UPDATE: Jan 2009
Although I said in the previous page that I was not going to work out the design of the two-sensor "true" AoA meter, I decided to revisit the issue.
The theory of the 2-sensor meter is very similar to the LRI style meter, but it doesn't have the same limitations of only working in a narrow range of calibration. To review the LRI, it is calibrated at 1-G, straight and level stall speed. In this configuration, let's say around 50 knots, there is a certain magnitude of ram pressure being generated (1.6 inches of water), and a certain degree of differential pressure due to the orientation (calibration) of the 2-orifice probe. Combine both together and you get total pressure, which is all the LRI actually measures. Stay close to 1-G and everything works well. However, put the aircraft into an accelerated stall condition, and the plane will be traveling much faster through the air (90 knots, perhaps?), and the ram pressure goes up (5.3 inches of water) but the probe is in the same configuration (meaning angle of attack), and thus is producing a very similar degree of differential pressure. You can see right way that an accelerated stall will "fool" the LRI into thinking you have plenty of total pressure (what they call "Lift Reserve") and have a safe margin from the stall AoA. No good at all for accelerated stalls!
The fix for this problem is to use two sensors and take a ratio of pressures. One sensor measures pure ram pressure (pitot tube), and the other measures differential pressure (AoA probe). The sensors measure pressure, output the result as a voltage, and the AD711 and AD633 integrated circuits do the math of dividing the two. The result from all this is that the magnitude of differential pressure is always corrected for the magnitude of ram pressure. Looking back at the math might make this easier to understand. Look at situation #2, the last line of the equation. See that long term of sin and cos? It is multiplied by V, or the pressure generated from velocity (i.e. ram pressure). If we divide that term by V then we're left only with pure AoA! I prepared some examples to show how the two sensors will work together translating airspeed into pressure, then into voltage, then through the division, to the final result that will be sent to the LM3914.
So using this formula, we'll develop a plan to guide the schematic design. I worked up a revised design spreadsheet to work out the resistor design. I've decided to use a 10-segment LED bar graph for initial testing, but I'll probably go back and change it over to a 3-LED display like the previous AoA Meter.
The schematic looks a bit more involved, but it's not really all that complicated. The BOM is from Digi-Key, and works out to about $70. The meter can be arranged nicely on a Radio Shack 276-170 prototype board.