and capable aircraft can be a dangerous, sometimes even deadly combination, and this accident would seem to reinforce that belief. And let's be clear that while this accident happened to involve a Cirrus, most any brand of high-performance aircraft will do.

An important part of instrument training involves making a competent go/no-go decision. Heck, it's explicitly called out in the Instrument Rating PTS as something the candidate must demonstrate. I sat on the ground a few years back with a Cirrus owner while we waited for the weather to clear. The radiation fog was thick and the surface winds were gradually starting to increase and mix out the fog, which made waiting all the more uncomfortable. But wait we did because, parachute or not, the conditions did not meet my minima for departure.

Turn in your hymnals to 14 CFR 91.175 and you'll find specific departure weather restrictions for aircraft operating under 14 CFR 135 and 121. Taking off under conditions with zero visibility and zero ceiling is not expressly forbidden when operating under part 91, but that doesn't mean it's a good idea nor does it mean that if you do so and you run into problems that you won't be scrutinized for violating 14 CFR 91.11 (Careless or reckless operation) - endangering the life or property of another. In our Me First society, it is easy to forget that our actions may indeed have adverse effects on others. This is where an instrument instructor's job of teaching risk management begins.

The accident pilot elected to launch with a reported ceiling of 400 feet and 2 miles visibility. Shown above are the takeoff minima published for Gaithersburg, which don't specify any ceiling or visibility. That means the 14 CFR 91.175 standards of 1 mile visibility for aircraft with two engines or less apply to part 135 and 121 operators. Technically the accident pilot was not prohibited from departing since he was operating under part 91.

The absolute lowest personal departure minima for a single-engine aircraft that I recommend to pilots I train for the instrument rating are pretty simple: The surface weather observation must be equal to or better than the highest circling minima (ceiling and visibility) for the airport, just in case an emergency return is required. In a twin-engine aircraft, I'm still pretty conservative and recommend the conditions be no lower than the highest straight-in minima of all non-precision approaches available at the departure airport.

When I flew freight in the Caravan, my company's procedures allowed us to depart in some really crummy conditions. On several occasions, I departed when the greater Bay Area was blanketed fog and with low IFR conditions at all nearby major airports. And you know what? It gave me the creeps every time I did it.

I've never flown into Gaithersburg, but a quick review of the available approaches show the following circling minima.

GAI NDB RWY 14 - 1 SM vis & 1380 feet MSL, 841 feet Height Above Threshold
GAI VOR RWY 14 - 1 SM & 1200 feet MSL, 677 feet HAT
GAI RNAV (GPS) RWY 14 - 1.5 SM & 1020 feet MSL, 481 feet HAT

In case you're wondering what I'm getting at, the low-time instrument-rated accident pilot took a pretty big risk when he chose to depart with 2 miles visibility and an overcast ceiling of 400 feet at an unfamiliar airport.

I've written before about my experiences with the door latches on a Cirrus SR22 that I used to fly. Quite frankly, I found the performances of these door latches stinks. Cirrus, in an apparent quest to make the aircraft seem as much like an automobile as possible, tried to implement a slam-and-shut-style automobile door. This just in: A high-performance single-engine aircraft is not a car. My experience showed me that the latches on an SR22 G2 must kept adjusted just right by a mechanic and the pilot had best ensure the doors are secured, top and bottom, before taking off. Interesting, the door latches on an older SR20 that I used to fly had a very positive door mechanism with a latching handle.

So a door popping open on a Cirrus is not uncommon and the AFM even has a procedure for handling it - abort the takeoff if you can, otherwise reduce your speed and land as soon as practical. A door popping open can be distracting as hell, especially to a low-time pilot, but the slipstream will keep the door mostly shut. You just need to reduce the airspeed and return to land. Of course, returning to land is going to be a lot easier if you at least have circling minima.

When I flew the Cirrus regularly, I followed all the recommended Cirrus Airframe Parachute System (CAPS) procedures. This included removing the safety pin from the activation handle before takeoff and installing the safety pin after landing. If you don't remove the pin, you simply can't be ready to deploy the parachute quickly in an emergency. I've read of several fatal accidents involving Cirrus where NTSB investigators, combing through the wreckage, found the CAPS safety pin firmly in place on the deployment handle.

Even though I followed the CAPS procedures and I regularly reviewed the deployment procedures, my mindset when flying the SR22 was that CAPS deployment was going to be an absolute last resort. The AFM gives some suggested situations where CAPS deployment is warranted:
  • Mid-air Collision
  • Structural Failure
  • Loss of Control
  • Landing in Inhospitable Terrain
  • Pilot Incapacitation
After the door opened, the accident pilot reported that his intention was to turn back and land at Gaithersburg. The accident pilot says the plane entered an unusual attitude and he let the airspeed get low, the aircraft stalled and started to enter a spin. The accident pilot said he had pressed the magic button (the autopilot Level button) to get the plane stabilized, but decided he couldn't wait for the magic button to do its magic. He was also concerned about entering restricted airspace nearby and was unfamiliar with the Gaithersburg airport environment. So he pulled the 'chute.

I'm glad he's okay and that no one on the ground was hurt, but this all seems so preventable. Low time pilots in high-performance aircraft with airframe parachute systems can learn a lot from this accident. "'Chute first" is a potentially dangerous and definitely expensive procedure. The hard questions need to be asked and answered on the ground, before the clouds are approaching, the door has opened, or the engine has quit and you feel the urge to pull that T-shaped handle.

12 comments:

Vincent, from PlasticPilot.net said...

John,

I'm no Cirrus pilot, but my experience is quite similar to this of the accident pilot (350 total time, used to high perfromance singles like Saratoga and Bonanza).

My personal take-off minimas are mostly same as yours: something that makes a return possible, prefereably with a circuit in VMC.

Two other remarks: airspace, and "magic button". While I understand the concern of the accident pilot about the nearby airspace, I would not care for it when in such a situation, which looks like an emergency. Pulling the chute - a destructive thing - to avoid airspace ? Second thing, the "level" button of the autopilot. Correct me if I'm wrong, but this is available only on the new version, the Cirrus Perspective, which has been around for something like one year ? This brings one question to me: how much experience had the pilot with the Cirrus ? Did he flew previous versions as well ?

As you see, not all "low-time" pilots are ready to take the same risks...

Dave Starr said...

Great advice there, John, the circling approach minima idea was new to me and is definitely a 'keeper' ... it's much easier to make decent decisions if you have a yardstick to measure with.

I also hadn't done the research on the Cirrus accidents where the safety pin was found still inserted after the crash. Years ago when i was a ground crewman for a USAF squadron, part of our runway safety procedures (done at the end of the runway just before the pilots received take off clearance, and appropriately called the "Last Chance" inspection) included the pilot removing the safety pin from the ejection seat trigger and displaying to the inspector on the ground near the cockpit. According to the POH he was then supposed top stow the pin in a special in flight storage pouch on the side of the cockpit wall.

Rumor has it that many aircrew members would reinsert the safety pin in the seat trigger mechanism after showing it, as the seat in that particular aircraft was only rated for ejections above something like 100 knots and 200 feet AGL ... unlike more modern 'zero-zero' seats.

The prevailing wisdom was, better to leave the pin in until well above the speed/altitude 'red zone'.

One day one of our birds suffered a violent engine disintegration of the left engine which also took out the right engine by 'collateral' damage ... as in shrapnel and flying debris.

Speed and altitude was well into the safe ejection envelope, but the pilot elected to dead stick into a farmer's field. He made an excellent controlled gear up touchdown but, with no directional control, was unlucky enough to collide with a huge oak tree in the center of the field, sadly with fatal consequences.

I think by now you can guess my punch line. When the first responders arrived on the scene their number one in-cockpit safety checklist item did not have to be performed. The safety pin, dutifully shown as removed at the end of tunway check, was now firmly back in place, rendering the ejection seat useless.

We'll never know. Did he decide to ride it in from the beginning of the emergency? Or did he die becuase in the few seconds of time available that a safe ejection was possible, he was too busy to fumble around to to find and remove the pin?

Not a Cirrus pilot (yet) and unlikely to be flying any ejection seat equipped heavy iron, but my thought is, if Uncle Sam or the aircraft manufacturer recommends the pin be removed, it probably makes sense to follow the book.

Anonymous said...

He is alive.

John said...

Vincent,

The NTSB report says the accident pilot had owned an Avidyne-equipped SR22 prior to acquiring the Perspective SR22 that was involved in the accident.

Anonymous, to me your "ends justifies the means" argument is simplistic and not at all compelling. This accident only resulted in damage to property, but it could have been a lot worse.

Rick Beach said...

John, your blog was forwarded to me by a friend because of my involvement with the Cirrus Owners and Pilot Association as their safety liaison.

Appreciate the conservative risk-mitigation tone of your post. However, one aspect deserves comment.

Your "mindset" that CAPS would be "an absolute last resort" may have killed a number of Cirrus pilots because they too thought the same way. By my reckoning, about half of the fatal Cirrus accidents (28 of 48) involved decision points similar to successful parachute pulls. That is, the accident pilots persisted in recovery or flying the airplane well into your stage of last resort -- only they never pulled the handle and crashed and died.

Until we acknowledge and appreciate the pilots who survive what you call "preventable" accidents, we are likely to perpetuate the deaths of people in Cirrus airplanes who did not use CAPS as a last resort.

The 53 people who died in those 28 accidents far outweigh the 30 people who have survived a CAPS pull.

Cheers
Rick

John said...

Rick,

Thanks for stopping by.


The problem is not the aircraft, it's the pilot's decision-making process. Even Cirrus acknowledged this (finally) when they upgraded the Avidyne MFD software to display a bunch of risk management questions that the pilot must click their way through. This was clearly a way to reinforce the idea to Cirrus pilots that they are not invincible.

I consider CAPS deployment a last resort primarily during the flight planning phase. For example, a hypothetical preflight train of thought might go something like this:

"Gee, the ceiling and visibility are pretty low for an IFR departure, but I can always use CAPS if I get into trouble."

Take CAPS out of that equation and you (hopefully) get a no-go decision. I think you'll agree that it's hard to wreck the plane and injure innocent people if you stay on the ground.

I agree that if you wait too long and a CAPS deployment is indeed the last, best resort, you may have screwed yourself. It's very easy to exceed the recommended parachute deployment speed. CAPS is just another system and like all systems, it has limitations. Pilots need to know and respect those limitations AND they need to do a brutally honest assessment of their own limitations.

The way I "appreciate" people who do risky things or attempt things beyond their skill level is to examine their choices in a way that is instructive to other pilots.

I'm glad that CAPS is saving lives, but some of the accidents make us wonder - "What were they thinking?"

gschlumb said...

I've flown into and out of KGAI IFR a couple of times, although not since the most recent ADIZ change. Even though it's a small airport, it is a deceptively high-workload environment because of the looming presence of the ADIZ. You get passed between several different controllers in very quick succession, and unless you can run your avionics without a second thought, it can be a LOT for even a competent pilot to handle. The fact that he elected to depart into conditions which would have left him ONLY an LPV approach to minimums (with a presumed downwind landing) in order to return to the airport is somewhat questionable, chute or no. Being a pilot with roughly the same number of hours (mostly in a Mooney rather than a Cirrus) and a relatively fresh instrument ticket with minimal actual the first time I flew in and out of KGAI, I probably would have elected to stay on the ground. (Mostly because John beat that into me during PPL training.) Assuming that most of his Cirrus time was with an Avidyne system rather than the G1000 in the accident plane, how much did unfamiliarity with equipment contribute to his disorientation? Like the Cirrus accident pilot who elected to fly over the Sierras in IMC at night in what turned out to be icing conditions (http://aviation-safety.net/wikibase/wiki.php?id=44566), this pilot had minimal time in the accident aircraft. That should factor in a large measure when pilots are establishing their personal minima.

Matthew Stibbe said...

I'm a Cirrus pilot with a few more hours than the pilot in this case. I fly from a non-IFR airport just inside the Heathrow control zone and so we are restricted to UK SVFR rules - 3km visibility and 1000 feet ceiling as minima for departures and landings. In many ways, I think this acts as a useful brake on more 'ambitious' departures. That said, I'm sure many IR pilots have made departures that they have looked back on and thought 'better not do that again' and perhaps that is part of the learning process. The key thing, I guess, is to have personal minima and to keep learning.

My flying buddy says 'this is supposed to be fun, isn't it.' This is another way for me to figure out if a particular flight is sensible or not. If I feel it will be stressful because it is close the minima, then Stuart's voice reminds me why I fly. It's not as if I have a wounded passenger who has to reach the hospital - I'm just going somewhere for lunch!

I also use a little risk assessment checklist which helps me try to get an objective view of the riskiness of the whole flight. Sometimes it's easy to persuade yourself 'it'll be okay' and it's harder to argue with numbers.

As for the pilot in this case, he made a bad decision to depart in this weather and, for him, I think a good decision to pull the parachute. Whatever the rights or wrongs of his departure decision, the parachute appears to have saved his life.

Eric said...

Back when my university had just added Cirrus SR20s to the fleet, my instructor was working on the standardization manual for the type.

The running joke was that every abnormal checklist could be completed by simply pulling the chute. That's what I thought of when I read about this accident.

Ryan Ferguson said...

The guys flying the Cirrus are lined up to by the single-engine Cirrus jet. Scary, huh?

John Ewing said...

Ryan,

You point out a valid concern and the same could be said of the new SR22 with flight into know icing (FIKI) certification. At least with FIKI it appears that Cirrus is ahead of the training issue and will require pilots to undergo training on airframe icing. Only time will tell if there an increase in icing related accidents in Cirrus with FIKI occurs.

Anonymous said...

Love the blog John, What comes to my mind right away is all this whizz-bang tech stuff and really what Cirrus did in my opinion was sell off this CAPS chute system as a safety selling point. The reality is alot of these people do not know that the original cirrus design was prone to flat spins and the FAA required the ballistic chute because of the undesireable spin characteristics of the original cirrus design. I believe also because spin recovery was much more lengthy and obviously a danger at low altitudes such as an aggrivated stall on base to final etc...etc...I wonder how many Cirrus pilots are unaware of that. I hope none. I'll Take a 172SP anyday over a cirrus in a spin or unusual attitude.

I'm with you also on the personal minimums. I like the old saying. "Better to be down here and wish you were up there. Than up there and wish you were down here. I think one of the simplist rules of thumb I learned from my first instructor was. If you are doubting it, then you probably should'nt do it.