Practically the first thing anyone wants to know about a new aeroplane is how fast it goes. This very basic question is curiously hard to answer.
Airspeed is measured by the pitot-static system. (The word pitot, by the way, comes from the name of Henri Pitot, who came up with the idea back in 1732. Pitot was a Frenchman, and so we pronounce the word pee-toe and not pit-ott.) The idea of the system is to compare the pressure generated by the impact of moving air – the “dynamic pressure” — with the pressure of stationary ambient air – the “static pressure.” The sum of the two at any location is called the “total pressure.”
Dynamic pressure was originally
displayed and measured by means of a manometer, a U-shaped tube with some water in it. One leg of the U is
connected to the pitot tube, the other
is vented to a sheltered location where the pressure is static. When moving air
strikes the entrance to the pitot tube, it displaces the fluid surface in the pitot leg downward and the surface in the other leg the same distance
upward. The difference in height of the two columns is proportional to the air pressure. Consequently, small pressures are often measured in inches of water; larger ones, like manifold pressure, in inches of mercury. A mechanical airspeed indicator works differently. An airtight case contains a flexible bellows made of thin metal. The bellows is vented to the pitot, the case to the static port. When pressure from the pitot is greater than ambient, it expands the bellows, and a system of gears and levers converts the motion into the swing of a needle. The accuracy of the reading depends on the quality of the mechanical movement, of course, but it also depends on another factor that is much more difficult to control. This is the so-called “position error,” which occurs because both pitot and static readings are affected by the changing velocities and pressures on the surface of the aircraft.
Dynamic pressure is the easier of the two to get right. Because velocity and pressure vary in inverse proportion to one another, local variations tend to cancel themselves out. That is why a pitot tube can be located on a wing, even though the wing is the part of the aeroplane that is designed to experience large pressure variations with angle of attack.
What can induce errors in the pitot pressure, however, is misalignment with the air stream. A misalignment of five degrees produces a negligible reduction in pitot pressure, but a misalignment of 15 reduces the pressure by 3.5 percent. To solve this problem, pitot tubes are often placed near a surface, like a wing, that forces the airstream into alignment with them.
Measuring static pressure is more difficult. Some pitot tubes have small holes drilled into their sides, separate from the dynamic-pressure plumbing. More usually, however, engineers hunt around for a location on the airframe where the local pressure remains nearly constant over the full range of angles of attack and flap settings. Often this turns out to be on the fuselage sides somewhere between the wing and the tail. Ports are placed on both sides in order to cancel the effects of yaw. A good static port location is one where the pressure is neutral to begin with and is not affected by angle of attack.
How is that hunt conducted? In order to evaluate a candidate static location you need a reliable reference. This was traditionally provided during flight test by a “trailing bomb,” basically a piece of flexible tubing thirty or forty feet long with holes drilled in its sides near the aft end. The tube trails far enough behind the aeroplane to be unaffected by its pressure field. Thin tubes taped to the surface of the aeroplane, terminating at various candidate static port locations, provide readings that can be compared with the presumably correct one from the trailing bomb.
A less outdoorsy method, these days, is to map pressures with computational fluid dynamics software, locating islands where pressure is both neutral and stable.
Serious flight tests are conducted with a “test boom,” a long, stiff tube with a gimbaled, weathervaning pitot on its nose – and a trailing bomb. Differences between test results and the readings of the airplane’s built-in pitot-static system – preferably, if the marketing department has its way, ones that make the airplane appear to cruise faster and stall slower – are plotted on a calibration chart and buried in the pilot’s handbook, where nobody looks at them again.
Ultimately, production pitot-static
systems receive their final validation in flight. Various techniques are
available. A nice simple one is to fly alongside an aeroplane with a
well-calibrated system. Another
is to fly over a measured ground
course at various speeds; but for that method to yield
wind is not a very precise thing to start with, good results the wind speed and direction must be known with pretty good precision. An objection to the measured-course method is that it must be used at low altitude in order to get good fixes on the starting and ending points, and that can be hazardous at speeds just above the stall.
GPS has made possible a new set of techniques for airspeed calibration. One method is a variation on the traditional ground course approach. A distant waypoint is established and the aeroplane is flown toward it, using the GPS to measure, say, a two-mile course. Repeat in the opposite direction. The true airspeed is more or less the average of the two GPS groundspeeds.
GPS can be used to ascertain both the wind’s direction and its speed, provided that the wind is fairly strong. The pilot flies in a circle at standard rate, monitoring the GPS groundspeed. When the aeroplane is flying directly upwind or downwind the groundspeed is at its maximum or minimum, so you have the wind direction. The wind component is half the groundspeed difference. The difficulty with this method is that the groundspeed barely changes over a 30-degree range of headings bracketing the wind axis, and so it is difficult to get a precise fix on the wind direction. On the other hand, and so an approximate wind direction is good enough.
A more mathematically complex method involves speeds taken on several headings. The procedure involves selecting an altitude (any altitude will do, since it’s indicated, not true airspeed that’s being used) and recording the outside air temperature, which you will later use to convert your indicated airspeeds to true. You set power and allow airspeed to stabilise. Record the airspeed, the heading, and the GPS groundspeed. Then turn 90 or 120 degrees, let the indicated airspeed stabilise (it should be the same on each leg), and record the heading and GPS groundspeed again. Fly three or four headings. Repeat this procedure for the complete speed range at intervals of ten or twenty knots.
If you Google “GPS airspeed spreadsheet” you will find several Excel spreadsheets ready to use. Enter the headings and airspeeds you collected; the average groundspeed, wind speed and direction will appear. (Note that the groundspeeds are not averaged in the usual sense; they are extracted from the data by a mathematical procedure that would be a burdensome chore if the spreadsheet did not take care of it for you.)
When I did this, I tried the circle method of obtaining the wind at the same time, and recorded groundspeeds for upwind and downwind segments flown along the wind axis. My estimate of the wind direction based on circling differed from the calculated average by about 10 degrees, but at the same time it was interesting to see that the wind component varied from one data set to another – not surprisingly, since there is no reason to expect the wind to be absolutely uniform, especially in a mountainous area such as the one where I am doing my flight testing. The airspeeds obtained by the two methods differed by about three knots.
What is the use of precise airspeed information, though, when the actual progress of the aeroplane from one place to another may be subject to the vagaries of winds, traffic delays or the tardiness of passengers? You get there when you get there. Furthermore, what difference does it make whether the airspeed indicator is right or wrong, as long as it varies in a continuous way from slow to fast? It all has to do with bragging rights, I guess. With respect to the speed of my own aeroplane I shall imitate the noble reticence of Rolls-Royce, who, when asked the power of the engines in their cars, would merely reply, “Sufficient.”