Interesting EFIS Observation

[Jamie]

Originally Posted by penguin

I think this phenomenon shows potentially a more significant problem as the attitude platform appears not to be very stable under any kind of failure condition. Clearly Dynon provides great equipment for VFR operations, but for regular IFR? To achieve their price point perhaps Dynon have had to make savings somewhere, perhaps this is one area? In my view the attitude solution should not require pitot or GPS aiding to remain stable.

Even the mighty Garmin G-1000/G-900X uses GPS aiding and pitot aiding as a backup.

[Brantel]

Your opinion stated here shows that you are clearly uninformed on how these things work...there are many post in the archives that will educate you on why the affordable Gyro's used in these experimental EFIS systems need aiding of some sort or how aiding is used to augment the system for better performance.

Like Jamie said above even the very expensive Garmin gear uses aiding...is it not suited for IFR????

The algorithms used in these AHRS's are very complex...You cannot judge a AHRS's performance from testing it on the ground while manipulating one or two of it's inputs without following the rules of physics for all its other inputs. The laws of physics for an airplane moving through space are much different than those of a box sitting on the ground with one or two of its inputs being manipulated unnaturally....

For Dynon (and others) pitot or GPS data is used in the attitude solution. If you wildly vary this input without the other natural reactions that should be there in real life, you will see an error.

Another thought:

Aiding is also used to decrease the time it takes for an AHRS to recover from being restarted of from being saturated and tumbling....

Which is better in IMC, an EFIS that cannot be restarted in flight or one that can?
Which is better in IMC, an EFIS that takes a very long time to recover from being restarted/being saturated/tumbling or one that only takes a few seconds?

Aiding can be a good thing to have in IMC not bad...

Originally Posted by penguin

Paul, excellent write up.

I think this phenomenon shows potentially a more significant problem as the attitude platform appears not to be very stable under any kind of failure condition. Clearly Dynon provides great equipment for VFR operations, but for regular IFR? To achieve their price point perhaps Dynon have had to make savings somewhere, perhaps this is one area? In my view the attitude solution should not require pitot or GPS aiding to remain stable.

Pete

[breister]

As just another note, I continually monitor both Ground speed and IAS during flight. The type of issue Jamie reported should never be a surprise, and if it starts to happen in-flight it should immediately be "Pitot Heat On" and "if no quick restoration, land immediately." I know approximately what my IAS should be at a given altitude and power setting, and so should every pilot who flies IFR.

The worst that is likely in this situation is to have a system not responding quite perfectly but trends accurately. If you aren't flying aerobatics in the soup, this should not be an issue, IMHO.

[GGreen]

I have no knowledge of what algorithms are ACTUALLY used in the experimental EFIS's. I would think they are using the Kalman filter or one of the plethora of newer algorithms though. These algorithms can detect when an input source is giving bad data, and then the software can cut that source out of the loop.

Another observation is that recovering from tumbling means the unit has to determine it's current orientation. This is greatly eased by another source of data, such as magnetic heading, GPS data, air speed data, or altitude data. Each of these can provide a good indication of one element of the solution which helps the algorithm to orient itself. GPS would be fastest as it provides more than one of the solution element at once, ie velocity, heading, altitude all together.

I'm assuming that when it is turned on, and there is no motion detected from any of these inputs, then logically it can be assumed that the airplane is mostly level, no significant pitch or roll. But then how does it work with a Cozy, which can have the nose gear up and a significant down pitch on the aircraft when systems are powered?

[Brantel]

Cuz they can sense gravity....

Never seen a modern EFIS that uses whatever position it is turn on in as its straight and level position...

Originally Posted by GGreen

I'm assuming that when it is turned on, and there is no motion detected from any of these inputs, then logically it can be assumed that the airplane is mostly level, no significant pitch or roll. But then how does it work with a Cozy, which can have the nose gear up and a significant down pitch on the aircraft when systems are powered?

[penguin]

Originally Posted by Brantel

... even the very expensive Garmin gear uses aiding...is it not suited for IFR????

The algorithms used in these AHRS's are very complex...You cannot judge a AHRS's performance from testing it on the ground while manipulating one or two of it's inputs without following the rules of physics for all its other inputs. The laws of physics for an airplane moving through space are much different than those of a box sitting on the ground with one or two of its inputs being manipulated unnaturally....

For Dynon (and others) pitot or GPS data is used in the attitude solution. If you wildly vary this input without the other natural reactions that should be there in real life, you will see an error.

Another thought:

Aiding is also used to decrease the time it takes for an AHRS to recover from being restarted of from being saturated and tumbling....

Which is better in IMC, an EFIS that cannot be restarted in flight or one that can?
Which is better in IMC, an EFIS that takes a very long time to recover from being restarted/being saturated/tumbling or one that only takes a few seconds?

Aiding can be a good thing to have in IMC not bad...

Which EFIS is better, one which only displays data when it knows it to be valid, or one that displays anything that is available?

I am aware of aiding and Kalman filters and have worked with systems that use these devices for many years. The whole point of a well designed Kalman filter is that it really doesn't care in the short term if one of the inputs is lost - once the filter is up and running (which can take a while) it can cope with losing inputs, for example position without GPS should be reasonably accurate for at least 15 minutes.

A system that shows an immediate deviation in output when one input is excited doesn't sound to me like a well written filter. Of course it is often difficult to tell exactly what is going on by just observing outputs, and none of these companies will divulge what is going on behind the screen.

If you want to trust your life to one box that must re-boot quickly for you to maintain control then that is your choice. I prefer to install an independent attitude indication (and ASI & Alt) that I can use if my EFIS should fall over (and I can also use to cross compare the EFIS output), and to let my EFIS come back on line when it has carried out proper airborne alignment.

Dynon provides a great system for VFR and occasional (life-saving?) IFR. Please understand the limitations of all the systems installed in your airplane and operate accordingly.

Pete

[Brantel]

Keep in mind these comments I have been making have nothing to do with any particular manufacturer. They are all different and all have their limitations. Even the very high dollar units.

One brand will turn monochrome if it detects that the attitude solution is out of whack, others put up a big red X, others do something similar. The problem is how does the EFIS know when its attitude solution is bad? This is not an easy thing to determine 100% of the time. Even the units with dual AHRS and realtime crosschecking have a problem with deciding which one is correct. Too sensitive and it will be red x'd more than it needs to be, too insensitive and it may not let you know it has a problem when it should. There are always tradeoffs to be dealt with.

Also, I would never trust my life to any single AHRS while in IMC no matter what brand it is, that just ain't very smart.

I do however kindly disagree that you can take any one of the current mainstream experimental EFIS maker's platforms (Garmin, GRT, AFS, Dynon, MGL) and label it as no good for IMC. Each of them have their own pros/cons but any of them can be used safely in an IMC environment where a proper panel has been designed with the appropriate backups. There are many ways to come up with what appropriate backups. How many backups you need depends on how much risk you are willing to accept.

So one or more manufacturers use pitot info as primary for their aiding with GPS backup. While others use one or a combination of pitot, GPS, magnometer in some way in their solutions for aiding.. manipulate one or more of these inputs randomly and see what it does to the attitude indication. They all can be tricked into wrong indications if you give them inputs that do not follow the laws of physics that the algorithms are based on....

Originally Posted by penguin

Which EFIS is better, one which only displays data when it knows it to be valid, or one that displays anything that is available?

I am aware of aiding and Kalman filters and have worked with systems that use these devices for many years. The whole point of a well designed Kalman filter is that it really doesn't care in the short term if one of the inputs is lost - once the filter is up and running (which can take a while) it can cope with losing inputs, for example position without GPS should be reasonably accurate for at least 15 minutes.

A system that shows an immediate deviation in output when one input is excited doesn't sound to me like a well written filter. Of course it is often difficult to tell exactly what is going on by just observing outputs, and none of these companies will divulge what is going on behind the screen.

If you want to trust your life to one box that must re-boot quickly for you to maintain control then that is your choice. I prefer to install an independent attitude indication (and ASI & Alt) that I can use if my EFIS should fall over (and I can also use to cross compare the EFIS output), and to let my EFIS come back on line when it has carried out proper airborne alignment.

Dynon provides a great system for VFR and occasional (life-saving?) IFR. Please understand the limitations of all the systems installed in your airplane and operate accordingly.

Pete

[breister]

Originally Posted by penguin

Dynon provides a great system for VFR and occasional (life-saving?) IFR. Please understand the limitations of all the systems installed in your airplane and operate accordingly.

Pete

Is there some particular reason you are singling out Dynon for your criticism?

Your opinion is your business, but realize that the arguments you give are not consistent with the relative merits / limitations of our other available choices. Vacuum attitude indicators are considered adequate for IFR and have NO corrective input to actual horizon. If a vacuum instrument tumbles IFR, you are likely toast. Thus, ANY of the current generation of experimental EFIS units, when functioning, are likely superior to that solution.

As to reliability, I've had several vacuum pump and instrument failures over my flying lifetime - I'll take an electric backup vs. a second vacuum instrument any day.

[Ironflight]

It was just an observation I was sharing guys....didn't mean it to be a grenade tossed in the room!

[penguin]

Originally Posted by breister

Is there some particular reason you are singling out Dynon for your criticism?

Because this thread is about Dynon. Clearly this is a very emotive subject (I have to say I'm a little surprised how emotive), also I have not supported vacuum systems - I too regard them as rather unreliable. But I do disagree that a vacuum failure = death. If your instrumentation system consists of vacuum gyros and an electic T&B then it strikes me that you should be competent to fly to clear air on the T&B - otherwise you're betting your life on a vacuum pump which doesn't seem very smart.

Brantel - I agree with most of what you wrote, but we will have to agree to disagree about the suitability of all of the manufacturers you list - I would only fly IFR with 3 of them.

My basic point is that we should understand the risks inherent with the equipment we choose to fit to our airplanes and accept those risks knowingly, rather than in blind ignorance.

Pete

Pete