My altitude manager project

[Teej]

Hey folks.

Been working on my own altitude management system and after some very successful test flights, I thought I'd float this out there and see if there was any interest in me enhancing it for possible sale. At this moment, a number of parameters are hard-coded as I set out to build this for my own benefit as well as a school project. (Despite my age, I haven't finished my degree yet. I'm in pretty much my final class...)

Keep the following in mind:

1: This is primarily for purposes of aerial photography. It would be of limited benefit (at best) to a learning pilot.
2: The use of a stabilizer (ap2000i, cp2, HC) is expected but not explicitly required - I have done some of my testing flights with no active stabilizer. Without a stabilizer, the benefits are greatly diminished.

Background...

Several years ago ('06) I built a camera ship and was looking at the options for a stabilizer. I ended up not buying one then because I really wanted something to do altitude control as well. Then I ended up drifting away from helicopters for a couple of years. I got back to them over the winter and bought a CP2 among other goodies to help enhance my camera ship. Of course there was still no altitude management included. I started thinking about building one, then saw Pistaj's progress with his. I almost got into his testing program until I found out his device wouldn't work with my setup without either switching to a different stabilizer or throwing yet another gizmo in the works to handle the mixing functions - my test mule is a trex 500 and my intended camera ship is a converted raptor with a CCPM kit. Pistaj's unit only works with 1 channel for collective.

So as I applied for my class being faced with the prospect of doing a web page for my project, I submitted the idea of an altitude manager and they went for it. Although research had been ongoing, actual work on the project started the last couple of days of January. Out of curiosity, I first considered using GPS, but then found out just how slow/inaccurate GPS is with regards to altitude (at least at low end / consumer grade GPS). I then tried using inertial measurement. I had it set up well enough that measurement error was on the order of thousandths of a G (an inch or two per second)...until the motor started turning. Although I had isolated and regulated the power to the sensor very well, it looks like I was still getting some electrical noise as well as vibration making the resulting readings useless. I haven't given up on the idea of using inertial measurement somehow, but it's far on the back burner now in favor of the pretty much universal solution - barometric pressure.

Part of the reason I didn't go with barometric pressure initially is because the only electronic altimeter I had at the time was my Eagletree altimeter, set in its "3rd party" mode. A great little sensor they put out, although they only advertise +/- 4 feet of accuracy. Plenty good enough for its intended purposes, but not so great for my application. My own ET altimeter seems to vary more than that. Ah well. My project surely isn't what ET had in mind.

After a bit of testing, I built my own barometer module with a current production sensor that gives me much more consistent results. I had it "almost working" over the weekend, but there were a couple of small bugs left. Today's flights (in strong, gusty winds) were very successful. Gotta see what it's like on a calmer day - may need some additional tuning. But it's working quite well.

In its current form, it's capable of reading 5 radio inputs ( coll + 2 cyclic, on/off for alt management, + ??) and generating 5 outputs(ie collective , 2 cyclic, gyro sens, on/off for stabilizer). I can (and have tested) control the helicopter, stabilizer (CP2) and my altitude manager on 6 channels, meaning with my AR7000 receiver I can work the shutter as well. Mostly for my own benefit I've got an eeprom on board for datalogging and outputting to the PC over serial. I have a very limited interface (just some runtime parameters) displaying on a Crystalfontz screen at the moment.

I could easily alter the software to only use 1 channel (collective) for altitude management. This would allow using the other 2 outputs as additional servo signal generators (governor, etc).

Any interest in me writting a user interface for this thing and putting it up for sale?

Teej

[nightflyr]

Sounds like an interesting project

[Teej]

Thanks. It's been a blast working on this. When I presented it for school I didn't know if I'd actually be able to pull it off in time. Quite a blast when today it worked flawlessly over ~ 7-8 batteries.

My Trex 500 has made an interesting test mule. Really likes to get taken by the wind. Hehehehe.

[nightflyr]

have any photo's of it??

[pril250]

Sounds interesting an exciting project indeed, will be keeping this channel open to follow progress and to see more of the final product in flight, size & weight, ease of installation & setup and control.

[Teej]

Well, at the moment it's very much a prototype. It's a circuit board ziptied to the side of my Trex 500 and made mostly of through-hole components. My pressure sensor is ziptied to the bottom of the rex. The unit is connected to the receiver via servo type wires. The CP2 is plugged into my board as if it were the receiver. Power comes from the receiver and is cleaned up on board.

Certainly if I expected to sell it, I'd have new boards cut for me and use more SMT to shrink it further, plus consider bringing the pressure sensor on board instead of attached separately (although being separate makes it easier to shield from rotor wash)

Weight is certainly not an issue. The 2 boards (main plus sensor) and wires to hook it up...I haven't weighed it but I'd guess it all weighs less than 1 AA battery.

The prototype can be seen here, surrounded by some EagleTree sensors, AR7000 and an accelerometer that I was trying to use and a compactflash memory card for size reference.

Apologies for this being an ugly graph, but here's a graph from some of today's data. Calibrated on left, data in pink is the altitude in feet above (programmed) sea level. On the right and in blue is the position of my switch - "1" means I had the CP2 on. When the line is at "2" the altitude manager was on.

No, the line isn't perfectly smooth at altitude because:

1: Sensors have noise. You can only filter so much before the data is useless.

2: It was windy. The heli was not held in place, it was "flown around" in forward, backward and sideways flight, pirouetted, etc. Extra airflow of course causes altitude changes, as do sudden changes in cyclic. What I am working on now is the balance between maintaining a specific altitude and making a "rough ride" for the camera. The movements here are very gentle. A few days ago it was set more aggressively and you could see the heli pump up and down. Heheheh.

3: I readily admit there's still some optimizing to do as alluded to in #2. This thing "works" but it's not "finished".

That said, the data you see where it's fully engaged, the variations are primarily within +/- 5 feet. The full peaks and valleys touch +/- 10 feet....and this is data from the onboard sensor, noise and all. The graph represents a flight of over 5 minutes. The first activation represents about 3 minutes of altitude hold. The second, higher hold was for 1-1.5 minutes.

[pril250]

Yep, see what you mean. We are patient people, wait for further progress.

[nightflyr]

Good lay out, minimizing it should be easy, and I would agree about the sensor being separate for mounting.

[Trooper]

Originally Posted by pril250

Sounds interesting an exciting project indeed, will be keeping this channel open to follow progress and to see more of the final product in flight, size & weight, ease of installation & setup and control.

+1 and of course cost.

Ian

[Teej]

Didn't realize how much fun this would be once it worked. Heh.

Hit the switch at 133 feet in a fairly strong climb (about 15 feet / second). Went up a bit further, then settled back down to 133 for several minutes. Played around there for some random shots.

Certainly not appropriate to fly a heli / mission you otherwise couldn't fly but sure makes it easier to park the camera up top!

[nightflyr]

Originally Posted by Teej

Didn't realize how much fun this would be once it worked. Heh.

Hit the switch at 133 feet in a fairly strong climb (about 15 feet / second). Went up a bit further, then settled back down to 133 for several minutes. Played around there for some random shots.

Certainly not appropriate to fly a heli / mission you otherwise couldn't fly but sure makes it easier to park the camera up top!

That would be the whole point of the system

[Teej]

A little low-altitude demo from this morning:

My device is switched on at 1:27 and off at 5:27. Run continuously for that 4 minutes. Pretty good lurch when shut off because I made sure I had some up collective set when I cut the system.



During the day I made a few tweaks to the system and was mostly doing higher altitude tests (120-130 feet)