Glass Cockpit Reliability

Primary flight displays are all but standard equipment in new airplanes, although some doubt their reliability. Avidyne says those worries are misplaced.


By Paul Bertorelli

In the panel of every state-of-the-art GA airplane, youll find an eloquent answer to the question, what if…? Its in the form of three steam gauges somewhere in the panel, an iron gyro, an airspeed indicator and altimeter fed by a conventional pitot/static system. By implication, the industry or the FAA must therefore think PFDs arent reliable enough to do the job without backup. Carrying the thread forward, one might conclude PFDs just arent that reliable to begin with.

Not surprisingly, youll hear a strong argument to the contrary from the manufacturers of PFDs, not to mention owners who are flying them daily in new airplanes.

We recently interviewed Avidynes Dan Schwinn, whose Entegra PFD is both the first and the leading primary flight display, although Garmin is rapidly catching up with its G1000 system. We flew with Schwinn in an Entegra-equipped Cirrus and spent a couple of hours discussing a wide range of avionics-related issues.


Weve received a few calls and e-mails from owners of new airplanes reporting failures of primary flight displays. Is this a tip-of-the-iceberg thing? In your view, how reliable is the basic technology?

You have to separate it into reliability and durability. When you look at electronic systems, they have a random failure rate, but they dont wear out. The backlight may get a little dimmer over 10,000 hours, but basically, these things dont wear out.

The mechanical systems in general have wear-out mechanisms and thats one of the major drivers. Thats why you talk about needing a new vacuum pump every 500 hours. So, in order to answer the question of which is more reliable, youd have to be able to separate the random failures from the wear out and I dont think I can make that assumption. At a certain number of hours, you rebuild a mechanical system and a certain number of hours beyond that, youre living on borrowed time. Theres no concept like that for electronic systems.

Will PFDs ever be certified with no mechanical back-ups?

Absolutely. The thing you have to keep in mind here is that when you replace a six pack, you replace six instruments. You can fly on maybe any two of them. So you can have a dual failure in a six pack and depending on which two it was, you might be able to continue to fly.

A single ADAHRs is pretty reliable, but its doing all of those things. When you get to dual ADAHRs in certain classes of airplanes, youre going to see OEMs certify airplanes with no mechanical back-ups. Its not going to be an FAA issue, its going to be market driven. The FAA allows dual ADAHRS without mechanical back-up.

Is that what the Eclipse jet will have? It has no mechanical backup.

Thats right, its two ADAHRs driving three screens. Theres a lot of crosschecking going on in that system and level B software. The Eclipse display has an attitude display of some kind on all three screens at all times. The dual redundancy on the attitude sensing and triple redundancy on the attitude displays means that if you work the numbers, it gives you a pretty high level of relativity. That will trickle down to smaller aircraft.

But what about a screen failure on the PFD side in, say, a Cirrus? If that happens, it takes everything down.

What happens when you put a lot of functionality on one screen is that the screen is probably more reliable than any one of those steam gauges. On a failure-rate basis, then, that makes it more reliable than all of them. But the screen does have the potential to fail and take out everything, while the steam gauges are completely unrelated systems. And thats why you end up with back-ups and thats what drives the whole reliability equation that you have to certify to.

Still, we heard from a Lancair owner who had his Entegra removed three times for repair.

The fact is, those guys exist, but they are statistical outliers. You cant really look at any one guys experience and know anything. But still, you cant let a customer get there. You cant let him have to go through a second installation. We just havent got that quite mastered yet and we are working on that.

Does that mean that the service network hasnt caught up with the market penetration of PFDs?

One the biggest problems we have is that the service center experience with these systems it low…and I dont want to make that sound like a negative. Theres a pretty good chance that if someone brings a unit into a service center, it will come back to us. The service center just doesnt know that the problem he may be seeing is, say, a wiring issue. We havent really provided the capability for dealers to get up to speed on these things and we need to do that. If you roll into a little airport and youve got an issue with your PFD, theyre probably going to say, Okay, well pop it out and send it back.

Thats one part of the problem. The other is that when he gets it back, he has magnetometer calibrations and any kind of stuff you need to do on a standard inertial system, which theyve never seen before. Theres better a probability of it being re-installed incorrectly than there would have been for a traditional instrument.

But if an owner cant get ready service, isnt that a form of poor reliability? How is this being handled, say, for the Cirrus?

Youve probably been assigned a Cirrus service center. You can go to any authorized Avidyne service center, but the thing is, we havent trained every service center on the Cirrus, so were working on a dealer program that allows them to get certified for various products.

We just yesterday had our second product quality support summit. We had the first one about nine months ago. We were really focused on the factory, getting our stuff built and installed right. We accomplished those goals. We know the products are getting installed properly at the factory. We know theyre not going out the factory door with problems.

But if you go on the COPA (Cirrus Owners and Pilots Association) site, youll see owners who arent happy. Probably the thing you most see is that a guy couldnt get it fixed right on the first try. Thats just so infuriating, so thats one of our biggest initiatives right now. We need to find out what happened. Did we have the wrong data on the failure? Did it not get installed correctly? Did he have the same failure twice?

Speaking of which, you used the term mean time between unscheduled removal or MTBUR. What is that number for the Entegra?

We wouldnt respond to that for competitive reasons. But there is one interesting number that we can share. Something like 20 percent of stuff that comes back to us for repair is classified as NTF or no trouble found. Right now, thats one of our biggest concerns.

Does that then suggest that there are bugs in this system you cant find or do owners report problems that really arent problems?

A phenomenon that weve found interesting…and the Cirrus guys have pointed this out to us…the pilots are very, very reliant on the PFD. One of the things weve discovered is that the sensitivity to anomalous behavior, real or perceived, is extremely high. You couple that with service centers who dont know the nuances of the product because its basically a new product to them, then you have people who are going to have more questions than they would about something thats been around for forever.

The true failures we have experienced, as you have discovered, have been-the vast majority-issues that come up that arent flight safety related or they appear at start-up time and they can get it fixed. Owners are annoyed because of the service center issues Ive mentioned.

Ill give you a for example: If you put the magnetometer back in wrong-which you can do, you can put it in offset in any airplane-the heading is going to be wrong. In an AHRS airplane, it couples to attitude, so you can have a three-degree attitude error because youve got your heading source wrong. And owner is going to notice that and find it irritating. Its those kinds of things that need to be taken out of the customer experience.

Some failures have to be real. What breaks on these systems?

One of the bigger items is cables. In earlier units, we had a cable termination that would come loose and the display would go dark. Also, air data units. This isnt so much a failure, but weve had air data units drift more than they should and they have had to be recalibrated. Thats probably our number one problem right now. Then you get into miscellaneous items. Number one was the display went dark, number two was NTF and number three was air data. The rest are statistical anomalies.

Are these design or production issues?

Well, the way you fix them is to change the design. For example, for the cable terminations, we changed the way the connector terminates. Was the design deficient? Theoretically, it wasnt going to last as long, so we changed it. If something gets assembled wrong, you may want to go with a process fix or just design out the chance of it ever happening again.

How reliable are the displays?

The displays themselves typically dont fail. The backlight, which is a fluorescent high voltage system, can flicker. Not so much anymore, but we had problems initially. The actual LCD is reasonably close to indestructible. Theyre built by the zillions for laptops and they have a little bit of ruggedization for our application so they just dont break. Its the same thing with the gyros. The gyros are from cars and they just dont break.

Speaking of cars and breaking, I dont expect to ever see the radio in my car break. Why cant the avionics industry deliver that kind of quality?

For a lot of owners, their standard is not the 30-year-old Cessna they used to have, their standard is the Lexus they just bought. They have a higher level of expectation in our industry than we have been able to deliver. Quite honestly, I think they have a reasonable level of expectation but we-and by we, I mean Cirrus, Avidyne and our suppliers and our dealers-dont have the ability to deliver to those kinds of expectations.

Why Not?

One primary reason is volume. Its really hard to deliver that kind of an experience at low volume. I dont know if its possible. But I do know that thats the standard for a lot of Cirrus pilots… I paid $80,000 for the car, I paid $400,000 for the airplane, I expect perfection out of this thing. They want a very high touch experience and that is not what we are getting them.

So, when you see Cirrus selling thousands of airplane instead of hundreds, its because they decoded that. Theyve already blown away everybodys expectations of what volume they could make. Theyve got the best-selling airplane on the planet.

In order for general aviation to get out of a declining, oddball, hobbyist sort of thing, youve got to upgrade the experience hugely. GA is becoming more a niche thing. The population has gone up 50 percent in the last 40 years, but GA has been dead flat. When you consider the population increase, its gone down.

Also With This Article
“Cirrus Accident: A Second Look”

-Paul Bertorelli is editor of Aviation Consumer and editorial director for Aviation Safety.


Please enter your comment!
Please enter your name here