The Autopilot Problem

New Avidyne digital autopilot

We have an autopilot problem and most of the blame for it belongs to the FAA. Last week the FAA administrator Michael Huerta acknowledged the autopilot issue, noted that autopilots can be a safety aid, and said the FAA would do something to solve the problem.

The autopilot problem is that the FAA regulations have almost totally restricted an airplane owners choice for installing an autopilot in a certified airplane. In some less popular models of airplane there may be no choice at all.

The FAA’s contribution to this problem is its certification rules. I know we blame certification for lots of aviation issues. But when it comes to autopilots, it really is the morass of certification that keeps modern autopilots out of most GA airplanes.

Long-standing certification rules require that any autopilot be extensively tested in every individual model of airplane before it can be approved. The autopilot flight testing covers essentially the entire normal operating envelope for the airplane. Autopilot failures must be introduced in flight to be sure the autopilot does not fly the airplane into an unsafe attitude or airspeed before the human pilot can recognize the failure and disengage the autopilot.

In almost every case the autopilot manufacturer must make electrical and mechanical changes to the autopilot system to meet the certification rules. For example, one airplane may need more torque from the servo to move its flight controls so that requires more electrical power and a different mechanical clutch and capstan to connect the servo to the controls. Each change must then be flight tested and documented, and the testing only satisfies certification rules for that specific model.

Unlike most other avionics equipment that is eligible for installation in a huge range of airplanes once certification testing is complete an autopilot approval is limited to a single model. The autopilot manufacturer has at best a very small potential market after the certification investment has been made.

Because of this cost and hassle newly designed, new technology autopilots are very rare. When they do come along such as the new Garmin GFC 700 that is part of its flat glass avionics system, the autopilot is only certified in current production airplanes because that is what makes economic sense.

So the owner of a well cared for airplane that is more than a few years old can install the most capable and up to date flat glass displays, traffic warning system, WAAS navigator, ADS-B and so on, but the only autopilot option for his airplane uses technology that is decades old because the cost of certification keeps newer autopilot designs out.

The FAA’s certification rules made some sense when autopilots were not very smart analog devices that needed careful adjustment of gains and servo velocities and torque and so on to fly properly and safely. But digital electronics and microprocessors changed all of that. A newly designed autopilot can be “smart” and can teach itself to fly the airplane and store that information in memory. Certification of a current technology autopilot could be low cost and simple if the FAA would adjust its rules to match advances in electronics.

Low-cost, compact and very capable autopilots are widely available to homebuilders, and newly designed equipment is becoming available at a rapid pace. These smart autopilots learn how to fly the homebuilt without the long drawn out process FAA rules require for certification. So the homebuilder can have a very capable two or three-axis autopilot at an affordable cost while the owner of well maintained and perfectly equipped certified airplane has little or no choice in autopilots that are certified for his airplane.

Autopilots are a proven safety aid when flying in the clouds or low visibility, or when flying solo in congested airspace. They give the human pilot time to be captain while the autopilot is a copilot who holds heading and altitude. The FAA is so convinced of the safety autopilots provide that it won’t allow charter pilots to carry passengers while flying single pilot without one. And an autopilot is a requirement to fly single pilot in jets. It’s the FAA’s own rules.

So maybe the fact that the administrator went on record saying that the FAA wants to encourage the development and installation of new autopilots to enhance safety is a change in the right direction. Autopilot makers are understandably wary. If they invest in creating new, digital smart autopilots that can be installed in a wide variety of airplanes without costly and unnecessary certification, and then the FAA pulls the rug out, they are screwed. And so are we airplane owners.

The FAA administrator can set goals, and he has, but let’s see if the people in the trenches who make and enforce regulation and policy were listening to him. I sure do hope so.

This entry was posted in Mac Clellan's Left Seat Blog. Bookmark the permalink.

34 Responses to The Autopilot Problem

  1. Fred Stadler says:

    Right on! My 50+ year old Cessna 310 flies great, but certainly could benefit from an autopilot, or even just a wing-leveler. Their absence means that copying a clearance in IMC is sometimes followed by “unusual attitude recovery.” Despite the wonders of modern electronics, the only autopilot certified for my plane is an equally geriatric model, far less capable than ones designed in this century.

  2. Kayak Jack says:

    … and exactly the same comments can be repeated about strobe lights on certified planes vs experimental. The very same part costs thousands of dollars more for a certified plane. “See and be seen” is greatly enhanced – even in daylight – with strobes. But, thanks to FAA’s strong policy of nonthinking, we have to use standard running lights, which struggle to put out 2 candle power.

    Someone needs to go into FAA with a large broom. They, like Congress, should not be allowed to start a new rule or law, without first deleting two old ones.

  3. William McIntosh says:

    I used to dismiss people who quoted Reagan’s famous “government is the problem” mantra as right-wing dreamers-political neophytes and Limbaugh listeners.

    I was wrong, and they were right.

    Today, whether it’s Part 23 cerification, autopilot certification, or yes, even strobe lights, the FAA’s cumbersome rules, dating back to the 19whenevers, are inhibiting further growth in the general aviation sector.

    Although I believe that from a technological and accessibility point of view general aviation has never been healthier, the truth is that the government moves far too slowly in the internet and information age to be a credible player that can truly act on new information and provide solutions, so not surprisingly, savvy corporations have learned how to market their special cases to Congress and the public, thus effectively bypassing the government and its onerous departments.

    An effective industry-wide marketing campaign would promote an “Aviation Reform Act,” highlighting the inhibiting effect that FAA outdated regulations have on the industry. It would also appeal to nonpilots, because aviation needs their support, too. It would counter Obama’s “corporate jet loophole” propaganda by pointing out and the business jet as creating the time necessary for decision makers to create more useful and better products, and to make better decisions, period. And a really good marketing campaign would introduce people to the everyday folks who build airplanes, and how more jobs could be created, but for our policies.

    But first we in aviation should agree on the reforms needed.

  4. Thomas Boyle says:


    I’m afraid we’re a long, long way from agreeing on the needed reforms. We still have too many people in GA, who grew up in the airlines and in the military, who have been trained to believe that the FAA’s regulations are a good and wholesome thing, and that the market could never do as well – or better – on its own.

    These people – many of them senior editors in our aviation publications – will cry “safety!” if changes beyond minor are proposed. What is needed is wholesale waiver, not just for avionics, but a transition to – very nearly – “anything goes” for private aviation. There’ a need for it to be a transition, because people trained to think the government is doing their homework for them, will need time to adjust to the idea of researching a company’s reputation. But – no government department approves changes to cars, amazing though that may seem. Crash testing is done by the insurance industry, not the government. Can you imagine insurance companies standing in the way of safety advances in aviation? Me either.

    Maybe, once our statist GA leaders have retired, if there are any GA pilots left by then…

  5. 172 Flyer says:

    you can find far advanced technology and safe technology in airplanes classed as “experimental”. It also costs a lot less than what can be put in our “certified” airplanes. Yep, the FAA is very slow to make changes, as it might also mean a few bureaucrats will lose their jobs, and we just can’t have that, can we?

    • Bill Tomlinson says:

      Spot on! What we need is a new class within the Experimental category. One that imposes the same restrictions against operation for reward, and offers the same freedom to “do as you please at your own risk”, but allows the aircraft to be built in a factory rather than having to be 51% built by the owner.

    • Jeff Boatright says:

      Actually, mission or protocol changes often mean expansion of positions in large organizations. The FAA is slow to make changes for many reasons. I don’t (yet) understand how that is caused by fear of loss of positions. Please clarify.

  6. M. Smith says:

    I have traveled to several third world countries where “anything goes” and business’s in those countries are not honorable and people don’t have the time to “do the research” they need on “every” product in the market place to ensure they are getting the best or safest product. People (who live in a well organized society) and think that the marketplace can sort things out for the benefit of the consumer have never lived in a third world “wild west” market place. The point is the regulations, tests, and requirements were put in place so that people wouldn’t have to research everything they purchase to the nth degree before they purchased it. They were put in place so people would not have to lose their life finding out that a product did not live up to the job it was supposed to do. Not everyone wants to be a “test pilot” for new products. Most people want to get into their aircraft and enjoy a nice flight without the worry of some untested manufacturers product turning that flight into a nightmare. So by all means encourage the FAA to change the regulations where it makes sense, but do not ask them to throw out regulations just because someone wants to upgrade their old autopilot. This is dangerous and stupid. My suggestion would be for the autopilot manufacturers to divorce the controller from the servos, pistons, and other control surface movement mechanisms. Have a specialized controller for each plane that moved the control surfaces according to a standardized input. For example the input could be zero to five volts for full range movement of the rudder from left to right with 2.5 volts being the center position. The planes controller would provide all the torque corrections or servo current and voltage corrections. Any autopilot that provide the zero to five volt signal range would be certified for that plane. Some people might say the deflection of the rudder is more on some planes than others, and I would agree, and they would go on to say for some planes you need more rudder than others, during a standard rate turn, and I would agree. To these people I say the controller would be the component fine tuned for the plane to provide the right input for the standardized autopilot signal. In other words, a signal voltage of 2.8 volts for example would result in the correct rudder deflection for a standard rate turn at a specific speed as a percentage of the planes calm air speed, (or something along those lines). Sure this adds another middle component into the equation but it frees the autopilot manufacturers and the middle component manufacturers would only have to worry about meeting the standard for each aircraft and not about all the control that an autopilot provides. So the guy with the old Cesna he finds the correct interface for his planes servos, trim, or whatever an autopilot would use and then he would be free to chose any autopilot.
    Don’t throw out rules and regulations that keep us all safe. Change them in a way that will keep us all safe. Also, autopilot manufacturers need to think different too and work as a group to provide standardization which is much easier to verify and test. Today they build whatever they want and cry when it costs a lot of money to certify it.

    • Bill Tomlinson says:

      “The point is the regulations, tests, and requirements were put in place so that people wouldn’t have to research everything they purchase to the nth degree before they purchased it.” Your point is valid, but could equally be satisfied by a bureau-of-standards approach rather than by compulsion.

  7. Thomas Boyle says:


    See what I mean?

    And if cars were certificated by the government too, just think how hard it would be for people to imagine a world with safe, modern, non-government-certificated cars!

    Which, of course, is the world we actually live in.

    • Jeff Boatright says:

      Not to be obtuse, but aren’t cars built to standards set by various government entities, in the US and in other countries? Maybe I’m misunderstanding your statement.

      • Rodney says:

        I think the idea is that cars have to have certian equipment installed to meet certian standards however if a company wants to but an airbag from one manufacturer in one model and a different one in another they do not have to thourghly test both models. Same with engines, the engine doesn’t have to meet a government spec only the emissions and if you want to put three different engines in a car you don’t have to go through three different certifications that can take years each. For a simple fixed gear fixed pitch airplane why does it require thousands of hours and many years to get certification? I would think a couple hundred flight hours covering behavior in the flight envelope and maybe a few months to go over the paperwork would be sufficient. It seems to me the FAA delays, refigures, wants changes and retests much more than is necessary stretching certifications out for years and years. Maybe that is because only older, wealthy republicans fly airplanes. :-)

        • Jeff Boatright says:

          Thanks for the explanation – very clear and it makes sense (to me, anyway). But, I’m not sure about the statement “a couple hundred flight hours covering behavior in the flight envelope and maybe a few months to go over the paperwork would be sufficient” in that I think some, and maybe much, of the certification process actually starts at that level, but then the process morphs due to many reasons, including those you cite (“delays, refigures, wants changes and retests much more than is necessary stretching certifications out for years and years”).

          Your last sentence seems to imply that the FAA works faster under a Republican administration. I haven’t seen any difference in speed or service level through Reagan (except for when he fired the controllers – a biggie), Bush I, Clinton, Bush II, and Obama (even with the sequester) eras. I may be extremely lucky, but for the few and minor issues where I’ve dealt directly with them, FAA personnel have been fast, thorough, and courteous.

  8. Thomas Boyle says:

    M. Smith,

    I propose a simple solution to your concerns.

    Let us have our “(almost) anything goes” regime, but let the FAA continue to provide certification services. People can buy FAA-certificated airplanes, or rely on other certifications (such as ASTM/LSA-type certifications), or rely on the manufacturer’s brand (are you really worried that Cessna or even Vans will let you down?).

    And let’s see which system succeeds in producing safe, modern, less costly aircraft!

    The only reason for the FAA to refuse, is fear of losing.

    • Jeff Boatright says:

      “…are you really worried that Cessna or even Vans will let you down…”

      Why yes, yes I am. Just like Ford, GM, Toyota, etc., left to their own devices will build and sell to maximize profits with less regard to safety until compelled, aircraft manufacturers will do the same. I don’t want to be the pilot who “discovers” the equivalent of the Pinto gas tank.

      Come to think of it, I guess I take it one step further, since I build and maintain what I fly…

      • Rodney says:

        Yes but you forget the Pinto was approved by the government and met all the applicable safety standards. What current certification standards don’t cover is yet to be discovered but their effort to cover everything (even the most remote of possibilities) is one of the things that makes it so expensive.

        • Thomas Boyle says:

          The Pinto wasn’t approved by the government. The government doesn’t certificate cars.

          It does have some safety requirements, but those are not at all the same thing.

          The Pinto was a failure, in safety terms. But if we had government certification of cars, we’d still be driving cars whose essential design dated to the 1950s. Any minor gains in Pinto avoidance would be overwhelmed by the loss of all the advances the un-certificated automotive world has brought since then.

          • Greg W says:

            Some of us do drive cars designed in the 50′s. My car is a 1954 Chevy, 19 mpg the same as the newer six passenger cars in my neighborhood. The airplane is a 1946 Aeronca same thing better mpg than most new production machines, the runway is maintained with a 1941 Ford 9n it works great as well, and is as reliable as the late 90′s vintage Kobota that is used at times. My point is newer is newer, NOT better, stone simple and reliable, strong machine are what this countries engineers created back when we built and designed here.

      • Thomas Boyle says:


        I have no objection to your desire to operate only government-certificated aircraft (although, it seems you don’t). I do think your faith in its benevolence rather, er, ill-considered, but there it is. For myself, I have far more faith in for-profit enterprises (they have a lot to lose) than I do in the government (it doesn’t).

        I want only to be allowed to buy a product from a willing seller, without having the government show up and arrest everyone involved. (Indeed, if I bought an airplane you had built as E-AB, relying on your reputation and my mechanic, we could do exactly that.) I have no objection to a requirement that the product be appropriately labelled, and could accept a requirement that it not be used for commercial purposes.

        And, yes, I could go along with a transition period. There is no private infrastructure, right now, equivalent to the Institute for Highway Safety, for example, and it would take time to put those in place. Of course, the FAA could compete in that market, and be the first in the game.

        • Thomas Boyle says:

          I should qualify – reputable, established for-profit enterprises.

          Yes, if you buy an airplane from a fly-by-night outfit, you should expect much the same result as you could expect if you buy a car, boat, medical drug or bridge from one.

          • Jeff Boatright says:

            “I should qualify – reputable, established for-profit enterprises.”

            Like Ford, GM, Toyota, Audi, etc. ?


            But I get what you mean and agree. Set standards, let the manufacturers figure out how to meet them. Works in most arenas, should work in GA – and largely does in the Experimental arena (though I’d argue that the Experimental ‘customer’ is more likely to be uniquely educated compared even to the average GA customer).

          • Thomas Boyle says:


            “Like Ford, GM, Toyota, Audi, etc. ?”

            Yes! All of those companies make fine vehicles. Personally, I prefer Toyota, Audi and Ford.

            DeLorean, on the other hand… :-O

  9. Joe Weaver says:

    Perhaps what we need is a simple exception or allowance that would let owners of certified GA airplanes, that are not used commercially, to install non-certified avionics. An “Experimental” aircraft needs to have a placard with that word clearly visible to all who enter the aircraft. The FAA could easily allow owners of certified non-commercially operated GA planes to install a placard on the aircraft that states “Non-certified Avionics are Installed in this Aircraft”, and let certified owners enjoy the bounty of low cost, high quality avionics that the experimental world now enjoys. As long as any non-certified avionics equipment is installed in the certified aircraft the placard must be installed & visible. When/if the owner decides to remove the non-certified avionics and install fully certified equipment, he/she can remove the placard with no change to the certified standing of the airworthiness certificate. This would solve a lot of issues for a lot of people. What would likely happen is that most, if not all, owners of non-commercial GA planes would end up installing non-certified avionics and never remove them. We’d all have that placard on our planes, and be much happier and safer.

  10. David Forster says:

    Is this initiative part of the results of the FAA-sponsored Part 23 Small Aircraft Review Committee recommendations?

    From what little I know, this committee recently submitted a final report and there has been legislation introduced to require its implementation by 2016, but I have seen surprisingly little coverage from the EAA on the subject. As an initiative which holds the promise of reducing cost and increasing the capability of avionics and parts for all light certified airplanes, I would have thought that coverage, if not support of this initiative would be a high priority. . .?

  11. Mac says:

    Hi David,
    The recommendations to change FAR Part 23, the rules governing certification of new “small” airplanes, is something EAA and all other aviation groups I know of support. The issue in certifying autopilots for older airplanes does not involve Part 23 directly because autopilots are approved on an STC, which stands for supplemental type certificate. An STC is a supplement to the original type certificate, in other words, coming after the basic airplane is type certified.
    The FAA could create rules to approve autopilots in older airplanes that don’t apply the complete standards of FAR 23. Few of the older airplanes in question are certified under Part 23. Most were approved under the earlier CAR 3 rules.
    The FAA has found ways under the existing rules to grant broad multi-model STCs for such things as WAAS navigators, flat glass PFDs, AHRS, ADS-B and so on without changing FAR 23. I think it can and should do the same for modern autopilots. Creating a new small airplane certification standard is a long term project, and a worthy one that would be a boost for new airplane manufacturing, but it won’t do anything for the existing airplanes, and a new FAR 23 will require years at best to be created.
    Mac Mc

  12. Avionics that goes into certified airplanes should be certified.
    1. Benefits need to compared to costs and risks. A new system that greatly increases safety does not have to be perfect. The standard for an airliner in commercial service is gross overkill for an airplane that carries up to four people, and maybe flies 100 hours a year.
    2. Certification standards and processes need to be updated to reflect current technology.

  13. Rodney says:

    As far as certification of aircraft goes it seems to me that many many aircraft certified under CAR 3 are still flying so why not go back to a similar standard for those simple aircraft or adopt a rule similar to the ASTM standard for LSAs when dealing with simple basic aircraft. Obviously an airliner is a different animal. For Auto-pilots the rules for getting a approval from your FSDO should be changed. It should be relatively easy to put in and adjust an auto-pilot to your plane. Adopt a 10 hour flight testing of all modes prior to carrying passengers. Document it and supply the documentation to the FSDO and get approval within a month or two. Unfortunately the current system makes any type of approval for any installation without an STC almost impossible. The attitude often seems to be. “if no one has made an STC for it or certified it there must be a reason. It must be unsafe.” or they are worried about getting sued. Changes could easily be made but no one is interested in making them.

    • Greg W says:

      Indeed the field approval process has been greatly restricted because of lack of knowledge at the FSDO’s, too many unsafe items had been approved. This was caused by us, they deferred to our judgment, as the people in the field, and we said far too often “it’ll work fine” when we really did not know. They now work from a rather short list of acceptable mod’s and anything else must get a “one time stc.” The engineering now must be from a paid private D.E.R (designated engineering rep.) as the FAA certification office will no longer do this. Perhaps a listing of avionics and autopilots meeting a T.S.O. standard could be “blanket” approved is similar to an approved installation. This would be like installing seat belts or aircraft skis.

  14. Tim says:

    I’d like to read all these replies but I am studying for my IFR written and really need to bone up on all the ADF and RMI questions right now.

  15. SkyGuy says:


    Certain older Cessnas had the option of a Brittain:
    1. Wing leveler.
    2. Wing leveler + tracker.

    Brittain is still in business in Tulsa, OK.

    They have a website.

  16. Bruce Stein says:

    I agree with your comments 100%. I have a light twin without an autopilot at this time. While it does make me a very sharp hand flying IFR pilot I would like to install a capable two axis autopilot. The problem is my lack of willingness to part with over 20K for and STC’d unit when an ‘experimental’ autopilot could be purchased for 8k that is as capable or more so that the FAA blessed unit. There is no reason that reasonable standards cannot be set that would allow for more options and therefore more competition in the avionics available to GA. As long as it has circuit protection and can be shut off in at least 3 ways I would have no issue with installing a unit not consecrated by the FAA.
    I also really like Joe Weaver’s suggestion regrading a placard in aircraft without ‘certified avionics’. If I could change my Airworthiness Cert to Experimental I would do it in a heartbeat and end the issue once and for all.

  17. SkyGuy says:

    See my post above about “cheap” autopilot.

  18. Patrick Underwod says:

    Mr. Boatright,

    With respect, I think you’re missing the point when you imply that, without the current level of government control and regulation in the US, we would automatically descend into third-world levels of corruption.

    You might take the following as a racist argument, but it is not. It is a cultural nd historical argument. Lower levels of corruption in the US than in the third world are not because we have a strong government, but because we have a strong culture of honest dealings between strangers that far predates the formation of our government (and, in fact, heavily influenced the way that government works). That culture is not strictly “American”, it is Anglo-Saxon and Christian.

    (I used to be dogmatically liberal, so just typing out that sentence fills me with trepidation. But there it is. And again, it has nothing to do with race; but it does go against the prevailing multiculturalist wisdom, ironically another product of our cultural pressure to deal fairly with those of other cultures, often without reciprocity; So consider me an ugly old racist if you like. Heck, I’m agnostic, but one would have to be willfully obtuse not to appreciate that the confluence of Anglo-Saxon culture and Christianity are historical determinants of American life ways.)

    Many third-world countries have beautifully constructed constitutions, with well-crafted laws on the books, and yet still suffer endemic corruption–because corruption is an accepted way of doing business in many non-Anglo cultures, and has been for thousands of years, regardless of governments and laws.

    It would be patently ridiculous to assert that Anglo-Saxons are immune from corruption. But our culture stresses more than most that people should empathize with others and deal fairly with them.

    Less government, including less, or at least more rational, FAA regulation, in the United States would (in general, and not without ugly exceptions) work to the strengths of the American people, not against them.

Leave a Reply

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


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>