Airplane certification: be careful what you wish for…

More to it than meets the eye

Every pilot knows that cocktail napkins were invented so aircraft designers would have a place to sketch their wildest dreams before they start trying to certify that dream. What most pilots don’t realize is that certifying that exciting new design is but a small part of the picture. There’s financing, engineering, production and sales and, in the end, profit. If the latter isn’t possible all the rest can be for naught.

This is why I, for one, take the proposed rewrite of Part 23 certification standards not with a grain, but with a round blue cardboard container of salt. (Read John Zimmerman’s post on the Part 23 proposal for a different view of the subject.)

In the course of 57 years of private flying I flew just about every airplane certified under CAR Part 3 and FAR Part 23 that replaced it. I listened to a lot of folks complain about the cost of certification and I heard others say it never was a hindrance. I also ran up on a lot of things in certified airplanes that made me wish it had cost just a bit more.

Citation factory line
Certification rules may be complicated, but they result in safe airplanes.

These rules have withstood the test of a lot of time, though, and a lot of airplanes have been developed using these standards. I don’t think anyone would fault the development of airframes over this time period. The record has been excellent and when there have been airframe failures, with few exceptions they have come when the airplane was operated outside the envelope. The main exception I can think of came with the Cessna 441 turboprop. A structural problem with the horizontal tail came to light after it was certified.

I have known a lot of engineers at airframe manufacturers and none has ever complained about any specific Part 23 rules. I have read all of it and if you would do the same I think you would come away thinking that it is pertinent. I always felt that it outlined the kind of airplane that I wanted to fly. Yes, it is cumbersome in a bureaucratic sort of way and, yes, it could be simplified, but it is still a recipe for pretty good airplanes.

If, after it is certified, an airplane is deemed to have an unsafe condition the FAA issues an Airworthiness Directive (AD) that brings it up to snuff. If it’s really dangerous, the issuance would be an emergency AD which usually requires compliance before further flight.

ADs might be considered the grade on the quality of a certification. The more ADs, the worse the grade. There are usually more ADs issued when an airplane type is new. Often, shortcomings are not found until the airplanes are turned over to the real test pilots, the owners. We fly them and find out about glitches during normal use as opposed to test programs. Then, when airplanes get old, ADs might start showing up again but these usually relate to wear and not certification.

I think the P210 that I got new in 1979 had as many if not more ADs than any other new type. For a while it almost seemed there was an AD du jour. The systems on the airplane were not even close to being up to standard. Simply put, it came out of the oven before it was done. It slowly evolved into a reasonably reliable airplane.

Problems during certification have almost always come because, like any regulations, these are subject to interpretation and an over-zealous FAA person can be a real pain. At one point, the interpretations varied so much among FAA regional offices that the FAA hatched a plan to even things out by making lead regions for certification based on expertise in that region.

For example, airliners come mostly from the Northwest Region so that wasn’t a good place to certify a light airplane. Best let the Central Region handle that. I think the FAA’s work on this helped though interpretation of rules still plays a big part, which is true with all laws. If it wasn’t, we wouldn’t need a Supreme Court.

I’ll give you an example of a problem related to one FAA weenie’s interpretation of the rules.

Mooney
Mooney? Unsafe, according to one FAA man.

I was at the Mooney factory in Kerrville, Texas, when they were working to certify the M20J (201). The FAA was giving them fits because on a go-around, when power was applied briskly, the engine would momentarily go about 50 revs above the red line and then come right back down. The FAA guy insisted that was not certifiable.

Roy LoPresti, who was chief engineer, asked me if I would help with this. I had a Cardinal RG at the time. It had the same powerplant as the 201. Roy asked if I would take the FAA man and one of his engineers flying and see how the rpm behaved in the Cardinal.

Guess what? The rpm momentarily went 50 over and then returned, just as Roy and I knew that it would. The Cardinal was certified in a different FAA region than the one where Mooney was located and that FAA person was still reluctant to sign the airplane off. Some negotiating with the regional office finally fixed this.

Problem solved, but it delayed the certification of the M20J by as much as a week while they worked the problem. Something like that would never be solved by a rule change. It took a kinder and gentler interpretation of the rules to get the job done.

In the Part 23 proposal, three areas seem to get special emphasis. All three relate to accident causes. One is important, two are far-fetched. Loss of control accidents, especially of the low-speed variety, are at the top of every safety list and this has always been and will likely always be a place where we work hard but with little reward. The other two address Vmc in twins and airframe ice. I’ll tell you later why I think these are far-fetched.

Could changes in the certification process help reduce the number of low-speed control losses? I don’t think so.

An in-flight loss of control is directly related to what the pilot does or does not do with the controls in the airplane. How the airplane responds to those controls is covered in the regulations but rather than setting this out in black and white, it is done in shades of gray because that is the only way it can be done.

Handling (or flying) qualities vary widely among airplanes. Even airplanes from the same manufacturer don’t always fly alike. Even the exact same airplane can have flying qualities that change markedly as the center of gravity changes.

Cessna elevator
That elevator is simple, but important.

Because a mishandling of the elevator control leads to a lot of low-speed loss of control accidents, let’s look at pitch control.

The rules say that an airplane must be stable in pitch. That means that if you pull or push away from the speed for which the airplane is trimmed, it will return to the trim speed. That is pretty simple except for two things. Moving the center of gravity aft diminishes pitch stability as does increasing power. The rules don’t say how stable the airplane must be.

One model of the Piper Cheyenne, the IIXL, offers a good example of how this works. Because one stability test must be done with climb power, Piper actually had to limit the allowable climb power of this airplane to pass the test. In other words you couldn’t use full power for climb.

What does this have to with low-speed losses of control? It is all related to the power of the elevator control to move the airplane deeply into a stall.

There is nothing new about using elevator control restrictions to try to cut down on low-speed loss of control accidents. Before World War Two Dr. Otto Koppen developed an airplane called the Skyfarer that was certified as being incapable of spinning. He used some patents developed by Fred Weick who developed the Ercoupe just a bit later.

The Skyfarer never made it into production. World War Two came along but hearsay suggested to me that it was not a pleasant airplane to fly. Fred Weick was able to develop the Ercoupe into a stall-resistant, spin-proof airplane that was commercially viable for a little while in the brief aviation prosperity after World War Two.

The Ercoupe did most of what is envisioned today by those who say the low-speed loss of control accidents can be designed out of an airplane. It didn’t really work, though, because the Ercoupe had a worse safety record than other two-place airplanes.

Ercoupe
The Ercoupe was unspinnable – but it still crashed.

It doesn’t matter what you do aerodynamically; there is no way to eliminate the increase in drag as an airplane is slowed down. An increase in drag means it takes more power to fly and eventually the sink rate might outnumber the horsepower and the sink rate might increase to a disastrous level. That is also true the slower you go with no power. Technically, control of the airplane would not be lost though control of the sink rate might be considered as lost. If the hit is hard enough, it matters not if the ailerons were effective at the moment of impact.

I remember seeing pictures of fatal Ercoupe wrecks that looked almost like spin-ins.

I think that pitch stability has a lot to do with low-speed losses of control. The rules now say that the stick force must vary with speed so that any substantial speed change results in a stick force clearly perceptible to the pilot. That is apple pie and motherhood stuff but it is not precise.

The current rules take a stab at this by addressing stick forces. A maximum of 75 pounds with two hands on the wheel, 50 with one hand, and 60 pounds if the airplane actually has a stick are the greatest forces allowable. An airplane with the maximum allowable stick forces would require some grunting and groaning to do much maneuvering. On the plus side, if it were being flown too slowly that would be quite obvious until the stick forces were trimmed off.

The P210 that I flew for 28 years offers a good example of loading and stability. The airplane has an exceptionally wide cg range, from a forward cg of 37 inches aft of datum to an aft limit of 52. At forward cg, it would test those maximum allowable stick forces. At close to aft cg, you had to use your imagination to think that a speed change resulted in a stick force that was clearly perceptible to the pilot, as is required.

Just as a matter of interest there have been many STC’d mods for the P210 and to my knowledge none have been able to certify to that 52 inch aft limit. Most settle for 50. The farthest aft I flew mine was 50 and, given the handling qualities at that value, I had no curiosity about the other two inches.

The Cheyenne had stability problems because it was developed from the Navajo. There was a big horsepower increase that was destabilizing. To counter this, the original Cheyenne was certified with a variable downspring that started exerting nose-down force when the airspeed dropped below 125 knots and reached a maximum force at 100 knots. They called it a stability augmentation system.

Piper Cheyenne
Piper tried a number of ideas to deal with pitch forces in the Cheyenne.

In some other airplanes, the spring is just there all the time so it’s a plain old downspring. Either way, what springs do is, in effect, create artificial control feel. The rules just say it has to be perceptible. It does not have to be natural feel. And the springs can make it possible to add power to the airframe or increase the cg range and still pass the test. A later model of the Cheyenne was offered without the stability augmentation system. It had less power and a narrower cg range.

Because it is easier to intentionally stall an airplane with light pitch forces than one with heavy pitch forces, it is also easier to do so accidentally. Because buyers demand the widest possible center of gravity range there is really no way to say the pitch forces are required to be a certain value. Whether a prescribed range of pitch forces would help would be open to question.

The Skyfarer and Ercoupe were both stall resistant. Being two-place side-by-side airplanes, a narrow cg range was acceptable. The up-elevator travel could be limited so stalling would be limited if not impossible. That can’t be done on a four- or six-place airplane where a larger cg range is required if the airplane is to have practical value.

There is a lot more to improving low speed handling qualities than tinkering with the regulations, or the wing, so best avoid just those low altitude losses of control with appropriate bursts of brilliance and superb flying technique.

Angle-of-attack instrumentation is mentioned as a way to help on this.

We have debated this here thoroughly, but I want to add one more observation on the subject and then shut up.

Not long after I started flying, the CAA (predecessor to the FAA) instrumented a Cub with a five light system that displayed angles of attack. If I remember, the goal was to fly the approach with the middle light on. The same was true for climbing.

I think a current Part 23 regulation came from this research. It says that a visual stall warning device that requires the attention of the crew within the cockpit is not acceptable by itself. That means even with AOA, the airplane has to have a natural or aural stall warning (either of which is really is a single angle of attack indication). It might help to increase the current five knot minimum airspeed for a stall warning. In maneuvering flight in an airplane with light stick forces, that five knot minimum might lead to a beep-boom end to the day. Maybe ten knots would work better,

When my father started Air Facts in 1938, a key area of emphasis was the stall-spin accident. So there is nothing new about any of this. That doesn’t mean it should not be continually addressed but the potential solution has always been found to reside in the pilot.

Aztec on one engine
Vmc demonstrations killed more pilots than the real thing.

Another focus area is Vmc in small twins. I think it is great they are addressing this because it has been a hot-button issue for me since July 22, 1958. On that day I lost friends in what I think was the first Vmc related fatal accident in a light twin, a Beech 95 Travel Air.

That accident was followed by an alarming number of similar accidents in Piper Twin Comanches and Beechcraft 95s and Barons. Some of the most experienced multiengine instructors in the country were lost in these accidents which were a direct result of what the FAA required in multiengine training (Vmc demonstrations at low altitude).

Those accidents slowly faded from the scene as the FAA required maneuvers were changed, as popularity of twins dropped, and as the main multiengine trainer became the Piper Seminole with a few Beech Duchesses also on the line. These airplanes have impeccable engine-out manners because they were designed with the previous problems in mind.

I’m all for improving the handing qualities of new multiengine designs but only a handful of new twins are sold every year so I hope a lot of sleep isn’t lost on this. It’s not a matter of the FAA being a day late and a dollar short, it’s a matter of the FAA being 58 years late and a dollar short so how come it is a big deal now?.

And there is one icing bugaboo on the agenda, supercooled large droplets, SLD. The FAA apparently discovered this when a turboprop airliner was lost because of airframe icing.

I just looked up heavy icing in my 51 year old copy of Aviation Weather. It is described as icing that continues to accumulate despite de-icing procedures. It is critical from the standpoint of flight safety. In other words supercooled large droplets are new buzzwords for something we have known about for years. Doing such is how bureaucrats try to justify their existence.

Ice is a completely different subject for light airplanes than it is for jets. I flew a light airplane for 28 years that was approved for flight in icing conditions and I am here to tell you that such an approval is a real booby trap for gullible pilots. They put a patch on this a while back with an AD that required flight manual additions that acknowledged icing conditions exist that the equipment can not handle. I think this tells you that a blanket approval for flight in icing in a light airplane is dangerous. It makes no more sense than would approval for flight in thunderstorms.

P210 boots with ice
Boots are nice, but you should avoid ice whether you have de-ice equipment or not.

That said, I sure did like having the ice protection equipment. I knew its limitations and followed the same procedures I used for years without ice protection: at the first sign of ice, do something to get out of that ice. Icing equipment in light airplanes is a valuable tool to use while you go for help.

So, addressing heavy or severe ice, or SLDs, is beside the point for most private pilots. The only thing that can help with that is action from the PIC.

A positive recent development not related to the Part 23 proposal has been the approval of some non-FAA approved electronic devices for certain airplanes under an STC obtained by EAA.

Hi-tech cockpits shouldn’t also be high-bucks because of tedious approval processes. The most significant development for general aviation cockpits has come from the iPad which you can buy anywhere and carry with you in any airplane. The aviation programs that are available for the iPad can bring an almost unheard of level of capability to an airplane that doesn’t even have an electrical system.

Really good and relatively inexpensive installed avionics have been available for experimental airplanes and there is great promise in the approval of this equipment in certified airplanes.

Back to Part 23. There is no harm to be found in a review and possible simplification of these rules. If common sense prevails, though, the airplanes that are developed will be much like the airplanes that came from the existing rules. There’s just too much good history there to ignore and when all is said and done I don’t think it will be ignored. Also, a change in the attitude of FAA personnel as they interpret regulations might do more to help than anything else.

Don’t count on flying becoming less costly, though, and don’t look for any improvement in the safety record. Noble as those causes may be, they have always been elusive.

9 Comments

  • Total agreement here. On Vmc qualities: After developing a twin engine Cessna Caravan (two PT-6s, one propeller), I came to the conclusion that in addition to benign Vmc qualities it would be beneficial to change the rules to require somewhat better single engine climb gradient requirements. My impression was that the surprise factor when flying an airplane that normally climbed at 2000 ft/min changed to one that barely made 400 ft/min created a nearly irresistible urge to pull back. Manufacturers have to increase the gross weight to just meet minimum requirements or they will not be competitive in the payload department. So I think it would be good to increase the requirements with a level playing field. Of course, then prospective buyers would whine that the new airplanes won’t carry as much as their old Baron.

  • Well, I differ from Richard. I understand his lack of faith in the FAA to make flying inherently safer or cheaper, but I think he misses the point of the Part 23 rewrite.

    The point of the reform is to significantly reduce the FAA’s direct role in deciding prescriptively how to certify a new or modified aircraft, replacing it with much less bureaucratic and sclerotic private industry consensus based standards, presumably to be developed (based in part on their experience with LSA) by ASTM. ASTM has been in the private engineering standards development business for many decades, quite successfully. ASTM, unlike the FAA, has been vastly more receptive to advances in commercialized science, technology, and engineering. Just look around at today’s computers, electronics, building materials, automobiles and consumer products as compared to 80 years ago when the technological basis for most of today’s light aircraft were implemented.

    Whether aircraft can be made aerodynamically stall proof is not the point. The point is that safer and cheaper and more capable and more efficient aircraft are much more likely to see the light of day in a regulatory environment built upon private sector consensus standards than the prescriptive, sclerotic, bowels of the Federal bureaucracy. After all, this approach has worked extremely well in the building industry, the energy industry, the consumer products industry, and so forth for generations now. That is why we have high tech materials, new products, and new designs constantly streaming out into the world of commerce, at ever greater efficiencies and lower prices … while the FAA is still proud to rule over an industry still dominated by 1930s era designs and prices and performance standards.

    Think iPhone vs. a 2016 Cessna 172, which is, other than the Garmin G1000 in it, almost identical to the 1960 model. Let’s see, the 1960 version of the iPhone was, what was it now? Oh yeah, there was no such thing, and nobody could have conceived of a power computer and wireless telephone back in 1960.

    The actual technology of aviation is not the issue with certification standards .. it’s the Federal process that has always been a huge drag on innovation and technological improvement. The proposed Part 23 reforms appear to me to be a very good first step, but only a first step, into the 21st century for American aviation.

  • Regulation is one mortal enemy of innovation. Another mortal enemy is standardization.

    As Duane has noted, free enterprise and free design has given us tools we could not even imagine 70 years ago when I first soloed and started to read Air Facts!.

    The FAA is populated by good people who want to do a good job, but regulation rarely improves anything, including safety. Knowledge works!

    Happy Skies to All

    Old Bob Siegfried

  • Into ally or otherwise, you are conflating very different aspects and presenting your scenarios as if they are the majority case. For the GA fleet, the impact of Part 23 on certification costs and on maintenace costs cannot be overstated – one has to look no further than the modern experimental fleet to find equivilant so that illustrate this cost impact, the most obvious being in avionics where the experimental fleet has access to more vendors, generally newer technology and greater functionality for significantly lower cost.

    The impact can also be seen in general maintenace for the ubiquitous 172. Timken wheel bearings – same part number from Timken, the PMA part supposedly has tighter tolerances but the lack raises questions at least to me. Get the part number Timken bearing and race at an auto store and you can replace the bearings in all three wheels for the price of the PMA bearing alone for one main gear.

    Plane Jane stall horn same aircraft. It is a plastic bicycle horn, the difference is the air craft version isn’t chromed and doesn’t come with a squeeze bulb. Cost from Cessna approx $100, cost from Aircraft Spruce for PMA part $42, cost at Toys-R-US approx. $2 ( this from two different APs).

    The Piper Cherokee had the major design features done on an envelope at a diner and was in production a year later. Today to certify (or recertify if a vendor were foolish enough to make a major design change) is typically a multi year project that will eat multi millions of dollars and will be no safer than if it had to meet typical ASTM and aeronautical engineering standards.

    We still fly engines that were designed in the forties because the costs to recertify them with modern engine controls (FADEC) are simply too onerous and would drive costs do high for new aircraft that a 172 would cost in the $1M range if it could be sold at all. Diesel is the only technology to take that on simply because it couldn’t work without it. The ROTAX 912 UIS hS many advanced features that give both better performance, better fuel burn, lower emissions, and improved reliability. Not available for the certified fleet (mainly trainers and LSAs). Only the 912U has been certified which is essentially the 912 design frozen at the time of the initial certification. Costs are too much to recertify for the improvements/technology.

    Dynon has gotten their eAI approved for use without going through the certification process, based mainly on service history and installed base (15000+), but currently has no plans to attempt the same for their glass panel system. Reason is they feel the demand in the GA market is too small to make it worth while to pursue another one off approval, plus the smaller installed base acts against getting the approval.

  • Dear Clayton:

    Thanks for your thoughtful comment.

    The Piper Cherokee was certified rather quickly under the same basic regulations that apply today. The difference was in the FAA of that day coming to the party with a let’s get this done attitude. Today there would be many more FAA people involved and each one would be working more to justify his job than to certify the airplane.

    That’s just the times. The Empire State building was built in a year back in the 1930s. The new World Trade Center took a lot longer and cost a lot more real dollars.

  • I think Richard was dead on in his accessment. One point that is not being discussed is that experimental aircraft still need the FAAs inspection and approval and there are a lot of people that step up to the challenge. Necessity has always been the mother of innovation and that has never changed. Take away all the regulations and it’s a race to the bottom as corporate entities love to pass the responsibility buck to whoever they can to maximize their profits. Typically when corporations find better simpler more cost effective ways to make something, it take competition for sales to get them to reduce their profit margins. Corporations typically loose sight of innovation and begin ignoring their engineers listening more to their marketing and accounting departments. MBA programs are a great way to take a mediocre engineer and turn them into a profit hound leading the company down a path to where the products don’t offer the quality or options that they use to.

    Never fear though, other countries are picking up the slack and making good products to bring to market. Their regulatory agencies make the FAA look like push overs. The companies abroad have been stepping up and meeting the challenges of their regulatory agencies instead of waisting time debating and arguing. They work with their regulatory agencies instead of fight and blame them. Now with bilateral agreements in place, the FAA accepts other countries regulatory work and they accept the FAAs work. It is essentially creating a level global playing field to bring more options and more competition to market. We will just have to watch and see how corporate America responds.

  • Call me a Luddite, but some level of regulation is required, IMHO. Now, those regulations should be performance based, not how to’s. I don’t want someone flying on a system in the clag that doesn’t meet some level of required performance. That said, in the field of avionics, it then becomes one of buyer beware, unless there is some level of standardized interface, performance, and reliability required. When it comes to basic aircraft structural reliability, the current regulations are adequate. A manufacturer can save a lot of money by limiting his structural life, and only analyzing or testing to that limit. Would you buy an airframe only good for a 1000 hrs? How about 5000? What about systems being time limited? Would you accept a reliability on a vacuum pump of only 200 hrs? or an alternator of 200 hrs for all electric panels? Bet you can get one of those subsystems cheap, due to cheaper internal component with lower reliability, and compliance testing requirements, but then you have to replace it. So what is the cost trade-off?

    • Actually, Richard, new designs for aircraft systems including engines and avionics are likely to be much longer life components than what has been standard in the industry forever.

      For one thing, electronics are vastly more reliable and more easily provided with backup redundancy than the old style mechanical systems such as vac pumps. A typical modern electronic system is built with MBTF (mean time between failure) in the tens of thousands of hours, whereas vac pumps have a typical lifespan of only hundreds of hours. And the cheapness of electronic systems, especially once freed of the onerous TSO certification process, means one can afford many backups in the panel. I would much rather trust my life to three separate redundant, and cheap systems than a very single expensive TSO certified system that can break as easily as a cheap one.

      And of course, unlike any mechanical system, electronics systems can talk to each other, share data, and share interfaces. And can be easily upgraded with software changes.

      Aircraft manufacturers are still making and selling 1930s tech aircraft engines, opposed 4 and 6 cyl engines with carburetors or fuel injection, magnetos, poor fuel economy, no redundancy, and TBOs ranging between 1,200 hours and 2,000 hours.

      While in just the last 40 years, automobile engines went from 1970s performance of maybe 50,000 miles before a valve job was necessary, cars rusted out in less than 6 years, and an entire new or overhauled engine, transmission, and drive train was necessary between 75,000 miles and 100,000 miles … to today when typical cars are able to go 300,000 miles, 400,000 miles or much longer, and basically the aesthetics go bad long before the mechanicals go bad. Not to mention that cars today are a heckuva lot safer than 40 years ago. The government had a role in that, but private industry, not the FDOT, made it happen.

      We could have that same improvement in aviation mechanicals and avionics if the FAA wouldjust get out of the way and let commercial operators do what they’ve already done in most other types of consumer products.

      If the government is controlling innovation – then there will be little innovation and improvement. The proper role of government is not to “control” or to “regulate” but to provide guidance to private industry in how to use private consensus standards to achieve the stated objectives. It’s not as if this is a new paradigm – the FAA is mere 50 years behind every other area of human endeavor in the USA.

  • I don’t think that the rewrite will help things much in many areas.

    It mostly doesn’t change some of the cumbersome procedural requirments (Part 21) that were addressed in the ARC report.

    Air Tractor has just published a negative comment on the Docket including the possibility of cumbersome and lengthy negotiations on means of compliance. I appears that they are so far the only entity to point out that the emperor has no clothes.

    There are some technical and legal gotchas in the proposal. There are areas where the new regulations appear to be more stringent than existing ones, without rationale , justification or cost/benefit anlysis, and areas where the ASTM standards conflict significantly with the new regulations. There a few areas where compliance may not be practical based on the most logical interpretation of the plain English requirment. For example, continued safe flight and landing must be possible after failure of any powerplant system component or accessory.

    Although I am not an expert in this area, I wonder about the practicality of the new Loss of Control requirments. The two alternatives seem to be spin (or stall) resistant aerodynamics and active systems, such as stick pushers. The first could involve penalties in other performance and handling areas, as well as suitability to larger aircraft which necessarily require a wider C.G. range. The second involves obvious penalties in cost, weight, failure potential and maintenance.

Leave a Reply

Your email address will not be published. Required fields are marked *