DIY Airboom, JLJ and GC 2013

Find here an introduction to BasicAirData Open Project 

Design constrains of a basic air data system is highly application dependent, herebelow you find a general arrangement that seem correct for many RC/FPV/UAV aircrafts.

You will  find indications about early stage requirements. The figure below refers to an airboom built by the authors, often the same measures can be take with separate devices.

What are we supposed to measure?

Basic airdata is all about measurements that can give information about physical characteristics of the air mass that surrounds the aircraft.

The air is assumed to be dry, so nothing must be done to account for moisture content. With the value of temperature and pressure and using commonly available data it’s possible to retrieve all the physical characteristics of dry air.

If the flight envelope is limited to low airspeeds, let’s say under 55 m/s, is needed only a slightly correction for outside temperature readings in so far as temperature ram rise [1][Ref1 formula 14] have a maximum value of 1,5°K @standard conditions.

The following data is needed for a minimal air data system:
Direct measured data

  • Outside air temperature OAT
  • Outside static pressure P_{s}
  • Total pressure P_{t}
  • Angle of attack, angle of sideslip AOA, AOS.

Wind vane design sensor is a direct method the skin pressure based methods require data processing and formally must be placed on the table here below

Calculated data:

  • Speed, IAS, CAS, MACH, TAS; Need difference between Dynamic pressure and static pressure, TAS need OAT

Quite often the speed is called Pitot speed or speed.

  • Barometric altitude
  • Rate of climb; Need barometric altitude

System components

Three elements are always presents:

  • the mechanical probe;
  • the pressure transducers;
  • the air data computer.

Redundancy and maintainability should be carefully considered.

Temperature probe considerations

Outside temperature can be measured by a RTD element,  the sensing element should be carried by a mechanical probe that allow contact with the airstream outside the boundary layer of the aircraft. Usually this requirement lead to place the sensor at few decimeters far the aircraft skin. Another big issue with temperature sensor is that the RTD must be shielded against the sun rays and conductive heat transmitted by the fixation flange. We should observe that typically when the vehicle is at rest the reading of temperature can be biased, in such cases an aspirated probe should be applied. At low speeds can be convenient also to use skin mounted OAT probes.

Static pressure and total pressure pressure ports

Pressure ports can be located on the probe or on the airplane skin. The static port is is the more sensible to installation position, the best location is on a nose mounted probe [Ref 2 page2-3][2]. Total pressure port is less critical, a good performance can be attained on a nose mounted configuration. Installing the probe on the fuselage or on/under the wing need for compensate for the position related error. Nevertheless also with a nose mount boom configuration the body of the probe itself can cause some aerodynamic interference at high AOA an AOS.

Some consideration on pressure lines tubing

Every airborne system have two main issues that must be consider to attain an accurate and reliable system. The dynamic response of the system is highly dependent on the pneumatic lag , this phenomenon is due to the propagation time of the pressure wave through the sensor lines. For static measure it have no impact but in case of high rate measurements can have a heavy impact on system performance, the lag magnitude is often about 5e-3 s/m. The other main issue is the volume machting for pressure ports. If the volume/lenght of different pressure lines, e.g. Static and dynamic, are different the reading will be affected by time jitter. Both these problems can be alleviated by shortening the pressure lines lenght, hence the pressure sensor will be located close to the mechanical probe exits. Volume coupling and line lenght reduction yield to an increased dynamic performance. It’s a must to provide a mean to drain condensed water from the pressure lines and to warrant that no water reach the pressure transducers. The tubing should be rigid as deformation lead to reading errors, use of standard RC silicone tubing is discouraged.

Pressure transducers

Sensors must to be rugged and easy replaceable. The desired accuracy should be granted over the entire operative temperature, acceleration and vibration range. All the sensors need to be recalibrated and tested on a regular basis. The pressure range of the pressure transducers is dependent on the plane speed range.

Air data computer ADC

This unit receive sensor wiring and tubing and contain sensors, electronic and communication interfaces. In a stand alone manner this unit is able to provide all the airdata to another connected unit, as as simple example layout think that the ADC is connected to a single flight computer. It is highly advisable to communicate by means of a digital bus to enable redundancy and real time failure detection and reconfiguration. It’s also wise to provide galvanic insulation or another form of decoupling to avoid that single defective unit lead to a bus failure or damage hardware of other online units. Safety oriented solutions are to be considered even on a small systems. In the case of use of some sort of flight closed loop stabilization /control/regulation system a lack of accurate airdata may lead to  a brutal loss of performances, in many cases to avoid total failure should be necessary that flight computer autonomously switch to a simpler control algorithm so ADC should communicate every anomaly to flight computer.  ADC elaborate sensor data with some equations, they will be defined in the appropriate instrument page and they will also involved in the uncertainty analysis. Be very careful when considering sensors and interfaces to use, analog electronics is by far most critical to use than digital, often require a moderately complex auxiliary circuitry as stabilized power sources, analog filters and so on.

Drafting Air Data requirements

Air data requirements are not airframe or mission independent. If we are unable to put down some typical mission profiles we should be tempted to use a shorcut, inspect documentation for status of the art similar equipment and try to match similar performances. Unfortunately in many cases copying specifications can lead to a big cost and or to miss telemetry objectives, so I’ll put aside this approach.
For example purposes let’s pretend to identificate the coefficient of lift and drag of our model in steady flight, relative uncertainty in the measured cl and cd 1%.

Work hypotheses [3]Chapter 4,

  • Unpowered flight descent
  • All the parameters that not will be not measured are supposed to be exactly known.
  • Sideslip angle is zero
  • Lateral and longitudinal dynamic decoupled
  • Small angles
  • Steady angle of climb and speed
  • Available sensor readings are Pitot speed and gamma
  • Only instruments uncertainty is considered, every other source as electronics or software are neglected in this preliminary phase

At first it’s needed to retrieve a good, validated, math model for the aircraft under study for the desired flight condition, in this case refer to[3] chapter 4.

The following values are assigned to constants:

mg=40\ N
Sw=0.2\ m^{2}
\rho=1.22\ kg/m^{2}
Reading from sensors, pitot speed and trajectory angle:
V=25.61\ m/s

\gamma =-0.09\ 1/s

Straight from formula 4.29 e 4.30 of reference [3]:

    \[Cl=\frac{mg}{\frac{1}{2} \rho V^2 Sw}\]

    \[Cd=-\gamma Cl\]

The uncertainty of Cl is dependant only on pitot measurement and the expression is: 

    \[ubCl(V)=\sqrt{\left( \frac{-4mg}{\rho Sw V^3}\right)^2 ub (V)^2}\]

Our design costrain, 1% total Cl uncertainty: 

    \[\frac{ubCl(V)}{Cl}\le 1\]

Hence

    \[ubv(V)\le 0.125 \frac{m}{s}\]

Now we calculate Cd uncertainty using a first guess ubgamma value
of 0,1/3606,28 radiants (corresponding to 0.1°):

    \[ubCd(\gamma, Cl)=\sqrt{(-\gamma)^2 ubCl(V)^2 + Cl^2 ubgamma(\gamma)^2}=2\ e-3\ rad\]

So relative uncertainty ubCd/Cd=3,88\%

You can find here a simple Scilab script for example calculations.
The total uncertainty for Cd is far from our initial requirement of 1% nevertheless the required accuracy for the angle measurement is quite elevated so it can be realistic to relax the design constrain. During the steady unpowered descent the angle of attack is equal to the angle of trajectory gamma, for this maneuver it’s possible to retrieve gamma from both the IMU or the airboom data. Measure the drag coefficient with the proposed equations require an extremely precise pitot and attitude data, fortunately the final Cd uncertainty can be reduced using system identification techniques and a better designed flight experiment. The data that have been obtained can be used to proceed to the instruments design phase, however for every different flight plan and aircraft configuration it’s necessary to check if uncertainty constrains still to be meet.

Range requirements

Static pressure P_{s}

Regarding the range of pressure every barometric sensor can do the job, the aircraft fly in the very low layer of troposphere so there is no range problem, only exception is when the unit take off from a high mountain airfield. There are many different devices out there, many already temperature compensated and capable to send out a filtered altitude digital value in meters. I have experienced two issues with this kind of devices, low response time and lack of control on the data process phase that is internal to the device; on the other hand they are easy to use don’t need for external circuits and are cheap and well documented. The barometer reading can be used to calculate the rate of climb, for this use consider a unit that can handle at least ten samples per second. Extra rate is useful for oversampling and averaging and to have reliable information during dynamics maneuvers. Regarding oversampling look an example datasheet from Freescale site, on table 2 line 2 you can observe that noise can by reduced with the use of this simple technique.
Without entering in the control details consider as example an airplane that have a simple closed loop altitude regulator. This plane is cruising at 35 m/s 100 AGL meters of altitude and have a barometric altimeter running at 1hz refresh rate. Suddenly the controller is required to bring the plane to a 200 meters AGL altitude. After a very long time the plane reach the desired altitude. The limitation on the sensor dynamics translate in a narrow control band. With a moderate wind gust presence the regulator will simply fail to level the plane. The situation can be mitigated with the use of more sophisticated regulators but it by far easier to choice a higher rate barometer. In a RC/FPV/UAV typical application the limiting factor is the servo frequency response.
Overall accuracy of pressure sensor should be about 4 pascal, this value used with the US standard atmosphere model lead to an uncertainty, for sea level altitudes, of about 0,3 m; at sea level the pressure is 101325 Pa and a 0,3 m 101321 Pa.

Differential pressure P_{t}-P{s}

Usually we call the device that measure the differential pressure between the total pressure port and the static pressure port a Pitot tube. In order for this to happen we need to assume that, in the low speeds a common aircraft flies at, the air is incompressible and thus this pressure differential is the dynamic pressure, denoted with q.
By definition q=\frac{1}{2} \rho V^2=P_{t}-P_{s}.

References

[1]
F. Trenkle, “AGARD FLIGHT TEST INSTRUMENTATION SERIES VOLUME 2. IN-FLIGHT TEMPERATURE MEASUREMENTS.” Feb-1973 [Online]. Available: https://www.dtic.mil/dtic/tr/fulltext/u2/758589.pdf
[2]
E. A.Haering, “Airdata Measurement and Calibration.” NASA, 1995 [Online]. Available: https://spaceagecontrol.com/nasa-tm-104316.pdf
[3]
D. G. Hull, Fundamentals of airplane flight mechanics. Berlin; New York: Springer, 2007 [Online]. Available: https://site.ebrary.com/id/10158140