When Cirrus unveiled the prototype that would become the SR20 at Oshkosh in 1994, the airplane looked like just another plastic homebuilt, albeit one that was more refined. But it had a twist. Amidst skepticism bordering on derision, Cirrus proposed a radical idea: in the event of an unrecoverable emergency, the entire airplane would have a parachute capable of lowering the aircraft to the ground in one piece, if not gently.
Never work, said some critics, never sell said others and even many who liked the idea harbored a deep-seated resignation that the FAA would never certify such a bleeding-edge idea.
Cirrus has, of course, proved the doubters decidedly wrong. Not only did the company certify the parachute for both of its models-the SR20 and the SR22-but just as the companys president, Alan Klapmeier predicted, the parachute has and is proving a keystone in a successful sales effort by Cirrus. Despite a cold shoulder from stick-and-rudder traditionalists, the Klapmeier brothers, Alan and Dale, have seen their vision proven.
The only troubling side of this GA success story is that the Cirrus airplanes have accumulated what can best be described as an unenviable initial accident record. The two Cirrus models have suffered at least eight accidents, four of them with fatalities, yet the company has fewer than 500 aircraft flying.
Crashes of new airplanes are nothing new, of course, and are thus hardly newsworthy. Yet the Cirrus crashes are news precisely because the aircraft are being watched so closely. When a new Cessna 182 makes a smoking hole, it hardly merits notice, while a Cirrus accident usually earns a blurb here and there in aviation magazines that normally ignore minor wrecks.
The Cirrus crashes have attracted raised eyebrows in the industry both because of the relatively high number and because thus far, the CAPS parachute has yet to save the day, even once. On paper, the Cirrus is one of the safest GA airplanes ever designed and the expectations for it are accordingly high. But in reality, the initial safety record has yet to prove out the design.
Low Numbers, High Rate
As weve often reported, comparing one models accident record against anothers is fraught with missteps. The NTSBs accident reporting tends toward sketchiness and the all-important hours-flown totals are difficult to come by.
Right up front well offer this disclaimer: both Cirrus and Lancair are too new to have accumulated a truly meaningful accident record. Therefore, we dont think its entirely fair to make too much of the data just yet.
On the other hand, Diamond has been in the game since 1994 and does have a longer-term record which we think is impressive. Further, because Cirrus could have been expected to do better than it has, were not inclined to sweep the accident record under the rug entirely and attribute it to bad luck.
With these provisos in mind, heres a quick look at comparative accident rates: With some 800 aircraft flying worldwide and nearing 1 million fleet hours, Diamonds fleet has posted an impressively low accident rate. In the U.S. and Canada, the Katana has flown about 800,000 hours and we count 24 accidents for an overall rate of 3 accidents per 100,000 flight hours, below average for the GA fleet.
Just this summer, the Katana had its second fatal accident in Leesburg, Virginia, for a fatal rate of .25/100,000 hours, again, among the lowest if not the lowest of GA aircraft in the U.S.
Lancair has had only one accident, the fatal crash of a pre-production airplane in January of 1999. To date, it has delivered a fraction of what Cirrus has, about 45 Columbia 300s. The fleet has accumulated too few hours to make a realistic stab at rate calculations.
Cirrus estimates its fleet has accumulated about 77,000 hours. As of mid-summer 2002, with Cirrus nearing delivery of 500 airplanes, we count eight accidents, half of them fatal, for an overall accident rate of about 10 per 100,000 hours or three times that of the typical GA aircraft.
The fatal accident rate for Cirrus aircraft is also high at 5 fatalities per 100,000 hours, compared to a GA average of 1 or less per 100,000 hours. Once again, we note that these calculations are based on small total numbers and are almost certain to decline as the fleet matures. Nonetheless, we still think it fair to examine the rate and ponder why its so high.
Diamond has accumulated a much better initial record and allowing for far fewer aircraft flying, so has Lancair. While it is true that Katanas motoring around in the pattern face different risk profiles than SR22s charging across half a continent, its also true that theyre flown by ham-fisted students and green-as-grass CFIs who can break airplanes with the best of them.
Examining recent traditional designs for which reasonably accurate accident data is available, we note that the Mooney Ovation first appeared in 1994 but appears to have suffered no accidents at all for its first four years in production. On the other hand, its actual rate may be as high as the Cirrus; some 200-plus are flying for an estimated total fleet hours of some 41,000. That gives the Ovation an overall rate of 9.7/100,000 hours.
Examining Cessnas new production, we found that the company delivered almost twice as many new Skyhawks in its first two years of renewed production as Cirrus has in its entire history. As far as we can tell, the new Skyhawk suffered about nine accidents during this period, two fatal. Cessna estimates that since new production resumed, Skyhawks have accumulated about 2 million hours for an estimated overall accident rate of 2.1/100,000 and a fatal rate of .3/100,000 hours.
These numbers, if accurate, are a bit better than the long-term accident records for the Skyhawk, which we think is good news. Some 43 Skyhawks of about 2100 built since resumption of production in 1997 have crashed.
Yet no one raises an eyebrow about the Mooney or the Skyhawk because neither is a Cirrus, which burst upon the market with lofty claims about new thinking in crashworthiness and safety. Fair question, then: Is there something wrong with Cirrus airplanes that weve simply missed, some underlying fatal flaw? The short answer is we dont think so but we still wonder if something is amiss.
A Pattern? Not Really
Normally, when an airplane has a fatal flaw, it shows up in accident patterns with shared characteristics. But that doesnt appear to be the case with the Cirrus aircraft, unless pilot proficiency/judgment issues are considered. And obviously, these must be considered to evaluate an aircraft fairly.
Other than a fatal pre-production test crash, the first Cirrus accident was a fatal VFR-into-IMC/CFIT in April of 2001. The pilot wasnt instrument rated. Two months later, a landing accident injured a pilot and destroyed an SR22 in Springfield, Illinois.
Later that summer, an SR22 suffered an odd engine failure and made an emergency landing in Gibson, Georgia. We call it odd because the pilot inadvertently entered a thunderstorm and reported that resultant G forces caused the engine to stop and it couldnt be restarted. The landing gear separated on landing but no one was seriously hurt.
In September 2001, an SR20 suffered an engine failure after a loose oil drain plug caused the engine to fail because of oil starvation. Two occupants were injured during the forced landing.
In March of this year, two pilots were uninjured in another oddball accident. The two were flying an SR20 on autopilot in IMC near Lexington, Kentucky. The pilot told the NTSB that the turn coordinator-which operates the S-TEC autopilot-was pegged to the left with no flag showing. He uncoupled the AP but by then, the aircraft was in a steep dive. It exited the base of the clouds and, unable to regain control, the pilot re-entered the overcast at which point he reported that the AI tumbled.
The two pilots attempted to deploy the CAPS but couldnt displace the activation handle and they later told the NTSB they couldnt tell if it had been deployed or not. The aircraft touched down in a field and came to rest against a tree. Subsequently, the CAPS fired on the ground. A post accident investigation revealed no instrumentation or autopilot failures, according to the NTSBs preliminary report, and the aircraft crash landed with a perfectly functioning engine.
In April, a factory new SR22 with just a few hours on it, crashed into a field near Parish, New York, killing both occupants, who were the co-owners of the airplane. They had taken delivery of the airplane only a few weeks before the fatal accident. Witnesses, according to the NTSB, described the SR22 as having entered a flat spin, the only recovery from which is deployment of the CAPS, according to the Cirrus POH. Its not known if the pilots attempted to deploy the CAPS and couldnt or if they simply didnt try.
The most recent Cirrus accident involved an SR20 at Angel Fire, New Mexico. The aircraft crashed into rising terrain after takeoff in high density altitude conditions, killing the pilot. As we go press, we have no clue as to the cause of the Angel Fire crash but density altitude may be the culprit. In reality, the accident picture is somewhat worse in that several additional crashes were classified as incidents. We havent included these in our count.
Regardless, the bottom line is not pretty for Cirrus: eight total accidents, four fatal and two in which the ballistic parachute system could have saved the day but either wasnt deployed or didnt function. Cirrus has admitted that the cable mechanism controlling the parachute could have yielded a hard pull. It has since redesigned the mechanism and retrofitted existing aircraft via service bulletins.
The cable issue aside, a review of the Cirrus accidents-at least those involving production aircraft-reveal nothing to damn the aircraft itself, in our estimation. The two engine failures appear to be induced by either inadequate maintenance or fuel exhaustion and thus far, the three production-version fatal accidents have failed to implicate the aircraft or its systems. If the Parish, New York accident was indeed a stall/spin, its the first such occurrence for a Cirrus. Although neither the SR20 nor SR22 are certified for spins, both have undergone some spin testing and even though spin recovery may be neither conventional nor as straightforward as in other models, the so-called equivalent level of safety is the parachute. Standard recovery calls for deployment of the CAPS. Thus far, theres no indication that this was attempted in the Parish crash.
The harsh reality of general aviation safety is that most crashes are caused by pilot error, not an aircraft mechanical failure or design flaw. While its true that certain design features may induce pilot error-such as poor climb capability or ineffective flaps or misplaced controls-we havent seen this in the Cirrus, nor have owners complained about anything other than minor recurrent maintenance issues.
Noting that the accidents in the production airplanes thus far appear to be the result of pilot proficiency/judgment errors, were inclined to wonder if the Cirrus marketing ethos attracts either inexperienced pilots or those who may place undue faith on the aircrafts safety systems at the expense of appropriate training.
We shopped this idea to Cirrus CEO Alan Klapmeier. Since agreeing with such a theory would be tantamount to customer bashing, we didnt expect him to touch it with a barge pole. And he didnt. However, Klapmeier concedes that Cirrus accidents have grabbed the companys attention.
Were hugely disappointed, obviously. Its certainly more accidents than we expected to see at this point, he replied when asked about the Cirrus accident showing thus far, but he also added that initial spike in the accident rate cant be seen as a fair assessment of the models long-term record. Statistics, he says, arent proof of anything.
The Cirrus parachute system has come under criticism for being variously unnecessary, impractical, too heavy and now ineffectual. We wondered if Klapmeier and Cirrus have second thoughts about it or are considering making it an option.
Klapmeier says no, the company is as committed as ever to the CAPS idea. But he concedes it may have an unknown effect on the way pilots operate the Cirrus airplanes due to what he calls the anti-lock brake syndrome. If you have anti-lock brakes, goes the reasoning, do you therefore drive faster on a slick road than you might otherwise because the brakes provide extra margin? Perhaps you do, says Klapmeier. But if you have an accident, would you want to get rid of the anti-lock brakes?
Is there something about the airplane that does that, that reduces the care that pilots take? Or does it change the characteristics that pilots have? asks Klapmeier. Even if thats true, he argues, removing the parachute makes no sense because it provides a range of options that vastly increase overall safety. For example, it offers a survivable response to structural failures, midair collisions and loss-of-control in IMC that no other system does.
In the Parish, New York accident, if a flat spin turns out to be the cause, would Cirrus consider recertifying the airplane with a spin test program and making the CAPS optional? Again, says Klapmeier, the answer is no.
The wing was deliberately designed to be stall resistant, which also makes it spin resistant but more difficult to recover from a spin once one has developed. Recertifying for spins would require modifying the wing.
We would not consider degrading the stall characteristics to improve spin characteristics, says Klapmeier. These airplanes arent going to be flown by retired fighter pilots. Instead, says Klapmeier, Cirrus is examining improvements in its training program, possibly to include a simulator and a decision matrix on when and how to deploy the parachute. Training may also have to be improved in other areas, he adds. In other words, Cirrus is not ignoring what it recognizes as an unfavorable initial safety image.
As for insurance concerns and reparability, Klapmeier concedes that Precision Airs Jim Stoia (see sidebar) is right, Cirrus hasnt developed an extensive field service and repair program yet.
Frankly, we are more than just a little overwhelmed right now. Well get to it someday. We cant now. In the long run, were convinced that these airplanes will be very repairable, Klapmeier explains. He says that Cirrus has met with underwriters and believes the aircraft will remain readily insurable.
From our perspective, many would-be buyers and some current Cirrus owners have noticed the accident trend to the extent that theyre asking whats going on with Cirrus? Its a fair question, borne more of curiosity than alarm.
In our view, Cirrus has acquired a minor smudge on its otherwise highly burnished image that has little or nothing to do with the airplanes themselves and everything to do with the usual lack of training and proficiency that blackens the eye of GA on a daily basis. But thats not to say Cirrus has no role to play.
We think the company most needs a long stretch of accident-free flying and an aggressive, highly visible training and proficiency program centered around basic airmanship and the use of the CAPS.
Second, we would like to see the company devote the resources to an energetic field service program so that many-not just a few-shops can repair broken Cirrus airplanes easily.
We think its a mistake for Cirrus to eventually catch up to this problem; it should be out in front, leading the way, as Diamond Aircraft has done.
Finally, in last months issue, we named the SR22 our Product of the Year. Nothing weve reported here causes us to reconsider that recommendation, despite the fact that we think Cirrus has work to do. Although the parachute idea remains unvalidated, we think Cirrus aircraft will accumulate a favorable accident record.