Best way to determine pitch?

[nicnac211]

I am currently working in a group to design an unmanned glider controlled by one axis (elevator) released from a weather balloon at 30k feet taking atmospheric readings on the way down.

The original idea to control the pitch of the aircraft was to use a pitch sensor and go from there. The problem is that the sensor is a fluid enclosed in a sphere, which is fine unless there is an acceleration put on the plane that could cause the sensor to give incorrect readings.

My question is this: What other sensors are used in UAVs to control pitch and what is the lightest and cheapest alternative?

The entire weight of the glider is 3 lbs due to the weather balloon weight restrictions.

[Joe2003]

The most used would be an accelerometer/gyro mixed using some form of micro controller.

You could use something like the fma copilot, they are cheap but I dont know what the effects would be like at high altitude?

[maxvtol]

The Co Pilot manual does state it's not quite as effective with cloud cover below 200 ft.

I would imagine there would be problems if there were clouds below and above, or going through a cloud.

Maybe someone more knowledgeable could comment on a barometric pressure sensor. I think that might be at least part of the solution.

[m1tch37]

I can't find the exact code, but this person made exactly what your after:

http://forum.sparkfun.com/viewtopic.php?p=30187&highlight=#30187

Look around the board, i have seen his code somewhere before.

Edit:
http://home.comcast.net/~michael.p.thompson/kalman/AVR_IMU_2.zip

Here is his code, it uses a AVR ATmega168 coupled with a SparkFun IMU 5 Degrees of Freedom. The readme couldn't make it any easier.
It produces the output: pitch= +0.3 roll= -2.6 and you get a second axis as a bonus.

I have not tested this code, and i would be cautious about using it in a plane before i did, but its a nice easy start.

Hope it help,
Mitch

[maxvtol]

Why wouldn't just an airspeed sensor work? Need more speed, nose down. Too fast, nose up.
Eagle Tree airspeed sensor (http://www.eagletreesystems.com/MicroPower/air.htm)

[m1tch37]

It could give you a relative pitch, but absolute, no. If headwind increases, you will be lead to believe you are speeding up.

[maxvtol]

It could give you a relative pitch, but absolute, no.
What other pitch is there besides relative to the air, and why would you want to use that? Constant glide slope would be difficult to maintain in a glider, especially from 30k ft to the ground.


If headwind increases, you will be lead to believe you are speeding up.

A headwind would have no affect on airspeed at constant pitch, groundspeed will be slower.

Edit:
My references to pitch should actually be called angle of attack. Pitch is the rotation of the plane about it's y axis which is controlled by the elevator.

[macboffin]

Only controllable in pitch? What about rudder, how you going to steer the thing? Or is it a throw-away? Onto someones head maybe?

[macboffin]

OK, being serious this time. Presumably you will want to equate the readings you get with altitude, corrected for density, possibly humidity, as well as temperature. Are you going to transmit these reading in real time or store on-board for later evaluation? I don't see any need to have pitch control, the natural trim of the glider will be largely unaffected by air density, and the glide slope will remain the same, it will just fly down that slope faster at higher altitudes, in other words descent rate will reduce as height reduces.
The primary control should be rudder, ( rather than aileron) for directional control so that you recover your investment, and don't make dents in the landscape. From 30k ft your potential glide range is large.Another consideration is wind speeds, which may well vary on the way down, as well as their directions.Head wind faster than gliding speed means you will have to get your running shoes on to meet the plane at arrival.
Lastly, what about air traffic and an unscheduled meeting with other traffic? What safe guards? At the least file a flight plan, fit strobes.
Plane getting too heavy? So a tied pair of balloons. Doubles the payload.

[murmur]

I'm with the rest of these guys: knowing pitch angle isn't going to do you a lot of good. An accelerometer to measure your angle of attack (indirectly) and a baro sensor to tell you altitude are going to be your quickest route to elevator control. For roll control, you'll most likely want to infer roll angle from heading changes (if you maintain your vertical speed at a fairly small percentage of total speed, there's a direct relationship between heading rate, lift (i.e. your accelerometer output), and roll angle). Heading rate can be measured with a gyro, if you maintain roll angle to small values.

If vertical speed gets to values that are a major percentage of total speed (~20%), or if you want to do a lot of rolling maneuvers, then you're going to need to add more complexity, e.g. GPS.

Dave

[Nick Reffold]

I would have thought that an AoA sensor is what you need to keep the wing at a predetermined AoA for either duration or range.

There were some discussions about how to design such a sensor a while back. An external vane, free to rotate on a shaft through the fuselage skin. Inside end of the shaft drives a rotary encoder.

I don't know if anyone has built a cheap one for hobbyists. Rotary encoders can be pricey. I was wondering whether the mechanics of an old wheel mouse might be adapted.

[murmur]

Nick, you're right, an AOA sensor would be a way to get around the need for a pitot (total pressure) sensor (which I forgot to mention, in my remarks, but it would be necessary if you didn't have an AOA sensor).

Nicnac, if you're going to wait for nice weather to fly this thing, then you wouldn't need a heated pitot tube (or a heated AOA sensor mechanism). Otherwise, either one of these will need some way of assuring that ice won't gum up the works.

Dave

[shanghai_fool]

I don't know if nicnac is still around but since the discussion turned to AoA, I thought I would post a question.
A while ago, I was web searching for an airspeed solution and came across an article that used a thin film platinum sensor that could be flush mounted (like just above the leading edge) and would measure very low airspeed (or angle-of-attack). This would be perfect for low speed aircraft like gliders, helicopters (VTOL's), etc. They use regular platinum wire sensors to measure air flow in HVAC ducts.

The theory behind this is basicly, heat the wire to a constant current and measure the output of the feeback loop that maintains it. Any cooling effect will cause the FB loop to compensate. Of course, a correction is applied by measuring ambient temp (varies with altitude).

I was trying to figure out how I could apply this to a model. Any ideas?

[murmur]

Shanghai_fool, that's regularly referred to as "hot-wire anemometry", and I think you could probably make some useful air-data sensors with those devices. You could place several of these at various positions on the leading edge of the wing (all at the same spanwise location), and look for the one showing minimum airspeed as an indication of AOA. Probably best done with some software.

For pure airspeed sensing, I wonder if you would need to compensate the readings from this sensor for altitude. I'm not sure, but I think perhaps the cooling effect on the sensor varies with air density as well as airspeed. In that case, you'd need a static probe hooked up to a pressure sensor to get altitude, as well.

Dave

[shanghai_fool]

Thats why the temp correction. It the variation of ambient temp that decreases with altitude ( at least up thru the troposhpere 10km).

I wish I could find the article I read but haven't yet found it. It looked a lot easier than a pitot tube and more accurate at low speeds.

I was more interested in that than the aoa since I am intrested in VTOL's, however, the same device could be used for both applications. I would like to get a thin film platiunum sensor and try it out. It does take more current ( a few ma as I recall) to heat the wire but is much more accurate at very slow speed.

[Gary Warner]

I am currently working in a group to design an unmanned glider controlled by one axis (elevator) released from a weather balloon at 30k feet taking atmospheric readings on the way down.

The original idea to control the pitch of the aircraft was to use a pitch sensor and go from there. The problem is that the sensor is a fluid enclosed in a sphere, which is fine unless there is an acceleration put on the plane that could cause the sensor to give incorrect readings.

My question is this: What other sensors are used in UAVs to control pitch and what is the lightest and cheapest alternative?

The entire weight of the glider is 3 lbs due to the weather balloon weight restrictions.

nicnac211,

Don't know if you're still following this thread, but look at my AOA project for some ideas. AOA Project (http://web2.airmail.net/warner1/)

With ideas like fluid in a ball, I think it's best that you dive a bit deeper into aerodynamics and flight stability.

[macboffin]

Air humidity will also affect platinum sheet or wire sensors. Thats a more difficult effect to allow for.

[macboffin]

It could give you a relative pitch, but absolute, no. If headwind increases, you will be lead to believe you are speeding up.
Aircraft expeience a headwind all the time.It's due to forward motion. Any movement of the airmass it's flying in, in the opposite direction, is a headwind.If it is a 50 mph wind, and the plane cruises at 50, it will stand still relative to the ground.Doesn't affect its flight or glide slope!

[murmur]

Thats why the temp correction. It the variation of ambient temp that decreases with altitude ( at least up thru the troposhpere 10km).

Pressure variation ends up being a bigger driver of density over typical full-size aircraft altitude ranges, but if your altitude is only going to vary 1000 ft, you can probably get away with just a temp sensor for calibration of the anemometer.

I was more interested in that than the aoa since I am intrested in VTOL's, however, the same device could be used for both applications. I would like to get a thin film platiunum sensor and try it out. It does take more current ( a few ma as I recall) to heat the wire but is much more accurate at very slow speed.

That's a good point; the total-pressure variations from 0 to 50 ft/sec are probably too small to grab reliably with a mass-produced pressure sensor.

Dave

[anci3nt]

nicnac, are you in the US? If you are, how do you plan to address the 'no aided decent' clause, or impact precautions in the FAA weather balloon reg?

Don't mean to poop on the party, just curious.