To Pull or Not to Pull

It’s hard to believe a full decade has elapsed since the launch of the GA glass panel revolution. But as I recall, the first relatively high-volume GA aircraft with a fully integrated glass cockpit was the 2003 edition of the Cirrus SR22. That was the same year that Diamond brought the Garmin G1000 suite to their DA-40. The race was on, and we haven’t looked back since.

While this technology is a blessing, it’s also more complex than traditional analog gauges. Each product line has it’s own failure modes and redundancies, it’s pluses and minuses. Those are the things which dictate how partial panel scenarios should be simulated. It ought to be based on the way failures are expected to occur in real life, right?

With the Entegra suite, for example, that means dimming the Primary Flight Display. It has no reversionary capabilities, so whether the offending component is the display itself or the AHRS, dimming the PFD forces the pilot to fly with (and without) the appropriate equipment, as I discovered during my own partial panel situation in 2011.

As the years have gone by, Garmin’s G1000 suite has claimed a larger and larger share of the market. You’ll now find the G1000 and it’s variants in everything from diminutive LSAs to Citations and Learjets. Because the Garmin suite has reversionary capability, dimming the PFD isn’t good enough. During a real-world failure of that screen, the MFD would automatically display the critical flight instruments. In other words, Garmin has turned display failure into a non-event. Therefore, the critical scenario in the G1000-equipped airplanes has been the loss of the AHRS sensors.

Since no aircraft I’m aware of contains a switch for shutting off the AHRS, instructors learned to disable the sensors by pulling the appropriate circuit breaker. This forces the trainee to fly off the standby instruments, just as they’d have to do during a “real-world” loss of the primary flight instrumentation. The system seemed to be working just fine.

But a few years ago, the FAA (or perhaps it was the FSDO in my area) decided that circuit breakers should never be pulled to simulate partial panel scenarios and passed that guidance along to Designated Pilot Examiners. The new recommendation is for instructors and examiners to dim the Primary Flight Display, even though that does not accurately model an AHRS failure in the Garmin installation.

In fact, Garmin has published similar guidance in their instructor/examiner guide. While they don’t expressly prohibit pulling a circuit breaker, it isn’t exactly encouraged.

Pulling circuit breakers—or using them as switches—has the potential to weaken the circuit breaker to a point at which it may not perform its intended function. Using circuit breakers as switches is also discouraged in Advisory Circulars 120-80, 23-17B, and 43.13-1B. Additionally, a circuit breaker may be powering other equipment (such as avionics cooling fans) that could affect the safe operation of other equipment.

The guidance from those Advisory Circulars is not new. But since the FAA changed the way examiners administer flight tests, it has forced instructors to alter the way they simulate partial panel scenarios with students. And my concern is that they are no longer being properly prepared to use the information in front of them during an actual AHRS failure. That’s a huge problem, because a partial panel situation in actual instrument conditions can be a serious emergency if not handled correctly.

I’m not the only one who feels this way. AVweb recently published an article positing that “realistic scenarios make more sense than challenging the pilot with failure modes that are unlikely to happen in the real world.” I couldn’t agree more.

This problem is specific to the Garmin suite and others which operate the same way. Avidyne’s Entegra utilizes separate GPS/NAV/COM radios, and those displays will remain functional during a partial panel emergency. That provides an additional place to display vital navigation data, as on the default nav page of the Garmin GNS4/500 series. But on the G1000 suite, everything is integrated into two main displays. There is no CDI available anywhere else, so if the AHRS fails, the Primary Flight Display remains critical because that’s where course guidance information will come from.

Also, consider the fact that the G1000’s HSI becomes a fixed-needle CDI when heading data is lost. Should we be training pilots to fly with that presentation? How are we to do that if the entire display is dimmed?

Dimming the PFD is also unrealistic because while the reversionary mode would bring up the HSI on the MFD, it would also display all the primary flight data that would be unavailable in a partial-panel situation. Dimming the PFD means giving up your only source of course guidance sufficient to be used for an instrument approach. Without the HSI, how is a partial panel aircraft supposed to navigate? I’ve queried Garmin about this and never received a satisfactory answer.

Some instructors have settled on using stickers, hanging cardboard cutouts over the display, or other such methods. I’ve tried those and found them to be problematic. Stickers leave residue on those $10,000 displays. They fall off. They interfere with the pilot while they’re being placed on the display — it’s not anywhere near as sudden or simple as the old 3″ instrument covers. Cardboard or plastic overlays are cumbersome and tend to shift or drop off in turbulence. And they’re a real pain to store when not in use. I bought one overlay, designed another, searched the Web for a better mousetrap, and still came away without a good solution.

As much as I hate to say it — because Garmin and the FAA do have a point — pulling that darn circuit breaker still seems like the best option. It’s often within easy reach of the right seater, and provides a higher fidelity simulation to the trainee.

While I grant that circuit breakers may not be designed for use as switches, how often are we really simulating this scenario? If the condition of the CBs is a concern, I can’t imagine a dedicated CB/switch combination with a heavy-duty guard to protect it from accidental deactivation couldn’t be installed or designed to solve this problem.

The difference in instrument scans between functional glass and partial panel can be significant. Whether you favor pulling circuit breakers or not, one thing we should all be able to agree on is that pilots should not be left to discover the intricacies of instrument failure until they experience it on a dark and stormy night.

  12 comments for “To Pull or Not to Pull

  1. April 28, 2014 at 3:25 am

    While it was on steam gauges, I always used sticky notes. (Not sure if this is what you mean by “stickers”.) Cheap, easy to keep under the pad on my kneeboard, reusable, and recyclable.

    The two key issues here are training for the reality of only one instrument failing and improving the student’s overall stick and rudder skill. The first is what you address very well here. The second is about making a better pilot. No one instrument tells the whole story. During my tour as a helicopter instructor, we found (anecdotally) students flew basic instrument maneuvers better without the attitude gyro because they spent more time watching those things, e.g. ball and VSI, the gyro doesn’t tell you.

    • April 28, 2014 at 9:33 am

      Sticky notes are kind of like duct tape: they can do almost anything! Unfortunately, the horizon line is so large that it would be extraordinary difficult to cover it and still leave the ADC information available. It also wouldn’t solve the problem of transforming the HSI into a fixed-card CDI, which is how the G1000 responds to a loss of the AHRS.

      I know exactly what you mean about trainees flying better without the attitude indicator than with it. Once that goes, they know they have to be “on their toes” and scanning the remaining instruments properly. Plus, there are actually fewer of them to look at, so the scan is simpler. Ironic, isn’t it?

      • April 29, 2014 at 3:24 am

        Not to belabor the point, but stickies come in an amazing array of shapes and sizes these days.

        More seriously:
        – The combination of the old attitude and heading gyros into AHRS makes sense in a lot of ways until you want to do the kind of training we are talking about. Too bad there can’t be an instruction mode in thef aircraft. Too much risk.
        – I think partial panel in a helo is different than in a fixed wing aircraft. In both cases, the gyro is bolted to the airframe, but the flight control surfaces (rotors) don’t have a fixed relationship with the airframe in a helo like the wings do in a plane. This makes the rate instruments much more important because they show where the whole aircraft is going, i.e. turning, climbing/descending, accelerating/decelerating.

  2. Mark S.
    April 28, 2014 at 2:08 pm

    Ron, I still remember clearly my shock that night when you pulled the AHRS breaker while I was flying under the hood in the DA-40 with you. I for one appreciate the realistic training I received that night.

    • April 28, 2014 at 9:44 pm

      I’m glad the training was helpful Mark. Hopefully you’ll never see that happen in actual IMC!

  3. kachius
    April 30, 2014 at 4:22 am

    Nice post! As a software engineer I can say that it is quite easy* for Garmin and the rest glass makers to provide a “training mode” simulating as little (or as many) real life failures they wish, without actually shutting down anything. Just “messing” with what will be displayed in each training mode. Instructor’s side could be left fully functional. PS:”quite easy” = technically. How much it will cost them or how much they will be willing to sell it is another issue.

    • April 30, 2014 at 7:50 am

      Good point. Anything is possible — after all it’s just computer code, right? I guess it’s just a matter of cost, liability, and regulatory hassles. I’ve flown quite a few glass airplanes, but never seen any sort of training mode, even on the non-certified models. Of course, most panels don’t have the reversionary capability that the Garmins boast (why is that, by the way?), but I wonder if such a mode has ever been considered.

      For jets and even turboprops, where most of the training takes place in a Level C or D simulator, there might not be a need. But most light GA training takes place in the actual aircraft.

  4. Siegfried Lenz
    May 1, 2014 at 12:20 pm

    Interesting question. The most Sound version would be some kind of ‘training failure mode’ allowing the instructor/examiner to fail specific parts of the display/system. I could imagine that manufacturers are reluctant to install this for fear of litigation if it were activated inadvertently (I remember a Flight on the right seat of a 210 dedicated to advanced instrument training where my teenage fiddling with the dedicated ‘Radio fail’ switches resulted in a go-around as we never got a GS indication….). Best solution might be a ‘training box’ handset-like device that can be plugged into the system for training flights and almost selection of failure modes.

  5. Anony Mouse
    May 1, 2014 at 12:49 pm

    IF one is going to do this, it needs to be set up so that it self-expires after 45 minutes (or some reasonably short period). That way, when the next person gets in the plane, things don’t start failing for no reason.

    Next, training mode w/ failues needs to be something you can’t turn on accidentally. So, how do you put in a password/keyword, or whatever, that enables this? Should it be done with a special memory card that is ROM and only contains — TRAINING MODE — so that it is not confused with “grab what’s there and load the data bases with it” type data.

    Yeah, I’m also a software developer. I work on z/Frames and know this to be true: “It is impossible to make anything foolproof, because fools are so ingenious.” –Robert Heinlein

    And so also being a pilot, I like this idea of training mode. But training mode getting turned on in IMC with moderate turbulence? NOW is not the time to accidentally enable test mode — particularly for someone who has not had failure mode training with this system.

    • May 1, 2014 at 4:05 pm

      You’ve nicely stated some of the reasons I had in mind when I wrote “too bad there can’t be an instruction mode” in my previous comment. Allow me to add one more situation: inadvertent IMC during simulated IMC under VFR. You would have a real emergency on top of a “fake” one. The instructor would likely be disoriented and the student unaware of the circumstances. Additionally, this scenario (simulated IMC under VFR) is most likely when the student is new to instrument flight and will not have the experience to help an instructor who has lost situational awareness.

      • May 2, 2014 at 11:58 am

        The training mode idea is an interesting one. I guess the real question is, which one presents less risk: having a simulated failure mode built into the G1000, just pulling the circuit breaker, or not training for AHRS failure at all?

        As for the inadvertent IMC encounter while simulating partial-panel under VFR, I don’t think that’s too much of a risk. For one thing, all these glass cockpits have a separate attitude indicator (usually analog) which would be unaffected. All that would be required would be 180 degree turn in the soup using that instrument. For another, neither partial panel practice or actual IMC are that commonly encountered. For example, about 3% of my flight time is actual IMC, so statistically any failure is 97% likely to occur in visual conditions.

  6. Graeme
    May 1, 2014 at 9:41 pm

    I don’t see why, since the whole suite is software anyway, why they can’t just program in a training mode where devices could fail? I see the problem. You can inadvertently flip a switch and not know you did, and now you can’t get your device back, but maybe if you entered a three digit pin, to enter the mode, like 123, and then you could fail and unfail it, or even simpler, make ANY soft key revert back to normal mode, once pressed.

Leave a Reply to SaferAviatorCancel reply


Get the latest posts delivered to your mailbox: