What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Look Ma, No Airspeed!

Vac

Well Known Member
Benefactor
26-minute familiarization video of a flight conducted without any airspeed references, only performance angle of attack cues:

https://youtu.be/r8XcXDHLIIM

No magic here. Just a good AOA/Energy Management system. Disclaimer: we generated the inset indexer video post flight from data, it's a bit jittery, particularly the slip/skid ball. There are several mis-speaks in the cockpit narration. No excuse, we'll do better in the future.

Reader’s digest version of how an AOA system should work for folks not familiar with our work:

System Requirements. For an AOA system to serve as a primary reference it must be accurate across the entire speed band of the airplane from Vmax to stall. It should measure actual AOA within 1/4 degree or better under static conditions and 1 degree or better under dynamic conditions. It must be responsive to high G pilot inputs and gust loads (better than 2G/second onset rates), but sufficiently damped and presented to the pilot as a “flyable” cue regardless of the wing it is fitted to. The system should accommodate flap position with proper sensors and a calibration curve for each flap setting, especially if the airplane has slotted flaps. Calibration curves should be normalized using data from within the same pressure field occupied by the AOA probe. For typical underwing locations, normalization should be based on data directly from the probe. Calibration points should extend from Vs to Vmax and should include EAS points for stall warning (based on FAR 23 criteria), an ONSPEED condition (Vref), L/Dmax and Carson’s speed. Use of multiple data points across the entire speed band of the airplane and regression analysis yield more accurate calibration curves.

Performance Cues. The parameters that an AOA display can convey to the pilot (visually and/or aurally) are Carson’s speed, L/Dmax, ONSPEED and stall warning. If a % lift display is also provided, weight adjusted maneuvering speed can also be determined. Maneuvering speed occurs at 100/aircraft G limit. For the 6G RV-4, this is a 17% lift condition. If the wing is generating more than 17% lift, the aircraft is below actual Va. L/Dmax occurs at 50% lift and ONSPEED occurs at 60% lift.

Energy Cues. From a general energy management perspective, a “flyable” system should also provide an easy means to determine Excess Specific Power, abbreviated Ps and pronounced “P sub S.” A simple, intuitive cue should tell the pilot if there is more thrust than drag or more drag than thrust for a given weight condition (actual gross weight times any load factor). If P sub S is negative, the airplane will either go down, slow down, or both, unless the pilot reduces AOA and/or increases power; assuming power is available and the ground doesn’t get in the way. If P sub S is neutral, thrust and drag are balanced, and the airplane achieves maximum sustained turn rate and minimum sustained turn radius—sometimes referred to as an “optimum turn.” An easy way to think of this is simply "bleeding, not bleeding" energy. Any slow tone means you bleeding energy--stop the bleeding by reducing AOA. If you are completely engine out, there is no "thrust" part of the equation, but the same "bleeding/not bleeding" push/pull logic applies if your objective is the best blend of glide and turn performance and optimum energy for approach/touchdown when maneuvering at low altitude. Accurate aural AOA cues convey essential performance and energy information to the pilot without the requirement to look inside the cockpit.

FlyONSPEED.org is a 501(c)3 non-profit, open source, all volunteer group dedicated to developing hardware, software and education resources for the EAB community. Our mission is to reduce loss-of-control mishaps, share lessons learned and have fun. We’d like to see a paradigm shift to folks “flying the wing,” knowing their energy state and maintaining positive aircraft control using proven military technology and techniques; and a data-driven approach to improving training.

Fly safe,

Vac
 
Last edited:
Definitely a lot of work to get it this far, and I'm looking forward to seeing how it performs in other than calm air.

Here's AOA recorded in the RV-9A in conditions not all that gusty:
https://www.youtube.com/watch?v=bWSzptdQyFk

Recall that the RV-9A is very responsive to gusts. One Van's engineer said that in turbulence, it "rides like a buckboard."
 
Definitely a lot of work to get it this far, and I'm looking forward to seeing how it performs in other than calm air.

Here's AOA recorded in the RV-9A in conditions not all that gusty:
https://www.youtube.com/watch?v=bWSzptdQyFk

Recall that the RV-9A is very responsive to gusts. One Van's engineer said that in turbulence, it "rides like a buckboard."

This is the key:

"System Requirements. For an AOA system to serve as a primary reference it must be accurate across the entire speed band of the airplane from Vmax to stall. It should measure actual AOA within 1/4 degree or better under static conditions and 1 degree or better under dynamic conditions. It must be responsive to high G pilot inputs and gust loads (better than 2G/second onset rates), but sufficiently damped and presented to the pilot as a “flyable” cue regardless of the wing it is fitted to. The system should accommodate flap position with proper sensors and a calibration curve for each flap setting, especially if the airplane has slotted flaps. Calibration curves should be normalized using data from withing the same pressure field occupied by the AOA probe. For typical underwing locations, normalization should be based on data directly from the probe. Calibration points should extend from Vs to Vmax and should include EAS points for stall warning (based on FAR 23 criteria), an ONSPEED condition (Vref), L/Dmax and Carson’s speed. Use of multiple data points across the entire speed band of the airplane and regression analysis yield more accurate calibration curves."
 
Definitely a lot of work to get it this far, and I'm looking forward to seeing how it performs in other than calm air.

Here's AOA recorded in the RV-9A in conditions not all that gusty:
https://www.youtube.com/watch?v=bWSzptdQyFk

Recall that the RV-9A is very responsive to gusts. One Van's engineer said that in turbulence, it "rides like a buckboard."

There's a "Kite" mode in OnSpeed. :D

Not really, but it has smoothing filters that should be set higher for a light wing loading airplane like a -9.
Eventually we'll be implementing adaptive smoothing that changes according to conditions to make it easier to fly the AOA in turbulence.

Attached a screenshot of part of the OnSpeed settings page where you can set the smoothing values.

In your video the plane is being thrown around pretty good. Pitch, Roll and Yaw are all dancing around quite a bit. The AOA is not wrong. It would be wrong if it was steady.
 

Attachments

  • Onspeed_smoothing.png
    Onspeed_smoothing.png
    69.8 KB · Views: 210
It must be responsive to high G pilot inputs and gust loads (onset rates more than 2G/second), yet suitably damped and given to the pilot as a "flyable" cue, regardless of the wing it is connected to.

-------------------------------
Minecraft education is beneficial because it fosters creativity, problem solving, self-direction, collaboration, and other life skills.
 
AOA probe

Forgive me if this has been covered elsewhere.

Reading the old test flying article in Sport Aviation about using AOA instead of airspeed to determine best angle, best rate and glide, I had the "bright" idea of taking the AOA % info from the Dynon D-100 EFIS serial output and displaying it on my Arduino display. Great disappointment. Looks like I only get non-zero numbers when close to stall speeds.

My AOA "probe" is a pull-rivet in forward-most rib hole, making it about 18-20 degrees to the pitot tube.

Calibrating it for stall warning was easy enough and works well for that purpose.

But is the lack of useful AOA info due to D-100 (or D-10A) info limits or due to my AOA "probe" location?

What kind of probe or sensor does the Onspeed system use?

Finn
 
Hi Finn,

That's actually a cool idea, and we went thru a similar learning curve when we started the project using a Dynon % lift signal to drive an Arduino. There are a couple of hurdles with that approach.

First is your sensor configuration. I have a Dynon probe on my RV-4 located at about 26% chord, 6" below the left wing on a Gretz mount. Where the probe is mounted and using local pressure for normalizing is important for accuracy. So, having your pressures come from two different locations is less than ideal, but we do have a similar system flying on an RV-12 using the Dynon % lift signal; so it may be doable.

The next issue is that the Dynon system is engineered to be a progressive stall warning system, and if it's properly calibrated it does that well. Dynon uses a simple two point calibration. It also derives percent lift. You can't accurately capture a pressure vs alpha curve with only two points. In theory, % lift at stall is 100, but if you download your data and plot out your Dynon calibration, that's probably not the case.

As part of our original experiment, we had some folks in the field calibrate their Dynon systems (DY and SkyView were calibrated in a similar manner and provided the same output). Turns out, it's very difficult to replicate a calibration and achieving 100% lift at stall just didn't happen most of the time. This is why Dynon AOA indications don't translate from one airplane to another, similar to indicated airspeeds--it's entirely calibration dependent.

This was a typical result. This airplane has a SkyView system with a Dynon probe located in the Van's plan position:

3c039a_a4204bac7b3a4a548a3344848adf0316~mv2.jpg


There are also non-linearity issues with using % lift. All of this is why we ended up abandoning trying to process the Dynon signal and measure angle directly from the coefficient of pressure.

If you haven't taken a look at our website, we've got lots of lessons learned there. We also have Arduino code on our GitHub site designed to read the Dynon signal. Drop me a PM and I'll give you contact info--happy to chat any time. Keep your experiment going :D

Cheers,

Vac
 
Last edited:
Total Energy

Vac,
Thinking about the calibration problem and getting a reliable low noise signal for Beta dot I wondered if any thought had been given to using a probe from the top of the vertical stabilizer to get a measure of total energy and airflow angle measurement. My suspicion is that prop wash would be an issue - just wondered if you have any comments on the idea or if you did any measurements.

Still waiting on Mouser for unobtainium for the teeny and IMU chips.

Keith
 
Onspeed probe

Finn,

The Dynon AOA probe is designed to provide a linear coefficient of pressure to AOA curve. Take a closer look at the shape of the probe and you'll understand. It's a very clever design. I haven't seen any data but that rivet on the leading edge will never provide a linear curve. Yeah you can make it work as a stall warning, but that's about it. I would test out that stall warning at a different density altitude and at a different Gload like in an accelerated stall.

Onspeed uses any of the Dynon, Garmin or Alpha Systems probes. Depending on where the probe is located on the wing the resulting CP to AOA curve isn't always linear so we're using up to 3rd degree polynomial curves to make it really accurate.

Another issue is calibration. During Vac's early Dynon % lift to Arduino flight (Onspeed Gen1 project) tests with different airplanes he found out that only one pilot was able to fly a proper calibration curve. A lot of airplanes stalled at 80% lift. Our new calibration procedure now involves flying a a controlled deceleration to stall at 1kt/sec. No one has a gauge for that so we built one.
It's very important that during calibration you enter the stall without spiking your AOA. Most pilots tend to be impatient (or maybe nervous :) ) and pull into it. Then the resulting spike will end up being recorded as 100% lift, but the airplane will stall at 80%.

Also, %lift is not the same as AOA.

Check out the attached plot of an RV-10 (flaps up) lift curve.

The calculated lift (based on the lift curve slope) starts deviating form actual coefficient of lift at high angles of attack. So if you draw a new line between stall AOA and some other point on the lift curve, you'll never get back the original curve. It will always be just an approximation.

Lenny


Forgive me if this has been covered elsewhere.

Reading the old test flying article in Sport Aviation about using AOA instead of airspeed to determine best angle, best rate and glide, I had the "bright" idea of taking the AOA % info from the Dynon D-100 EFIS serial output and displaying it on my Arduino display. Great disappointment. Looks like I only get non-zero numbers when close to stall speeds.

My AOA "probe" is a pull-rivet in forward-most rib hole, making it about 18-20 degrees to the pitot tube.

Calibrating it for stall warning was easy enough and works well for that purpose.

But is the lack of useful AOA info due to D-100 (or D-10A) info limits or due to my AOA "probe" location?

What kind of probe or sensor does the Onspeed system use?

Finn
 

Attachments

  • rv10liftcurve.png
    rv10liftcurve.png
    140 KB · Views: 98
Thanks Vac.

Your graph seems to show results similar to mine. AOA % only comes alive (non-zero) with air speeds below 85 mph or so.

So if I want to get serious about this I need to tap my own differential pressure sensor into the AOA and Pitot lines (and probably relocate the AOA port) and add a flap position sensor.

Don't know what "alpha curve is" but will definitely take a look at your website.

Finn
 
Thanks Lenny,

"coefficient of lift"?

CP? Cp? Pressure coefficient? Pressure difference between Pitot and AOA port?

I'm obviously in over my head here and need to study up on all this.

Yes, definitely need to test the Dynon stall warning under different conditions.

Thing is that I typically fly way above Vx and Vy speeds. Uncomfortable at the apparently high pitch angles. Trying to determine Vx, what I thought was 45 degrees apparently was only 20 degrees per the D-100 logs. Need to get more comfortable at those angles.

Finn
 
Last edited:
Using Dynon AOA indications

Max,

You may be able to use the Dynon system to give you an L/Dmax and on speed cues--depends on your calibration and whether or not you want to experiment a bit. I encourage you to do so, since any AOA (even if it's only progressive stall warning) cuing is better than none. You won't be able to get the McDonnell tone pattern or display that we're using in the video; but it will improve your SA, especially for approach and landing.

If you look at the plot I posted above, you'll see the 50 and 60% lines, that .5 and .6 times the maximum calibrated % lift. In this case 80%. Assuming an accurate, linear lift curve is captured (it's not which, is what Lenny pointed out above), on speed would occur at 48% Dynon-calculated % lift and L/Dmax would occur at 40%. That's how our first generation experiment worked--the start of the "fast" tone occurred at L/Dmax and in this example, when the Dynon transmitted a 40% lift condition. We didn't set the tones based on Dynon % lift, however. We set them by flying the various speed appropriate for weight and flap setting for L/Dmax and on speed establish a known angle of attack and then utilized that Dynon computed % lift.

You can approximate the same thing by calibrating the Dynon system and then flying known parameters for L/Dmax, on speed and stall warning. Whatever those visual/aural cues are are those conditions in that airplane. On speed is a high enough AOA, there may be some Dynon tone option that captures it--so that's what you're listening for during approach and landing:

3c039a_ee0ea2de15ee4f699ad7e64ba2ae5670~mv2.jpg


Right now, if you want a full-up ONSPEED system it's still build only. Unfortunately supply chain woes have spilled over into our world as well and some parts are currently made of unobtainum. We are currently working on a new version of hardware. We realize that most folks don't have the inclination to build their own electronics.

Fly safe,

Vac
 
Last edited:
Is the stock Dynon AOA able accurate enough to fly onspeed or do you have to build one?

Max

From Lenny's reply the Dynon pitot/AOA probe will work fine.
It's the Dynon EFIS that introduces limitations.
If you look at their flyonspeed.org website and from previous posts I think that they decided to use their own pressure sensor.

Still reading to find out how they determine actual AOA for calibrating the pressure difference AOA. Can't use a vane on the fuselage (as you see on jets) due to prop slipstream. Might put a vane (with position sensor) under or in front of the wing temporarily.

Edit: Dang! Vac beat me to it!

Finn
 
SA?

... but it will improve your SA, especially for approach and landing...

Vac, what is SA?

For newbies like me unexplained abbreviations really prevents understanding.

Like in a earlier post. CP. If it was written Cp I might have figured out it was pressure coefficient, which I assume is pitot tube pressure divided by AOA probe pressure. But you know what they say about assuming.

Finn
 
Finn,

You rock--no worries! Yeah, I'm a walking TLA generator (three letter abbreviation)...

SA = situational awareness, or the pilot's perception of what's happening.

FWIW, I have the same issues estimating pitch visually. At test weight, I'm at about 45 degrees nose up (feet on the horizon) for a full-power stall, and, as you point out, even Vx is more like a jet than the typical GA airplane. The good news is that you are generally clear of the mythical 50' obstacle before you even get to Vx pitch :D

Here's everything in one, painful eye-test chart. We are indeed computing AOA based on the coefficient of pressure:

3c039a_aa0ae5e166834c6385283a5bacece102~mv2.jpg


This chart is a bit dated, and we are indeed working on our own probe configuration. The Dynon and Garmin probes work great, but are subject to error at sideslip angles. Gets interesting above about 6 degrees or so; so we are seeing what's in the art of the doable using a two-port spherical probe. No idea if we'll be able to crack that nut or not.

All of the physics are explained here: http://www.tc.faa.gov/its/worldpac/techrpt/tc18-7.pdf

Our primary focus at the moment is automating the calibration process. We learned in our original experiment that folks in the field have a hard time calibrating systems, and since accuracy is so important we are working on having the computer do the work.

Thanks for diving into the web site and keeping me honest with the abbreviations and assumptions. Feel free to email or PM any time as well.

Cheers,

Vac
 
Last edited:
AOA probes

Lenny,

As much as I can appreciate the Dynon Pitot/AOA probe design, I hate having stuff hanging out in the breeze (and the cost).

I was thinking about the system (Lift Reserve Indicator?) that had ports on top and bottom of the wing. Any chance that would provide a linear pressure-differential to AOA relationship over the full speed range?

Finn
 
Wing ports

Finn,

A couple of comments to the above posts:

Here's the Vac decoder ring :D (SA is right on top there for example): https://www.eugeneleeslover.com/aviator_slang.html#S

You can use a vane, but remember that there's a downwash under the wing. There's one even in front of the wing when using a boom mounted vane. So you'll need to account for that and calibrate it if you want real angles. Also the vane is susceptible to some other physics effects during dynamic maneuvering. It gets affected by pitchrate and rollrate and G effects from the inertia of the vane itself. It's all calibratable but it's not as simple as slapping a vane under the wing and reading AOA with its potentiometer.

As for wing pressure ports, AFS was doing that originally with their AFS AOA Pro and Sport but they switched over to the pitot probe AOA.
In the wing port setup linearize the curve by using this formula (Ptop-Bottom)/q, where q is dynamic pressure which is Ppitot-Pstatic. So you'll still need to have a proper pitot static setup for it.

Our new probe is a work in progress. We're flying it on two airplanes for now. My Zlín Z-50 and Vac's RV-4. The goal is to cancel out beta effects. Slips and skids at high AOA get particularly dangerous (skids near critical AOA result in a violent spin entry, Vac posted some videos of that on his Youtube channel), and AOA needs to be accurate in those conditions. Right now none of them are. It's a pretty complex problem. In theory and in CFD the probe performs well, but under the wing things are vastly different.

Our goal is to have accurate AOA in all conditions, including maneuvering and under high G onset rates. (btw G onset rate is how fast the Gs come on in G/seconds, also called jerk or jolt)

The onset rate goal is met. The Zlín is a highly maneuverable airplane with its barn door control surfaces, and it's cool to see (and hear) that the AOA signal actually leads the airplane at high G onset rates. Takes some time for the airplane to catch up with the change in the relative wind.

Lenny
 
Thanks Lenny,

The whole idea with the vane would be to calibrate pressure-sourced AOA. So there goes that idea. How would you calibrate a vane-AOA anyway?

So the top/bottom port idea would require two differential pressure sensors. So you might as well have a streamlined pitot/AOA probe. 3D-printing one might make that be economical. Just keep shooting down my ideas :D Actually, I do appreciate it because it stops me from going down non-practical paths.

Yeah, would never have considered slips influencing AOA probe.

I'll just keep following you guys.

The main reason I looked at the Dynon "Angle of Attack
-- 00 to 99, percentage of stall angle" info was I already had that serial signal going to my Arduino Mega for combined data logging, so it was just a few lines of code to extract and display that.

Finn
 
Here's how you'd calibrate the vane for up/downwash. Under some strict conditions (wings level, unaccelerated level flight) AOA = Pitch. So you can fly 4-5 airspeeds, hold each one for a few seconds (we call that a trimshot) and note pitch angle from your Efis. Then you calculate a linear or polynomial regression from vane angle to pitch angle and you got your AOA curve.
The problem with that is that at high angles of attack you need a superpilot, like Vac, to hold a 0 VSI and a precise airspeed. Once you start climbing or descending the pitch formula induces huge errors in the AOA.

To get around that you can just do AOA=Pitch-FlightpathAngle, but now you need an accurate flightpath angle, which is asin(VSI/TAS), but TAS is temperature dependent so you need OAT. You see where this is going :)
You can just do the same thing for pressures and skip the vane. This is how we calibrate our system at the moment.

We have a much smarter and lot less pilot skill intensive "Calibration Wizard" we're test flying now that uses a guided deceleration from Vmax to Vstall. It detects the stall, calculates the stall warning, onspeed and other setpoints, regresses the curve and saves it all into the config.

You're on the right track with your setup, get a Dynon probe, calibrate it properly and then you'll have a fairly linear lift curve. You can even output Onspeed audio with the Arduino Mega. That's exactly what our Gen1 code does.
The other benefit you'll get is that the Dynon probe is biased toward higher AOA and you'll get accurate airspeed as you approach the stall. Look at the airspeed calibration table of a Cessna 172 (attached below). Huge errors at high AOA. The Dynon probe is designed to fix that. The Garmin one is too btw.

sD8g3.png



Lenny
 
Last edited:
...
You're on the right track with your setup, get a Dynon probe, calibrate it properly and then you'll have a fairly linear lift curve. ...

Nah. If I'm anyway going to replace my standard Van's bent 1/4" tube with a Pitot/AOA probe I'll rather wait for yours that is less sensitive to slips/skids.

What are the issues you're running into when going from the theoretical CFD modelling to actual install under the wing?

I'd certainly be interested in trying to 3D-print it if/when the .STL or .gcode file is published.

Finn
 
What are the issues you're running into when going from the theoretical CFD modelling to actual install under the wing?

Finn

This illustrates pretty well what happens:

41436-flow.jpg


Look at the flow angles on the bottom of the wing where the probe would be located and compare it to the parallel horizontal streamlines way in front of the wing (relative wind).

Our probe doesn't provide a linear output and because of that it won't work with the Dynon efis. It also doesn't measure dynamic pressure used to calculate airspeed. You'd need to have an Onspeed box to process the pressure data or have your own pressure sensors. Happy to send you the STL file for printing.

Lenny
 
AoA probe

Not sure I see the issue. Any probe will experience the non-straight streamlines under the wing.

Regarding your probe, I assumed that the two holes in the model pictured on your website forum was for pitot and AoA respectively.

I thought the problem with slip/skid/bank errors were due to the flat surface on the Dynon probe where the AoA opening is located and that simply rounding that would solve the problem.

In other words, the Dynon pitot opening is on a rounded surface, the AoA opening is on a flat surface of the probe.

Looking at your probe model again, if the two ports are side-by-side and connected together for average AoA pressure when slipping/skidding/banking, why not add a third opening for pitot?

Going through the Rogers paper referenced on your website, I have a couple of questions and comments.

Why 45 degrees and not 30 for the AoA port? Of the airplanes we're interested in, most stall well below 30 degrees.

For anyone not interested in making your own system, please disregard the following.

On our experimental airplanes we have no FAA problem with tapping into the pitot and static system. So the (Pfwd - P45) / (Ppitot - Pstatic) is feasible to get linear AoA results. (Pfwd / P45) / P45 requires two differential pressure sensors which may explain why your board has two differential pressure sensors and one one-port (gauge) pressure sensor for measuring static. The ones you use on your board are now $60 ea. (no longer $20).

Maybe to make things clearer:
Pfwd = Ppitot (pressure measured at pitot tube port)
P45 = Paoa (pressure measured at AoA port)
According to the Rodgers paper, two ways to get fairly linear results:
AoA = (Ppitot - Paoa) / (Ppitot - Pstatic) (requires two differential -- two-port -- pressure sensors)
or
AoA ~ (Ppitot / Paoa) / (Paoa) or (Ppitot / Paoa) - 1 (requires two one-port pressure sensors)

(Ppitot - Paoa) could be gotten with just one differential pressure sensor (pitot connected to one port on the sensor, AoA to the other port), so if you have access to dynamic pressure (Ppitot - Pstatic) from another source (like I do in my engine monitor), just one differential sensor would be required. Edit: However this may not work due to static port errors.

Finn
 
Last edited:
Wow Finn, you're all in!

Lemme' take a stab at your questions. Keep in mind I'm the stick monkey in this program, and I'm sure the engineers will chime in if I screw this up...

The aerodynamics under the wing are actually pretty complex. We are working now to develop a good CFD model to use for analysis. Remains to be seen whether we'll get there or not. As Dave's paper shows, placement of the probe is important. We engineered the printed prototype to utilize the existing Alpha Systems mount and made it the same length as that probe--it may simply be a matter of stand-off distance beneath the wing; but we don't know yet. Science in progress.

Any pitot tube has issues with sideslip. The original Alpha Systems probe analyzed by Dave and his team in the wind tunnel, showed error above 6 degrees of sideslip angle. I can actually hear sideslip error in the Dynon tube when I'm flying the AOA tone. You can replicate this with your airspeed indicator. Just watch the indication as you slip the airplane or induce a yaw input. So it's actually the Pfwd (pitot) pressure that is changing as a result of side slip, not the AOA pressure on the Dynon tube. And, depending on configuration, there may be a change in static pressure as well as sideslip angle changes.

More openings on a spherical probe make sense, but we've only got two differential pressure sensors in the box. We're trying to work with what we have, since, as you point out, those are pricey transducers (and the supply chain disruption isn't helping prices or availability--Econ 101).

The 45 degrees is just an expression based on Dave's original work. The Dynon smooth face is about a 30 degree angle. We just kept the original terminology. The difference in angle has nothing to do with the AOA range of the airfoil/airplane.

Lot's of learning curve in this project. We've been thru three different iterations of our basic coefficient of pressure equation. We are currently using P45/Pfwd (lower AOA port/pitot) and don't use aircraft static pressure for the AOA solution (we use that for IAS and altitude calculations). Dave's original formula was Pfwd/P45 for normalization, but the Alpha System pressures are positive throughout the full AOA range of the aircraft. Not so with the Dynon, so we had to overcome the dreaded "divide by zero" error that makes the computer tilt as the AOA port pressure (P45) transitions from negative to positive thru zero as AOA increases. We originally screwed that up by using non-local pressure with our first couple of formulas. Now we have circled back to basics as Pfwd never goes to zero or negative (unless you are in a tail slide, at which point any AOA tone isn't of much use ;)).

As long as you can derive the two local pressures from the sensor and can compute the coefficient of pressure, you should be good to go. You don't want to use static pressure, since that is a non-local source (e.g., static ports on the aft fuselage in the case of my RV-4).

One of our design goals is automatic calibration. Turns out that's a pretty complex problem using an inexpensive IMU. As a proud History Major, matrix calculus gives me a headache, hence "complex problem..." The engineers are all over it, however :). An accurate IAS and altitude solution assists with that, hence our sensor configuration. The configuration also provides the capability to have a CAS display in the cockpit, which I find handy for test work.

The system has a secondary function as a data recorder, so the indigenous airspeed and altitude information can be helpful there as well, especially for airplanes not equipped with an EFIS. Also drives the optional visual display. The "energy" display is designed to show airspeed, alpha and G--kind of a handy "if I only had one gauge in the rear cockpit what would I want" display if there was a student in the front pit.

The nomenclature has evolved over time. No rhyme or reason, other than someone typed it and we've learned the language. I'm like you, I develop algorithms in plain English first, then translate to code speak.

Cheers,

Vac

P.S. Here's another project from the skunkworks. Might call it the "uber cheap" sensor. Three tubes: pitot, static and P45 (AOA). Not approved for use in icing conditions ;)--oh, and not tested yet either.

3c039a_b3d349695e804d3fb7b904821110232c~mv2.jpeg


No reason you couldn't glue or strap a second tube under your current 90 deg pitot and fabricate something similar...I've seen some folks rolling their own like this for lift reserve indicators...
 
Last edited:
AoA probe locations

Vac, love your replies! :)

I have a knack for getting myself side-tracked. What I should be doing is flight-testing. Given that I'm having a hard time finding days with optimum conditions, I figured I could could speed it up by getting Vs, Vx, Vy and glide equivalent AoA numbers that would then be good for any weight and density altitude and thus shorting that part of flight testing. Unfortunate that the D-100 does not have a flap position input, but I guess if calibrated with flaps up it won't stall before warning with flaps extended :)

Adding a tube to the existing pitot is not a bad idea to get a Dynon D-100 non-normalized AoA output. But I'm not giving up on my pull-rivet yet. ;) Not until I've done some more test flights at different density altitudes and weights and see how bad it is.

Still haven't gotten a good grasp of all the factors. I know the reason we have the pitot at a distance from the wing is to get it out into the free air stream to make it less sensitive to AoA. But, thinking outside the box, now I'm wondering if putting a pull-rivet (or small opening) in the leading edge for Pfwd would result in Pfwd/P45 data that would be somewhat linear and independent of weight and density altitude. Obviously not usable with the D-100 and more a subject for future experimentation and building my own system.

For now, I'll just use the Dynon D-100 and add aural tones to my Arduino.

Again, thank you for your and Lenny's replies. Very educational!

Finn
 
Last edited:
Back
Top