Designing an Instrument

[HELModels]

I am trying to design a simple instrument that will indicate tilt(pitch or roll). I've narrowed some of it down to a fluid with some kind of buoyant object inside. As the container is tilted the object seeks the surface. I need the buoyant objects to actually be about .1 mm in diameter, so that a bunch of them will go past a sensor as they seek the surface. I define the size as .1 mm because the sensor sees about 4 sq. mm area.

1. what kind of material is buoyant in water, very small?
- Do you think I could crush and grind styrofoam?

2. Where can I get a rectangular tube made out of clear thin plastic?


Thanks for any help

[JetPlaneFlyer]

If the instrument is statically mounted then what’s wrong with a pendulum?... keep it simple is my motto.

If the instrument is to be mounted in a plane then you would need something like a gyro because flight dynamics would mean your idea (or a pendulum come to that) would not accurately measure 'tilt' from the horizontal.

[HELModels]

This is about as simple as it gets, really. You mean in a coordinated turn, there would be no motion?

[JetPlaneFlyer]

This is about as simple as it gets, really. You mean in a coordinated turn, there would be no motion?

Partly yes, in a perfectly coordinated turn there would be no motion at all but in a less than perfect turn the 'buoyant things' could go either way, not necessarily in the direction that would indicate the actual roll direction (or pitch 'direction' come to that). A plane in flight is exposed to constant dynamic forces which would make the rising bubble instrument meaningless as far as bank and pitch indication is concerned... In negative 'g' situations the buoyant objects would actually go down!

It also does not sound in the least bit simple to me.. The 'buoyant objects' would simply sit at the top of the tube or float on the surface (unless negative 'g').. So to get them to 'float' up past a sensor you would need some mechanism that constantly fed the objects into the bottom of the tube so the ‘sensor’ could record their direction of ascent.. and something to skim them off the surface too. Or do you have something else in mind ?
Then you would need to process the output of the 'sensor' into some meaningful signal.. Can you even get a suitable sensor that would be small and light enough to go in a model?. On a power model the whole thing would probably get shaken into froth anyway unless you had excellent vibration isolation.
It all sound horrifically and unnecassarily complex.

What about a simple and cheap RC helicopter gyro which is specifically designed for this very propose (i.e. detecting angular motion)?

What’s the object of the exercise anyway?

Steve

[HELModels]

The sensor is an optical sensor with all the complicated math inside. The output is simple TWI with dx, dy outputs. If the vessel contained the bouyant particles and the sensor were placed on top, then when the particles were at the top the sensor would see them. I've already done some simple experiments with a bottle(square w/flat face) of "froth" - soap. I shook and rattled and tilted the bottle and the sensor indicated direction and magnitude right on my computer screen. The liquid would slosh and push the bubbles along the vessel walls. I wanted some kind of buoyant object that wouldnt pop and go away, but the thought about false indications due to dynamics vaguely ocurred to me.

The apparent complexity of output is solved. Finding the right fluid, vessel, and understanding the dynamics is where I am now.

[HELModels]

What about a simple and cheap RC helicopter gyro which is specifically designed for this very propose (i.e. detecting angular motion)?

What’s the object of the exercise anyway?

Steve

A $5.00 mouse chip is cheap and small. The object was to see if I could make a pitch and roll indicator. As for buoyant particles, I think the frothy soap might actually be pretty good, but I need a smaller vessel with a flat clear face.

I might actually have a simple accelerometer and not an "angle indicator". The more force I put into a direction, the bigger the magnitude indicated.

[Montag DP]

A $5.00 mouse chip is cheap and small. The object was to see if I could make a pitch and roll indicator. As for buoyant particles, I think the frothy soap might actually be pretty good, but I need a smaller vessel with a flat clear face.

I might actually have a simple accelerometer and not an "angle indicator". The more force I put into a direction, the bigger the magnitude indicated.Remember that whenever the instrument is accelerating, the particles inside are subjected to inertial forces. This means that in flight your instrument would read the net of inertial forces and the buoyancy force, rendering it more or less meaningless. If you could get neutrally buoyant particles, then you'd have a type of accelerometer.

In steady state the idea sort of works, except that all the particles will quickly move to the surface and then the reading will be gone.

[HELModels]

Remember that whenever the instrument is accelerating, the particles inside are subjected to inertial forces. This means that in flight your instrument would read the net of inertial forces and the buoyancy force, rendering it more or less meaningless. If you could get neutrally buoyant particles, then you'd have a type of accelerometer.

In steady state the idea sort of works, except that all the particles will quickly move to the surface and then the reading will be gone.

I appreciate your input. I seem to remember you have made some of your own instruments, IIRC.

There's no substitute for trying stuff, though. I took some styrofoam and broke it up into beads and put them into a glass vial fillled with water. With the mouse taped to it and a magazine wrapped around to prevent the mouse moving relative to the vial, I tilted the vial. The beads zipped to the other end and the mouse indicated. Tilt the other way, opposite indication.

I admit to not fully understanding the forces that will neutralize it.

[eflightray]

An instrument using fluid ?

Reminds me of a bit of film I saw of a bloke who used to do the air shows in the States, (may still do). He would fly a twin doing aerobatics, ........with both engines off using inertia, and a glass of water, (the fluid bit), just standing on the dash. He would do a roll without spilling any water, amazing.

Sorry for the slight off topic, it's funny how something bring back memories.

Isn't the mechanical tilt switch the basis for the cheap car alarm sensor ? I also remember there was once a tilt hand controller for computer games that replaced the joystick.

Rambling again. :D

[Montag DP]

I appreciate your input. I seem to remember you have made some of your own instruments, IIRC.

There's no substitute for trying stuff, though. I took some styrofoam and broke it up into beads and put them into a glass vial fillled with water. With the mouse taped to it and a magazine wrapped around to prevent the mouse moving relative to the vial, I tilted the vial. The beads zipped to the other end and the mouse indicated. Tilt the other way, opposite indication.

I admit to not fully understanding the forces that will neutralize it.I think you're thinking of someone else, because I've never made my own instruments.

Your ground test would work, as the mouse indicated, but the problem is that in the air there are a lot of other forces present which will mess things up, as you already know. It is an interesting out-of-the-box idea, though. Maybe it could be useful for something. :confused:

[nauga]

Sounds like you're describing an inclinometer. Most full-size airplanes have one in the form of a slip-skid indicator, or 'ball'. In a coordinated turn, i.e. one with no sideslip or sideforce, the ball is centered in the race, regardless of the airplane attitude - centripetal acceleration keeps it there. That's why full-size attitude sensors use gyros or one form or another (in addition to other sensors).
http://www.friebe.aero/images/artikelvarianten/image_2569_1_thumb.jpg
It's just a ball in a fluid-filled race. In other-than-level attitudes, the ball moves depending only on acceleration, not attitude.

Nauga,
a strapdown kind of guy

[nauga]

Reminds me of a bit of film I saw of a bloke who used to do the air shows in the States, (may still do). He would fly a twin doing aerobatics, ........with both engines off using inertia, and a glass of water, (the fluid bit), just standing on the dash. He would do a roll without spilling any water, amazing.Bob Hoover. He also had a ping-pong ball (table tennis, maybe?) on a string below the glass, and the ball always stayed in the same position (relative to the rest of the cabin parts) regardless of attitude.

Nauga,
who used a stopwatch instead of a ping-pong ball

[JetPlaneFlyer]

HEL,
Using the eye of a optical mouse is really quite a clever idea :) ..
However i think there are easier ways to execute it than using a liquid with stuff floating in it. Could you use a pendulum with that swung past the eye? This would have the potential of giving you full proportional inclination readout rather than a crude 'on-off' type output that you would get as lump of 'floating stuff' passed the eye.

As has been said before though.. this would not work for a plane in flight due to dynamic forces.

Steve

[HELModels]

What is that they say? Ignorance of the law, or laws of physics, is no defense.

I've got one more trick to try and that is to put the mouse right smack on the display of a digital camera. It might be able to see enough pixels to detect motion.

[HELModels]

HEL,
Using the eye of a optical mouse is really quite a clever idea :) ..
However i think there are easier ways to execute it than using a liquid with stuff floating in it. Could you use a pendulum with that swung past the eye? This would have the potential of giving you full proportional inclination readout rather than a crude 'on-off' type output that you would get as lump of 'floating stuff' passed the eye.

As has been said before though.. this would not work for a plane in flight due to dynamic forces.

Steve

O.K., lets talk about this pendulum. I could rig up something. Just about any surface with minute texture is visible to the mouse. This pendulum would indicate what?,,acceleration?

[JetPlaneFlyer]

O.K., lets talk about this pendulum. I could rig up something. Just about any surface with minute texture is visible to the mouse. This pendulum would indicate what?,,acceleration?

Yes it would indicate acceleration. But in a flying plane the reading would be a mix of acceleration due to gravity (always downward) and acceleration due to flight dynamics (any direction).. So i'm afraid it would probably be quite meaningless.

[HELModels]

I've measured motion due to gravity with the soap froth. With the sensor over the air space of the bottle, I tipped the bottle upside down, right side up and then let the soap drip down the side. I didnt understand it at first, but the cursor just kept drifting, drifting until the all the soap was off the sides. This was gravity. Maybe I've got a quirky hour glass.

[HELModels]

I took the mouse and placed it on my LCD display with the lid horizontal so it wouldnt slide around. I then played a flight video and what do you know, the mouse cursor moved with the video. It was very small movements, but it moved. One problem is that the object distance is off the plane of focus, so only the grossest changes are detected.

This is promising actually since there is no inertial effects to neutralize motion - what you see is what you get.

Oh well, gotta go cut grass

[BMatthews]

Hel, have you ever been in an airliner and walking back from the biffy when the plane made a turn? I have. The only two sensations that let me know that we were banked and turning were that I felt a bit heavier on my feet and when I looked out one side I could see ground and out the other side sky. If I had been holding a drink the liquid would have been sitting perfectly level in the cup that was being held parallel to the deck. And at the time of this the airplane was in a rather steep turn with a bank of around 20 degrees.

So the only thing your indicator will tell you is what forces are in action in relation to the mounting point of the outer case. In this case the model the instrument is fastened to.

[nauga]

Ever swing a bucket of water around in a circle? If the water level could be used to measure attitude, the water would dump out when you swung the bucket...yet it doesn't if you swing it fast enough.

Nauga,
who has gotten wet demonstrating this

[HELModels]

Hel, have you ever been in an airliner and walking back from the biffy when the plane made a turn? I have. The only two sensations that let me know that we were banked and turning were that I felt a bit heavier on my feet and when I looked out one side I could see ground and out the other side sky. If I had been holding a drink the liquid would have been sitting perfectly level in the cup that was being held parallel to the deck. And at the time of this the airplane was in a rather steep turn with a bank of around 20 degrees.

So the only thing your indicator will tell you is what forces are in action in relation to the mounting point of the outer case. In this case the model the instrument is fastened to.

Right, and if the pilot got sloppy and kicked the rudder pedal, or the stewardess in his lap bumped the ailerons, your drink would slosh. This thing will register that, so it could be used to say "step on the ball".

The soap froth bubbles come close to neutral buoyancy. It takes a few hours for all the bubbles to rise to the surface. If some fluid had bubbles and the bubbles sought the surface at the some rate and a negative G hit, wouldnt the bubbles speed up, or vice versa if positive G?

[nauga]

Right, and if the pilot got sloppy and kicked the rudder pedal, or the stewardess in his lap bumped the ailerons, your drink would slosh. This thing will register that, so it could be used to say "step on the ball". You are correct, but that is acceleration, not attitude.

Nauga,
fixed in space

[HELModels]

Thanks nauga. Your turn and slip indicator is what helped me see it.

[BMatthews]

If side slipping or other uncoordinated flying factors is what you're after then your simple G force indicator will work fine. But if the goal is to actually sense the attitude of the model with reference to the plain of the Earth then no, you're on the wrong track. As mentioned this gizmo you're working on will sense accelerative forces in the plains of its ability. Beyond that it won't.

Depends on what you're actually trying to achieve with this gizmo, doesn't it. If the goal is to connect it to an autopilot circuit then yes this will be a good thing to have. But only as part of a more complex and comprehensive package that will include a gyro stabilized reference that will act as an artificial Earth plain.

[HELModels]

Thanks, Bruce. The original intent was attitude, but it didnt do that. Everyone has helped me see what it actually does indicate.

[nfhill]

I'm aware of two practical 'simple' ways to sense relative attitude to the ground. The FMA Copilot uses IR sensors to detect the horizon. Way back, circa 1980, a well known RC guy invented a way to do it with op-amps.

It turns out that the earth and the ionosphere form a spherical capacitor and the electric field density is great enough to be easily detectable. Connect a simple op-amp as a comparator and the voltage, as a function of relative altitude, of the inputs is usable to sense attitude to the ground. With three voltage sensors, it's possible to detect pitch and roll. This was patented and an auto-pilot system was marketed to RC hobbyists. It didn't succeed as a product, but it did work.

Nick

[nauga]

Way back, circa 1980, a well known RC guy invented a way to do it with op-amps.Maynard Hill. He's still at it - a few years back he led a team that flew an R/C airplane across the Atlantic, nonstop and unrefueled.

Nauga,
with long legs

[HELModels]

I'm aware of two practical 'simple' ways to sense relative attitude to the ground. The FMA Copilot uses IR sensors to detect the horizon. Way back, circa 1980, a well known RC guy invented a way to do it with op-amps.

It turns out that the earth and the ionosphere form a spherical capacitor and the electric field density is great enough to be easily detectable. Connect a simple op-amp as a comparator and the voltage, as a function of relative altitude, of the inputs is usable to sense attitude to the ground. With three voltage sensors, it's possible to detect pitch and roll. This was patented and an auto-pilot system was marketed to RC hobbyists. It didn't succeed as a product, but it did work.

Nick

That's neat and simple. I read something about the latter method, but forgot that. I seem to remember it involved a radioactive element that is hard to acquire, especially these days. There are some new techniques that involve something called the direction cosine matrix.

This link is for an autopilot that uses 2 accelerometers to deduce the angular rotation.

http://www.rcgroups.com/forums/showthread.php?t=1057001

The concept seems to be that the 2 accelerometers wont sense the same force the same way, therefore the difference is due to a rotation that a gyro would sense. Accelerometers dont sense anything when forces cancel and gyros dont sense anything when there's no rotation, but fused together accelerometers and gyros paint a complete picture. What I'm thinking is that my optical accelerometer x 2 could provide 3 axis of accelerometer and 3 axis of gyro.

An idea I had is that a gelatin would wiggle and jiggle and if placed so it could wiggle and jiggle in the focus of the sensor, then it would directly register an acceleration without sources of error that need to be accounted. Any way I do it will involve some type of data fusion.

[nfhill]

That's neat and simple. I read something about the latter method, but forgot that. I seem to remember it involved a radioactive element that is hard to acquire, especially these days. There are some new techniques that involve something called the direction cosine matrix.

Yes, it did require a radioactive source to ionize the air at the sensor. This source is as easy to acquire today as it was then. Ionization type smoke detectors are quite common and each contains a radioactive source for the same purpose. Ionization type smoke detectors measure current flow across a small ionized air gap.

Nick

[HELModels]

Nick,

Do you happen to know the name of the marketed RC version? Do you know the patent number?

Never mind if you dont, I found article titles in AMA hall of fame.

[nfhill]

I think the patent is #3,873,050 on March 25 1975. Maynard Hill has a number of patents on similar technologies.

I don't remember any details on the product itself. My memory that there was an actual product is not entirely certain. I've been expecting to be corrected on that by some other old timer.

Nick

[kcaldwel]

I think the patent is #3,873,050 on March 25 1975. Maynard Hill has a number of patents on similar technologies.

I don't remember any details on the product itself. My memory that there was an actual product is not entirely certain. I've been expecting to be corrected on that by some other old timer.

Nick

I don't remember it as a product, although I do remember it in the magazines of the day.

Here are some schematics, and letter from Maynard that seems to indicate he didn't think it was stable enough for a commercial product. He did use a piezo gyro in his trans-atlantic plane:

http://www.rcsoaring.com/rcsd/RCSD-1986-08.pdf

Kevin

Edit: For $0.50 you can get a copy of a published paper:

http://www.archivale.com/catalog/product_info.php?manufacturers_id=2&products_id=2362&osCsid=1d2dc90c6b44d98edea71042e434f568

Edit/edit: There was a semi-product, just not by Maynard. Page 8 here:

http://www.rcsoaring.com/rcsd/RCSD-1984-07.pdf

[HELModels]

Thank you both for that. I also found something on electrostatic meters. A really simple one is made from a FET, 9v source, with a bit of wire for an antenna, and an LED. No mention of an alpha source.

Edit: I read the patent and once I got past the patent lingo, the description of how it works is pretty good. The potential difference between the plane's ground and the voltage of the sensor is proportional to the bank or pitch angle. The Earth's static field is locally horizontal and below 3000 feet varies about 180 V/meter.

[HELModels]

I think the patent is #3,873,050 on March 25 1975. Maynard Hill has a number of patents on similar technologies.

I don't remember any details on the product itself. My memory that there was an actual product is not entirely certain. I've been expecting to be corrected on that by some other old timer.

Nick

The thing I find curious about this device is that there is a bunch of stuff(internet) about FETS, ISFETS, MOSFETS that need 9V as a source and use a piece of copper wire for the probe. They all claim to be very sensitive to the electrostatic gradient.

Is it necessary to have an alpha source as a probe?

[vintage1]

The thing I find curious about this device is that there is a bunch of stuff(internet) about FETS, ISFETS, MOSFETS that need 9V as a source and use a piece of copper wire for the probe. They all claim to be very sensitive to the electrostatic gradient.

Is it necessary to have an alpha source as a probe?

It probably was before FETS were around.

We used to really struggle to get anywhere near the sorts of input impedance - especially at DC - that a FET gives 'just like that'.

Its a shame they weren't invented first.