[Dynon]

Dynon is releasing v5.1 of software today for all current EFIS units (D10A, D100, D180), and version 1.0 for the new D6 and D60. These versions of software add a feature we are very excited about: GPS Assist.

As most Dynon customers know, the Attitude system in Dynon EFIS units use airspeed to help it calculate which way is up?. Dynon has always believed that the use of airspeed gives us superior attitude performance, due to the fact that we measure airspeed directly, instead of estimating it. Airspeed is a crucial number inside any AHRS unit, but there are multiple ways to measure or estimate it. No estimate is ever perfect, so direct measurement can really help performance, especially in aircraft that have high horsepower to weight ratios. One advantage of direct measurement is reflected in the fact that we can be powered up while in flight and only need a few seconds to get aligned, whereas other systems can need minutes on the ground, motionless, in order to perform correctly.

The use of airspeed does, of course, carry one large caveat, which is that attitude performance degrades if a pitot becomes blocked by ice or another obstruction. While this is a rare situation in a properly equipped IFR aircraft, it is clearly not one that you ever want to encounter.

Dynon has been working hard to change this situation. We have considered a lot of different solutions, and we continue to believe that direct measurement of airspeed gives us superior AHRS performance. Given this, we decided the solution was to supplement airspeed with GPS data when it was warranted. GPS data is particularly nice in that it is commonly available, and it is easy to know when the data is valid and when it is not. We feel we?re in good company with this method. Garmin uses GPS as a primary source in the G1000 AHRS, and if GPS fails they switch to pitot and magnetometer. They require ?external? aiding in the same way that we do.

This GPS assist system has been thoroughly tested, and has been flown heads down for hours on end with only GPS as the reference. Based on our testing, we believe all IFR maneuvers (and even more aggressive ones than that) can be safely and accurately flown when using GPS assist. We think this is a great addition to the Dynon EFIS line and are excited to have this available for all current EFIS systems.

To use GPS assist, you must have GPS hooked up to your EFIS and properly configured. Many of you already have this connection made in order to use the HSI, autopilot, or other functions. If you do, there?s nothing you need to do besides updating to v5.1 in order to use GPS assist. If you do not have GPS hooked up yet, instructions on how to do so can be found in the EFIS install manual. The latest manuals can always be found at http://docs.dynonavionics.com


The addition of GPS assist gives the EFIS some new behaviors:

If GPS data is available, it will be used when the system considers it to be better than airspeed. This will be indicated on the screen with a magenta "GPS ASSIST" and the GPS ground speed will pop up below the IAS number, labeled "Gspd". This is also a great warning that your IAS number is probably inaccurate.

If GPS data fails while it was in use (ie when pitot had already failed and then GPS fails), "CROSS CHECK ATTITUDE" will appear on the screen. You can continue to use the attitude as a reference, but please be vigilant in cross checking the attitude against heading, turn rate, VSI, etc.

If no GPS is hooked to the system, no changes in behavior occur.

Many old annunciations that were "black and white" attitude screens now are annunciated with text below the aircraft icon.

If you wish to manually test GPS assist, the easiest way is to give your EFIS zero airspeed by venting the pitot line to the atmosphere in the cabin. Be aware that if the EFIS is your only IAS gauge in the plane, you will now have NO IAS gauge. Please fly safely. If you have backup steam gauges or an IAS gauge, please make sure you block the disconnected pitot line so they continue to work.

 

[breister]

Fantastic news, Dynon!

In over 2,500 hrs of flying including in pointy jets I've never had a pitot system fail - but there is always a first time. This will provide good peace of mind.

Right up until the GPS satellites start failing in 2010...

 

[Geico266]

Right up until the GPS satellites start failing in 2010...

Dudes at Dynon...... Any comments on the GPS system as mentioned? Do you believe these "weaknesses" will result in GPS problems for Dynon systems? Aircraft in general?

 

[Dynon]

We do not need a very accurate fix from the GPS. Even the most basic of fixes will do what we need. Given that the only thing I have really heard about GPS is a possible reduction in the constellation and the attached reduction in accuracy, this should have no effect on us.

 

[breister]

And apologies due on my part for spreading rumors. There is indeed a possibility announced by NASA / Air Force that there may be some degradation of GPS signals if they cannot get their launch schedule sorted out, but NOT a total loss of service. I made that comment tongue-in-cheek; sorry, sometimes I forget that that does not come across well in writing.

Since the Dynon is only looking for average ground speed to aid their attitude, there is little to zero likelihood that the GPS degradation will be drastic enough to prevent accurate attitude reference in the worst-case scenario.

 

[Delta Bravo]

DYNON 5.1 Issues

I have installed, removed and reinstalled the new firmware version 5.1 on my D100 three times. In each case the instrument is unusable as it displays wild pitch and roll fluctuations that are triggered by normal acceleration both on the ground and for troubleshooting purposes, in flight.

I have uploaded the back up version 4.0 and the D100 has returned to normal operation.

Anyone else have problems with 5.1?

Thanks

Don Bodnar

 

[Dynon]

Don,
If you could send us a backup of your unit after upgrading to 5.1, as well as your 4.0 backup, we should be able to figure it out. It sounds like it's corrupting your calibration table for some reason.

We've heard of one other case of this. With 8K+ units out there, it's not a confirmed common problem, but it's clearly an issue we're looking into.

 

[roee]

AHRS aiding: airspeed vs. velocity

Dynon folks, it sounds like you're taking some great steps in a good direction. It's great seeing your product evolve! I'm still a bit confused though about your explanation and reasoning for using airspeed in the first place, and I'd like to understand it.

...Airspeed is a crucial number inside any AHRS unit...

Is that really true? I think that an AHRS could be designed without any notion of airspeed or other air data or aerodynamics at all (could operate just as well in a vacuum), and that at least some AHRS out there are in fact designed that way.

Please correct me if I'm wrong, but I believe that the function of tracking attitude using imperfect inertial sensors (accelerometers and/or rate gyros) and magnetic sensors is actually more directly related to, and better aided by, the aircraft's known 3D velocity vector through space than its airspeed. Attitude is only related to airspeed and to any other air data indirectly, by making assumptions about the aerodynamic behavior of the aircraft and the atmosphere. In other words, airspeed, even if known to perfect accuracy, is actually still just a poor substitute for the aiding input that an AHRS really wants, which is a true 3D velocity vector. And incidentally, a very accurate 3D velocity vector can be readily attained from GPS. So it seems that even under nominal conditions where no faults exist, GPS would still provide a more useful aiding input than air data to the attitude solution.

I'm not trying to be a heckler here. I truly find this subject fascinating from a technical standpoint, and I'd like to better understand it. If my statements above are incorrect, please do correct me, and please explain my error in detail.

Respectfully,
-Roee

 

[Aviator168]

Rose,

Airdata is important if there is no GPS. However, GPS updates are slow. Folks at Dynon, I think, is trying to find the correlation between the 3D velocity vector and the airdata with the aid of the GPS.

 

[Dynon]

Well, we're not going to give away all our secrets, but the need for velocity inside an AHRS is well known. Here's the most simple reason you need airspeed: A standard rate turn for say, 1000 degrees. The roll gyro is worthless over these 5 minutes, since it's imperfect. Accelerometers are worthless because you're an awesome pilot and are well coordinated, so they just show all acceleration as "down." So now what do you do? Well, airspeed and other data helps here.

To comment on GPS, GPS doesn't work in accelerated flight. You can have the nose way above the horizon while you descend in a stall, and in a 6G turn, the vector of the plane is not the direction it's pointed. Acceleration in flight is related to TAS (which is your speed through the air mass), not to your vector in 3D related to the ground. I can pull G's in an airplane even if my ground speed is zero if the wind is not, so GPS isn't the perfect answer. You can't use GPS to independently correct roll over time because it gives you no info about the roll angle of the aircraft (GPS reads the same upside down as it does right side up).

I think the reverse question is actually the more interesting one. Ask an AHRS manufacturer that claims to not need airspeed or GPS at all how they do it, what situations cause them to misbehave, and how they recover in flight if the estimate goes wrong (like after a spin). If they are not "using" airspeed, then they are estimating it internally, and there are things that can throw the estimate off. Some systems say "we use airspeed when we have it, but we don't need it." Clearly in this case performance is reduced somewhere, or airspeed would never be used, so it's worth asking about that more specifically as well. A pitot frozen at zero is still a perfectly valid airspeed, so is this "wrong airspeed" the same as "no airspeed"?.

One of the things we pride ourselves on here at Dynon is always being 100% transparent about the sources of data that our AHRS uses and what can cause the pilot issues.

 

[roee]

I think we're coming at the AHRS design problem from slightly different scopes, and that's where the difference is creeping in. Sounds like you've taken a very airplane-centric approach to your AHRS design, and I think I understand now how you're using speed, whether true airspeed or GPS velocity, to aid your AHRS. So in that context I understand your comments. The use of 3D velocity that I was alluding to is somewhat different from what you're doing, and is a more general approach applicable to any vehicle with full 3D freedom of motion in translation and rotation. No assumptions about the aircraft traveling generally in a forward direction or having aerodynamic characteristics of any kind.

About GPS: GPS works just fine in accelerated flight, it doesn't care. What you get from it is a time series of 3D position fixes, regardless of acceleration, and as you pointed out, regardless of aircraft attitude/heading. Now the first order differential of that gives you a time series of 3D velocity vectors. And taking another differential gives you a time series of 3D acceleration vectors (or acceleration-induced "G" load vectors). You don't care about absolute position, or even about absolute velocity (frame of reference is arbitrary, neither the ground nor the airmass matter here). What you do care about is the differential of velocity, which is acceleration. Add to that the fixed force of gravity, i.e. 1 G straight down, and you have the aircraft's felt "G vector". Of course this G vector is relative to GPS's fixed frame of reference, i.e. the earth, not relative to the axes of the aircraft. So by itself that doesn't tell you anything about the aircraft's attitude/heading. But a corresponding G vector relative to the axes of the aircraft is exactly what you do get from the AHRS's accelerometers. Take both of those vectors, throw in a vector from the magnetometer, crunch a bit of math, and you've got an attitude/heading solution.

Of course it's harder than that in reality, being that all these sensors are imperfect. But they actually complement each other's imperfections really well, which is what makes it all possible. Inertial sensors have high update rates (sub-millisecond) and good short term precision, but have long term drift, and by nature of being rate sensors, lead to error accumulation. GPS has a relatively slow update rate, and moderate short term accuracy, but works on absolute position and has no long term drift. That's where the artistry of filter design comes in, weighing in the inertial sensors for quick responsiveness and weighing in the GPS data to correct long term drift. (And I've mostly neglected the magnetometer and rate gyros in the discussion to try to keep it simple).

Ok, too much typing... I'd love to discuss this subject further, preferably with a beer in hand! (Oshkosh?) Anyway, I just want to say again, that my post was not a criticism so much as an intellectual inquiry. Thanks for taking the time to respond.

And I do applaud Dynon for being upfront and discussing this stuff openly. Every instrument has its set of vulnerabilities, and it's essential for the aircraft builder and pilot to know and understand these vulnerabilities to use the instrument safely.

 

[Aviator168]

It could be a lot simpler. Mount two gps receivers on the aircraft. One in the front and the other on the tail. You will get your attitude reading. A study has been done with this setup and an accurcy of less than 0.5 degrees (forgot the link. You have google it). Last time I checked. A GPS receiver is less than $60US. Mount an extra one, you will get the roll reading as well.

 

[roee]

Multi-GPS attitude

It could be a lot simpler. Mount two gps receivers on the aircraft. One in the front and the other on the tail. You will get your attitude reading. A study has been done with this setup and an accurcy of less than 0.5 degrees (forgot the link. You have google it). Last time I checked. A GPS receiver is less than $60US. Mount an extra one, you will get the roll reading as well.

Yes, many studies have been done on using multiple GPS antennas at a fixed geometry to determine attitude. But it's a bit more complicated than just throwing in three off-the-shelf $60 receivers. To get good attitude accuracy, the receivers have to be using the same clock, same choice of satellites, etc. so that common mode error sources can be canceled out and only the relative signal phase differences to the antennas will factor in. It's a simple (and good!) idea, but there's a lot more that actually goes into making it work well. This is another fascinating topic, good for at least a beer or two!

But even with this idea working at its best, on its own, it still has some distinct weaknesses:

1. It is totally dependent on GPS. Loss of signal means immediate loss of attitude solution.

2. It has good static performance, but not so great dynamic performance.

Now the good news. Here again, the strengths and weaknesses of multi-GPS attitude dovetail perfectly with the strengths and weaknesses of inertial-based AHRS. So if/when multi-GPS attitude technology is deployed (and I think it will be), it will very likely be coupled with inertial sensors, integrated in similar fashion to the AHRS aiding that we've been discussing here. Whether we then call it an "inertial AHRS with multi-GPS attitude aiding" or a "multi-GPS attitude system with inertial aiding" is mostly a matter of semantics.

Sorry for the thread tangent, but this is great stuff!