Are Glass Panels Really Unsafe

[N395V]

Someone mentioned the "huge" price premium of glass over steam guages.

Just got out the catalogs and looked for steam guage bargains with at least recognizable manufacturers names and did some price comparison.

I assumed the following:

Both steam guage syatems and EFIS based systems would require similiar cost in transducers, wiring, and plumbing.

The steam guage system would require a vacuum pump and the EFIS a processor.

I used my basic G4 BMA Sport as a Comparison although it could be done for less by using other systems.

For a basic VFR equipped aircraft: A BMA Sport EFIS system for flight, engine instruments, and GPS NAV:
$8,200


Steam Guage Basic VFR aircraft with basic in panel GPS NAV

$7,150

Add the equipment necessary to bring the BMA sport up to full capability,
autopilot, fuel flow, all cylinderCHT/EGT, coupled AP and then do the same for the steam guage plane and the results are:

EFIS $11,700
Steam $ 12,150

Make them both IFR capable just add the price of your desired backups to each syatem.

There is no $10,000 penalty for a glass panel in an experimental and there appears to be no subjective evidence that a glass panel is any less safe than a traditional panel.

As an aside I found that wiring the EFIS was far less of a PITA than wiring and plumbing a steam panel.

 

[Rainier Lamers]

It is my opinion that terrain view and synthetic vision on these instruments is a huge waste of processor time. Sure it will help you fly between mountain peaks but I find it distracting to view and in the event of engine failure it is not going to tell you anything about fences, power lines, signposts, trees, houses, cars, ruts, gullies, rocks, etc.
To purchase an EFIS with this in mind would indeed be foolhardy.

Wow, that is quite a statement !
So Chelton, BMS and us are wrong in your opinion ? OK, your're perfectly entitled to it.
If I interpret you correctly, you would rather guide down your aircraft using a six pack in zero viz than look at a synthetic terrain (which BTW can have obstacles showing as well).

Well, that is certainly a new view to me.
Yes I am biased - but I make no excuse for that.

The synthetic terrain will not help you close to the ground if you are still not able to see anything. That is obvious - but that is not the purpose of it at all.

Forgive me for being unrepentent on this one - I'll stick to my fancy EFIS any day. It might not guarantee me escaping a crash landing but at least it gives me a good idea just where that will be and I can make an early choice. In addition the infrared night vision option we are currently working on is not exactly worthless either. The terrain look ahead vertical profile shows me exactly if I'm going to clear that ridge and by how much using my current real decent rate and ground speed - even if I am completely blind. That is a **** of a lot better than seeing nothing at all.
As to wasted processor cycles - I'll gladly waste them on whatever I can dream up to help, aid or in any other way increase my flying pleasure. What else should I do with them ? Stop the CPU to conserve power ?

Is the display distracting ?
yes, you are right it is. In a way. Every pilot I have so far taken for a flip takes one look at this and instantly sees the value it can give you. The value is not just in zero viz but perfectly apparent on a normal day on a normal flight. No guessing where the airfield is in the haze - you see the **** thing right there on the screen complete with runway designations in 3D - before you see it out of the cockpit. You see mountains and the general outline of the landscape even if much of it is hidden in clouds or fog. I can "see" an airfield right though a mountain ridge and know exactly where it is in relation to where I am - just as if I where to have x-ray vision to see right though the ridge.
Not once have I come across a negative comment about this (until now that is) - but just in case, we have put a litte setup in the 3D setup menu where you can simply switch the 3D stuff off. I bet that is the most under used setup we have in the whole instrument.

In my opinion, there is no such thing as too much information a pilot can get, regardless of the current situation. The challenge is to present this information to the pilot it such a way that it is cystal clear and prioritized. Information must be presented in a way that does not require interpretation or assumptions. Many modern EFIS systems do an admirable job here.

EFIS systems really started off being "fancy" moving maps or more basic navigation systems with some primary flight and attitude stuff thrown in for good measure. They have progressed far beyond that and you have not seen the end of it. We have only just started...

Rainier
CEO MGL Avionics

 

[nicolcarstens]

Times / technology / prices do change!

Hi Milt,

I went cheap for my backup electric horizon. It went TU before 2 hrs of flight time. Should have bought a good one. Got what I paid for.

I don't want to start a fight here, just want to highlight one line of thought... with regards to "you get what you pay for". (obviously we all have our own definition of what "cheap" is - a discussion on its own)

But irrespective or your definition of "cheap", times/prices/technologies are changing. That goes for all of them: GPS receivers, color displays, AHRS, EFIS, etc.

All I would like to say/stress: make sure you update your definition of "cheap" and "reliable in spite of reduced cost" every now-and-then...

For example: you know that there are FAA certified AHRS systems (no aiding required) running on MEMS sensors these days? You have any feel for (today's) prices of these high quality sensors? (Granted: you can still manufacture a BAD product from GOOD sensors!! And not ALL MEMS sensors are any good!!)

You familiar with (BAe) Silicon Sensing? http://www.siliconsensing.com/index.jsp?articleid=9

I will leave you with that thought...

Kind Regards,
Nicol.
Engineer with MGL Avionics.

PS: Ask my girlfriend and you will hear: I am the last guy to buy cheap (if I can afford not to)! I rather spend the "extra" money IF i KNOW that it will give me a better product... BUT why pay more for less, or even the same?

 

[Yukon]

Those who live in Glass houses.........

Hey, you guys are forgetting one of the main factors in airline electronics reliability. 3 (or more) generators! For all practical purposes, a total electrical system failure is statistically impossible. Just doesn't happen.

The MTBF of our toy electrical system is probably measured in hundreds of hours, and is definitely the weak link in our system. Adding a 10 amp vacuum pad alternator is a good move, but not much power. If you are using pitot heat, you are out of luck.

Lightplane IFR should be (mostly) limited to punching through an overcast and
high minimums approaches. Single-pilot, single-engine hand flown instrument approaches to minimums are not good for your health. Airlines are required to
use the autopilot and or flight director for low-vis approaches. This is where the safety comes from, not from the glass screens. Glass presentations of aircraft attitude, airspeed and altitude are little different than the mechanical kind. An airspeed tape is no better than an analalog presentation, just more compact.

No George, Southwest doesn't use autothrottles or V-Nav. It is their contention that these systems degrade pilot proficiency, and it appears David
D concurs. Still have the 6-pack presentation too!

 

[nicolcarstens]

It is my opinion that terrain view and synthetic vision on these instruments is a huge waste of processor time. Sure it will help you fly between mountain peaks but I find it distracting to view and in the event of engine failure it is not going to tell you anything about fences, power lines, signposts, trees, houses, cars, ruts, gullies, rocks, etc.

I am sure the wives of those pilots that have been killed flying into mountains would strongly disagree with you...

(and then like the boss said: one could display these items too if they were included in the database)

But: have you ever done a "hit probability study" for cars, houses, trees, etc vs mountains given you are flying at 300 ft AGL? (not that I have done this one...) Point being: let's worry about the big things first!

I am not saying go look for trouble either, but we all find our way into it (some time or the other) without trying hard. If you can see altitude, speed, attitude, RPMs, mountains, etc all on ONE screen, surely that must add value to a low hours pilot? (vs having to dig out a paper map, fly the aircraft, read contour lines, read altitude, etc, etc)

Cheers,
Nicol.
Engineer with MGL Avionics.

 

[Yukon]

OOPS!

NTSB Identification: DCA07WA043
Scheduled 14 CFR Non-U.S., Commercial operation of Ryanair
Incident occurred Sunday, May 13, 2007 in London Control
Aircraft: Boeing 737-800, registration: EI-DCX
Injuries: Unavailable
On May 13, 2007, at about 1129 UTC, a Ryanair Boeing 737-800, EI-DCX, experienced a partial, temporary loss of flight displays after departure from Stansted. The flight crew declared a PAN and returned to Stanted instead of continuing on to Stockholm. There were 176 people on board. The Air Accidents Investigation Branch United Kingdom is conducting the investigation.


Full narrative available
Index for May2007 | Index of months

 

[rv8ch]

Oops!

Hey, you guys are forgetting one of the main factors in airline electronics reliability. 3 (or more) generators! For all practical purposes, a total electrical system failure is statistically impossible. Just doesn't happen.

That makes me feel better! Nothing can go wrong - go wrong - go wrong - go wro...

 

[L.Adamson]

I am not saying go look for trouble either, but we all find our way into it (some time or the other) without trying hard. If you can see altitude, speed, attitude, RPMs, mountains, etc all on ONE screen, surely that must add value to a low hours pilot? (vs having to dig out a paper map, fly the aircraft, read contour lines, read altitude, etc, etc)

Even the many thousand hour pilots are still whacking rising terrain. Since I've been interested, and keep track of this "phenomenon" for the last thirty years or so, I have quite a long list.

Doesn't matter whether the flight was IFR, or a VFR flight suffering a daytime whiteout condition, or an unseen cloud at night. Not to mention all of the common missed approach accidents.

In that brief moment of spatial confusion, a visual image of imminent doom with quickly rising terrain, can go a long way!

L.Adamson --- KSLC, where until recently, you could count on at least three flight into terrain accidents every year!

 

[rv6ejguy]

I can remember reading at least 4 reports to date of complete glass panel failures on big jets in the last 5-6 years or so and I'm sure there have been many more. Triple redundant or not, anything mechanical or electrical CAN fail.

I've read of many more GA GPDs taking a dump with only a few hours on them, one three times in about 400 hours- luckily all under warranty and all under VFR conditions. Many of these are clearly not ready for prime time- especially experimental ones. Along with these issues have been endless glitches and gremlins with a fair number which goes on for years in some cases.

If you are relying on this stuff IFR, I'd urge people to think about their electrical systems and backups, this part is just as critical as whatever electrically powered flight instruments you are using. VFR, no biggie if your GPD goes down, IFR- big deal. I've got twin alternators, twin batteries and twin busses just for VFR flight.

One thing that I've seen with GPDs with people flying VFR is that their heads are in the cockpit more pressing buttons and looking at the display compared to steam gauges- can't be a good thing. Many guys also seem prouder about their panels than their airplanes and I'm amazed at what some day VFR RV drivers will spend on glass, guess they have money to burn.

The nice thing about steam gauges is that they are already obsolete when you buy them and still have the same status 20 years later! Glass is becoming old news in less than 5 years. Another big bill to update then.

Know how to use the thing too. There have been some accidents IFR recently even among trained crews with approaches and missed approaches where they didn't know how to use the glass.

Finally when one gyro or gauge take a dump, you still have the rest to fly with. When your GPD goes blank, you are probably grounded until it is fixed.

Just my 2 cents, 3 layer Nomex suit on...

 

[Yukon]

That makes me feel better! Nothing can go wrong - go wrong - go wrong - go wro...

Mickey,

This appears to be a glass problem, not a generation problem. What say you?

 

[L.Adamson]

One thing that I've seen with GPDs with people flying VFR is that their heads are in the cockpit more pressing buttons and looking at the display compared to steam gauges- can't be a good thing. Many guys also seem prouder about their panels than their airplanes and I'm amazed at what some day VFR RV drivers will spend on glass, guess they have money to burn.

Actually, as a VFR RV glass panel driver, you punch in the route and destination. Click on the A/P, and then take occasional quick glances at the airplane symbol/moving map while making instant comparisons on your sectional, as well as XM uplinked weather in the distance. The rest, and most of the time, is spent scanning for traffic, as well as scenery, landmarks, and birds, that you never seemed to notice before.

I've seen this scenario too, and don't really go along with the head stuck on the panel idea.

L.Adamson

 

[rv6ejguy]

Actually, as a VFR RV glass panel driver, you punch in the route and destination. Click on the A/P, and then take occasional quick glances at the airplane symbol/moving map while making instant comparisons on your sectional, as well as XM uplinked weather in the distance. The rest, and most of the time, is spent scanning for traffic, as well as scenery, landmarks, and birds, that you never seemed to notice before.

I've seen this scenario too, and don't really go along with the head stuck on the panel idea.

L.Adamson

Ahh. smart man who knows how to operate the stuff and keep a lookout. Seems like too many new GPD owners punch buttons for no reason other than to apparently impress someone sitting beside them- look what this baby can do. Wheeee...

That's very nice but what about that Cirrus at 12 o'clock and 1/4 mile...


I'm thinking more along the lines of what Richard Collins has been harping about past few issues, especially flying IFR with glass- better know it thoroughly before launching into actual IMC. A daunting task with some of the units I have looked at. Them's is thick manuals!

 

[rv8ch]

Problems

This appears to be a glass problem, not a generation problem. What say you?

Me say I don't know yet, but as you rightly pointed out, lack of electricity can be a major problem for all glass cockpits. Electrical failure can come from many causes, here's one example: http://en.wikipedia.org/wiki/Air_Transat_Flight_236

 

[nicolcarstens]

Backup battery... (or batteries)

Me say I don't know yet, but as you rightly pointed out, lack of electricity can be a major problem for all glass cockpits.

True. But a modern EFIS (with all of its sensors) can go a LOOOONG way on one 12 Volt 7 Amp Hour SLA battery (hours to put some numbers to it!).

For example: an Enigma running off a single 12V/2AH battery will give you up to four hours to get out of trouble!! Let's assume you have a really bad battery and it only gives you one hour... that is a lot of time (on a small battery). Not so?



(Yes, I know: backup batteries also fail... Fit three: one onto each of your 3 EFISes! )

 

[L.Adamson]

That's very nice but what about that Cirrus at 12 o'clock and 1/4 mile...

The Cirrus with it's Avidyne glass panel & "skywatch" traffic avoidance system, has known where I am for several minutes now. No problem!

 

[Kevin Horton]

I am sure the wives of those pilots that have been killed flying into mountains would strongly disagree with you...

(and then like the boss said: one could display these items too if they were included in the database)

A friend provided the test aircraft for a NASA study on the effectiveness of synthetic vision (SV) for preventing controlled flight into terrain. The test aircraft was configured with an EFIS with on the right seat. There was also an MFD that could display threatening terrain ahead in red. They flew with quite a few test subjects, and had them fly a simulated IFR departures in mountainous terrain (Reno, NV). They varied the type of terrain display (EFIS with no SV, EFIS with SV, no SV but MFD with terrain display, SV + MFD, etc). My friend was disappointed to find that SV alone provided worse results than he was expecting - there were quite a few cases where he (as safety pilot in the left seat) had to terminate the test point before the subject pilot (under the hood) flew the aircraft into the mountain peak on the nose, with the mountain peak clearly displayed on the SV.

Universal did some testing with their prototype SV display, and found that they had to change the vertical terrain scaling on the EFIS to achieve acceptable terrain awareness by the evaluation pilots.

I'd be interested to know if MGL has done any testing to validate the effectiveness of their SV display, or is it simply assumed that SV will help increase safety.

 

[rv6ejguy]

The Cirrus with it's Avidyne glass panel & "skywatch" traffic avoidance system, has known where I am for several minutes now. No problem!

Right you are, but this Cirrus just had a complete GPD meltdown.




One thing to be aware of on the backup battery deal is the AH ratings. AH to dead is meaningless, you want to look at the AH rating to 9V which is where most avionics sign off. Most big backlit color displays draw a fair amount of current.

 

[nicolcarstens]

One thing to be aware of on the backup battery deal is the AH ratings. AH to dead is meaningless, you want to look at the AH rating to 9V which is where most avionics sign off. Most big backlit color displays draw a fair amount of current.

Very true. Your words of wisdom/caution are 100% appropriate on this point. Thanks for pointing it out to those that might be in the dark... (pun intended )

I provided an example for those that have no feel for the numbers. The exact numbers are not the issue here: that you can keep your modern EFIS running from a battery, for a GOOD period of time, was the point - don't require 3 alternators for all backup systems!

 

[rv6ejguy]

BTW: Enigma signs off at around 6V to 7V. Current consumption at full backlight and all sensors running is around 1 Amp

One amp is very good. Anyone know what a G1000 burns?

 

[Dynon]

Well, since this has just become a huge advertising thread for one EFIS company, let me just suggest that everyone buy a Dynon EFIS. You can get an internal, fully supported, self-managed backup battery that will last for hours without you needing to design your own backup electrical system If you're really looking for redundancy, you can hook that backup battery up as well into our emergency power input.

 

[SvingenB]

Redundancy

The reliability of a system with tripple redundancy compared with a single unit can easely be a million to one (or much more/less depending on operational requirements and some other factors). If the single system has a MTBF of 1000 hours, failure mode total blackout, then a tripple redundancy system can on average run for 1000 million hours before total blackout. It is something to think about when looking at the reliability the airliners want/need in their glass.

A distributed system (several independent instruments, like a panel with old fashioned steam gauges for instance) will also inherit similar reliability considering you still have enough information from your gauges to land safely if one or two of them fail.

 

[Rainier Lamers]

Well, since this has just become a huge advertising thread for one EFIS company, let me just suggest that everyone buy a Dynon EFIS. You can get an internal, fully supported, self-managed backup battery that will last for hours without you needing to design your own backup electrical system If you're really looking for redundancy, you can hook that backup battery up as well into our emergency power input.

Sorry, for the advertising. It's not meant to look that way. I have asked Nicol to delete or edit his post accordingly. Happy ?
This thread has had some very interesting inputs from all sides. I don't think it is possible to do this thread justice without some practical input. If I were representing Dynon, I'd still post similar comments to show "the other side".

We effectively have three parties participating in this thread:

a) The doubters
b) The believers
c) The guys that design and make the goods

You will agree that that is an excellent combination for all parties concerned ? It gives us valuable insite into opinions (which will, one way or another, find its way into our and perhaps your products). The other participants (and the many lurkers) get the benifit of information as to the cutting edge plus a wide and very welcome range of opinions.

I think this is a very good thread and it seems it is benifitial to everybody. I'd really like some of the other EFIS makers to join constructively - surely they have opinions as well ?

Rainier
CEO MGL Avionics

 

[nicolcarstens]

Hi Kevin,

Interesting link you provided. Thanks.

I'd be interested to know if MGL has done any testing to validate the effectiveness of their SV display, or is it simply assumed that SV will help increase safety.

We have never claimed to be experts in the field of "the effectiveness of synthetic vision"... I expressed (marked it): an opinion (like most of the issues surrounding experimental aircraft safety - statistics and opinions at best!).

Having done our own tests: it works for us. And there are a LOT of customers that are VERY pleased with it. People that don't like it at all are welcome to switch it off.

If I read your response, I read "SV alone provided worse results than he was expecting". In other words: "not as effective as expected"? Based on the little information we have, I would hardly call this experiment "unbiased science", don't you think? Your friend might have had to take over earlier (more often?) if the guy was using a paper map and struggling to find himself (while flying the plane), not so? But granted (an experiment nevertheless): it might highlight that the benefit of having "SV alone" might not be as big as one might assume! Point taken.

One last comment on the topic from my side: you don't have to go "SV alone"... why would you? Limited panel space? It is just one more tool that will make flying safer - our opinion!!

Without looking at numbers too much, do these picture not tell you a story? If I was surrounded by clouds, or had a layer of clouds between myself and the runway, I would have felt a lot more comfortable seeing the first picture!

vs

Even my mother than can't read a map, can understand this. If it was a large international airport, I would love to have had the "Avidyne 'skywatch' traffic avoidance system" data as overlay on the display too!! (h#$l, I asked for it 10 min ago! )

Don't get me wrong (and let's not start fighting about the merits of SV): I am not saying an EFIS that is not offering SV is no good! A lot of people are very happy flying with zero glass (let alone SV)! Clearly it is not required (if you know what you are doing, and you have good backups in place, there might be zero need for it!!). But to call "terrain view and synthetic vision" a "huge waste of processor time" ... hmmm ... we don't share the opinion! (... and we are NOT alone...)

PS: picture of a Chelton system used in the interest of non-adverstising

 

[rickmellor]

If the single system has a MTBF of 1000 hours, failure mode total blackout, then a tripple redundancy system can on average run for 1000 million hours before total blackout.

This type of redundancy will reduce the probability of total system failure but it won't extend the longevity of the individual components. All of the components of the triple-redundant system are running in the same environment for the same duration as all the other components and are thus subject to the same failure rates as all other similar components. For this system to reach 100 million hours of up-time, each individual component would need to be rated at 100 million hours. Once you account for variation about the mean for MTBF, your redundant system will only improve the probability of actually attaining it. It doesn't extend it.

-Rick

 

[Rainier Lamers]

Dead on !
MTBF is probably the must missused and missquoted item when it comes to judging reliability. It only has validity if every single component in the system has been taken into account with its own MTBF. To make matters worse, each items MTBF is only valid under exactingly specified conditions. This, and the fact that MTBF in its intended way cannot judge the reliability of increasingly "soft" systems makes the whole thing a little of a tea-leaf reading excersize. Many common electronics components (excepting MIL spec stuff) are rarely quoted with MTBF figures and if, it is doubtfull they have real meaning or are even traceable.
If you want to know about how reliable a panel is, one of the relatively good ways to get at least a vague idea is to use Google and a few stategic search words and phrases. I personally have more faith in that than in any quoted MTBF.
When I was involved with MIL-spec designes - we often found components that just did not fullfill the required MTBF, no matter how one would interpret it - while at the same time ordinary "cheap" components very often proved much supperior. This was especially true in the days of TTL chips.
Often we ended up using "normal" components and then qualifying the resultant design "as a whole" to specs (which includes MTBF).

Rainier
CEO MGL Avionics

This type of redundancy will reduce the probability of total system failure but it won't extend the longevity of the individual components. All of the components of the triple-redundant system are running in the same environment for the same duration as all the other components and are thus subject to the same failure rates as all other similar components. For this system to reach 100 million hours of up-time, each individual component would need to be rated at 100 million hours. Once you account for variation about the mean for MTBF, your redundant system will only improve the probability of actually attaining it. It doesn't extend it.

-Rick

 

[SvingenB]

This type of redundancy will reduce the probability of total system failure but it won't extend the longevity of the individual components. All of the components of the triple-redundant system are running in the same environment for the same duration as all the other components and are thus subject to the same failure rates as all other similar components. For this system to reach 100 million hours of up-time, each individual component would need to be rated at 100 million hours. Once you account for variation about the mean for MTBF, your redundant system will only improve the probability of actually attaining it. It doesn't extend it.

-Rick

I do not agree. Life expectancy is completely different than MTBF. Consider a Lycoming with TBO (life expectancy) of 2000 h. The MTBF may very well be in the 100,000 hours range or more, but ONLY within the TBO of 2000 hours. MTBF is statistical measure of reliability for a number of units, not the life expectancy of components.


Edit:
Just did a search on the net, and there seems to be much confusion about this . MTBF (the way I use it) means Mean Time Between Failure. Some call it Mean Time Before Failure, but that is what I would call MTTF or Mean Time To Failure. Then again, some say that MTTF is MTBF for components that cannot be repaired (which of cource is correct, but redundant in my opinion). MTTR is Mean Time To Repair, so effectively there is MTBF = MTTF + MTTR. For production units, for instance a power plant, this is used to calculate the availability, A = MTTF/MTBF.

Some also equate MTBF with life expectancy. This is definetely NOT correct. If MTBF was equal to the life expectancy, for instance 2000 h, (it would be more correct to use MTTF ), then only half of the units would actually meet the life expectancy, meaning you cannot expect that an arbitrary unit will live that long with any certainty.

I also read that the typical MTTF (where failure = death) for a human being in it's 30's is typically about 900 years

 

[rickmellor]

Well, I guess I'm a glass-half-full kind of guy.

Here's a nice study study that takes a deep dive into MTTF for hard drives. While there are some key differences between electromechanical drives and today's solid state EFIS systems there are still some key parallels to be drawn. In particular they show that exponentially distributed failure rates in redundant systems isn't really seen in the real world, largely due to consistent environmental stresses imposed on co-located equipment. There's also a cool bit in there about how drive failures follow a decreasing hazard rate function ... i.e., the more recently a failure occurred, the more likely the next is to occur. This also makes sense from the consistent environmental stress perspective.

http://db.usenix.org/events/fast07/tech/schroeder/schroeder_html/index.html

From the perspective of redundant EFIS / avionics installations, this type of information leads me to believe that the best redundant installations physically isolate the system components as well as logically / electrically isolating them. For instance, if you're deploying dual AHRS units in a plane, stick those suckers as far apart as possible even if it means that the backup will be in a less than ideal location. The physical isolation and the differences in temperature, vibration, humidity, etc. may just play a key role in ensuring that the overall system remains viable in the event of a failure.

-Rick