These are our experiments with radio control aircraft, most of them homebuit and more than a little bit crazy.

Click on any pictures here for a larger version.

New motor!
IFO Super
IFO Succeed
IFO Fail
Scratch 75 wing
Not as light plane
Lighter plane, take 2
Lighter plane
Tiny motor go bye-bye
In-flight camera
New toys
Back in the saddle
Building a wing
Them's not ailerons!
The kindness of strangers
Wally wing?
Still gaga for brushless
Love this little LensRC motor
Good plane! Hard ground!
Remember what I said about basics?
Lots of gliding
New wings, glide test
Back in the air
Back to basics
Glider musings
Electric musings
Great flying day, but windy
Successful gas flights with a simple "stick plane" design
Messing around with electric parking lot flight
First stick plane ala Zack
First gas flights, mixed success
October 10, 2004

A few months ago Zack and I decided to build an airplane out of one of those cheap styrofoam gliders. So we finally got one, a Cox .049 2-stroke engine, and a radio/servo kit leftover from a previous (totaled) model.

The first couple nights we attached very basic balsa control surfaces:

We threw that around a parking lot until we were comfortable with the controls and balance issues.

Then we added a little electric prop:

It was pretty disappointing, it was geared down for a much lighter aircraft, and it would only extend the glide path a little. We also broke one of the gears when we rammed it into the side of a church, whoops.

Then we attached the Cox .049, but we were still doing parking lot flying, so it was just a test weight at first:

We snapped that plane in half (that's why we used the cheap foam gliders), so I put the controls on a smaller plane:

It flew like a brick, but the control surfaces were very powerful. I want to go back to that, try it with a motor on it, and see if it's actually fun, if a bit touchy..

We went out to Zack dad's farm so we could fly with the Cox engine. First order of business was learning how to start the Cox engine -- we'd had some disappointing experiments already:

But we got the hang of it and Zack and Kava became an unstoppable starter team. You can't see the blades in this pic because they're moving:

This is our flight feild:

And this is how our first flight ended:

Zack taped a balsa splint on and the plane had two successful flights after that. It proved to be a very capable plane, very responsive, very overpowered, capable of flying into the mild wind we had, turning aggressively, and doing loops without any difficulty whatsoever. I turned out to be a fairly sketchy pilot, easily confused about which direction the plane was pointing, but it was still one of those magical experiences. Kava proved to be an able climber, but we eventually had to leave the plane in a tree overnight until we figured out how to get it out (32ft fiberglass ladder, 20ft metal pipe).

Anyone know where to get a cheap strong 40ft+ collapsable fiberglass prod?

October 11, 2004

Zack started putting wings on sticks and came up with this model:

October 17, 2004

Last week was one of those weeks where you discover that a single charge on the nicads can last through 3 models, because the models are so crapulent that they get scrapped (or scrap themselves) almost immediately. The lessons learned:

But this design proved quite stable, once we gave up on the ailerons. We just need a bigger better motor (gas) and a field and it will no doubt be quite the sleek flyer:

After a number of landings (it turns out the tail contacts first, despite our best efforts):

October 24, 2004

We had a great set of flights today. While we were throwing around some of our more crazy experimental aircraft in the parking lot across from my apartment a while back, one of my neighbors walked up. His name is Matt and he is an airplane mechanic at the local airport. And he's building a really cool scale biplane. So today when we are burning in our new motor he shows up and decides to go out to the farm with us. He coached us and, following his example, we had about 5 successful flights, with very few amazing crashes.

First, here's an experimental "front rudder" or "keel" we tried to prevent it from drifting to the side:

Here's another experimental stage. We built a huge flat wing out of our first set of styrofoam wings. It appear to be indestructible, so we put the first tail we found on the floor on it. It wasn't nearly powerful enough to control it, but it was neat to have such a light and bare model as a proof of concept:

At the end of last night's experimentation, we came up with this plane, shown here with an electric motor for parking lot flying:

It was mostly uneventful to outfit that thing with a gas motor and throw it at the field. The first 3 flights were exciting (and nearly identical). It rolled to the left...hard, so it followed an arc up, then smashed exactly nose-first into the ground. No damage, luckily. It was kind of distressing when we discovered we'd just hooked the rudder control up backwards, so when the pilot tried to correct for the roll, he just made it worse.

Here it is about ready to take off at the field today:

Being thrown:

In flight:
zoomed in

We got some VIDEO of that flight. Click here for video.

The first flights were all with some relatively high quality wings we'd purchased from the hobby shop that had been intended as replacements for a cheap kit. They were very light and also about as small as we really wanted to do with this motor. After we did a little bit of damage (repairable) to their leading edge in a tree landing, we decided to go ahead and throw on the flat wings we had assembled the night before. They seemed to fly pretty well, but the pilot lost orientation (or maybe one of the control lines snagged on one of several protrusions?) and the plane did a nose dive from above the tree line. The motor was burried two inches deep into the dirt:

We had previously considered that wing unbreakable:

It was getting dark and it would have taken a good 15 minutes to repair that damage, so we went ahead and called it a night.

The biggest things we learned:

October 31, 2004

Today Zack, Matt, and I went out to Steve's farm to fly the same basic plane we flew last weekend. It was a beautiful day, but a little windy (pretty consistent wind though). Our first flight we had Matt on the controls to test out our new ailerons. Zack had glued a strip of balsa onto a bamboo rod which went through some straws down the length of the wing. Alligator clips attached to control rods clipped onto the bamboo rods at the center of the plane, twisting the rods and moving the aileron surfaces at the end of the wings.

The design worked great on the ground, but in air, the wings had way too much flex. In the wind, the wings were bending almost imperceptibly but the ailerons were fluttering constantly. The effects averaged out, so they were still useful as control surfaces, but it made for a sketchy flight. The exciting part was when the plane got into a bit of a spiral battling the wind with the crazy control surfaces, one of the wings snapped just past our balsa reinforcement rod (nothing else broke in the landing though). We're going to have to do something to reinforce these previously-unbreakable foam wings.

We flew Steve's simple electric -- it is one of those two-channel (turning and throttle) V-tail trainer kits. Very simple, very light, not really meant for windy conditions. It actually held its own just fine though. I've been having a lot of trouble getting good performance out of my electric setup, but that tiny electric plane was plenty powerful. Once it was up there, our number one problem was that the control surfaces didn't work very well when the throttle was down, and when the throttle was up it was climbing like crazy. Would have been a very nice glider if it had had some sort of pitch control (elevator). I think I'm going to have to try making a plane just a little bit bigger than that and flying it like a glider.

Then we put the smaller (about 3 foot, instead of the 4 foot we'd been using) wing on the gas plane just to see if it'd work. As expected it went a lot faster and actually handled just fine. Maybe was even a little bit less a victim of the wind. The large wing on the electric plane, however, did not work. For some reason it made the plane roll like crazy, even though we had the wing in the center.

It's a real pity I forgot to bring my camera, because in the wind we had lots of great very-low-speed (or actually hovering with the electric) flights that would have made great photos.

We all had a bit of time at the controls and had some great flights. Only two tree landings and fairly minor damage to hardware.

Eventually had to stop flying because the tank started leaking like crazy (I think a seam was just being pulled open by a bent aluminum mounting bracket). Also some moron glued the engine upside down after it broke off.

Steve let us borrow an ancient .35 (we have been using .049) engine. It is extremely enormously fatally overpowered for the planes we've been flying, and it has absolutely no throttle control. I'm not sure there's any way to use it without killing ourselves, but I think the best bet is to put a huge piece of blue insulating foam on it -- something that (by all reason) shouldn't be able to fly. The only question is how to reinforce it (glue wood strips to it?) so it doesn't break up right away.

December 4, 2004

I've been putting a half-assed effort towards putting together a very light electric plane. I've got a joke of a motor with a small prop that looks like it was designed for rubber power, and a too heavy NiCad battery pack and some three and a half foot wings (about 1.3 sq ft), and (for now) rudder/elevator control. It flew level for a while, but rudder impulses were so powerful they rotated the entire body of the plane relative to the wing (which was held on by a single rubber band). I really like the idea of using a totally underpowered super light and tiny airplane. I built a primitive balance and it seems to weigh less than 9oz, making for a very light wing loading (about 7 oz per sq ft). Most sailplanes have a much wider wingspan (for a better "aspect ratio" -- thinner wings) but also weigh much more.

Here is my light electric beside its wing (it broke the rubber band)...not sure what happened to the elevator.

I'd like to eventually put a refit brushless CD-ROM motor and a couple of lighter Li-ion batteries, for a totally ridiculous amount of power in the absolute lightest package possible. But to do that I want to build my own brushless motor controller so I can get a better understanding of what's going on, and that will take time and effort.

I also bought kind of a flying "zip zap" (the cheapest smallest r/c car ever). It has barely a foot wingspan, and is made out of mostly a single piece of very durable foam. It has two tiny motors and steers like a tank (putting the stick to the side cuts one of the motors). It was amusing for about two hours.

We broke the props, so Zack fashioned some balsa ones that almost worked:

Here's the plane:

I've also been assembling my kit plane with Zack. It is a few carbon fiber rods connected with kevlar and cyanoacrylate (super glue) and shaped basically like a kite. It will be covered and fly like a big flying wing with huge elevons. It's gonna weigh about 8oz and with the huge wing surface it should be pretty fun.

Here's the beginning of the assembly, most of the carbon fiber frame:

My dad explained how the signal from the radio works -- it is essentially a pulse train (encoded in an FM signal), one pulse per servo, and the length of the pulse determines the servo position, between about 0.5 and 1.5ms, with a longer pause before the first servo to keep it in sync. I was thinking maybe I could hook up a microcontroller to my radio and give it computerized mixing functionalities, and ghost channels (I could get a 6 channel receiver and have the computer control flaps from a switch or something). With just a little fiddling with my oscilloscope and a websearch it was easy to see the pulse train being output at 0/+10V levels from the trainer port on my transmitter. Hooked a resistor from the output to its input, and flicked the trainer switch and it picked up its own signal -- should be really easy to work with. The trainer connector even gives me +10V from the batteries for power.

December 11, 2004

I did some reading online and it looks like the magic wing loading for small (1-meter) sailplanes is in the 3-5oz/sqft range. I've got these 43" wings I've been using that weigh by themselves about 2.5oz and are about 1.3sqft. So I put together the lightest flight pack out of the equipment I have.. 2x HiTec HS-55 servos (rudder / elevator), 4-channel HiTec feather receiver, and I built a battery pack out of 4x 180mAh nicads I had (4.8V) sitting around for a computer project.

Total weight: about 5oz
Wing loading: 3.9oz/sqft

This is the 1.2oz battery pack I made:

Zack came over and we gave it a toss once the battery had charged a bit. It was perhaps promising but on impact the front half of the stick snapped clean off. We were trying a rubber band impromptu way to reattach it when the back half of the stick snapped clean off. What happened was we built a plane out of a 3/4" square stick, and it was indestructible, so we built one out of a 1/2" square stick, and it at least held up to a couple glide tests, so I built one out of 1/4" square balsa, and it snapped on its first flight. There's optimism for you.

But we were undaunted and used rubberband hacks to kind of hold it together. It was a joke in flight, whenever it would jolt at all, the rubber bands would let the stick with the battery on it swing far off to the side. But eventually we just didn't care and Zack tried a frisbee launch that was actually a lot of fun...I could detect that we had control. Not enough to do anything with, at least not with my reflexes, but it was neat to think of it as just kind of a variable aerodynamic object even though it was operating well outside of conventional straight-line flight.

But that could only be fun for about 5 minutes, so we went back inside and thought a bit. Zack wanted to use a big light piece of foam as a fuz, so he was demonstrating how strong it was when it broke in two. So we just reached in the closet for more foam. Zack picked up like twenty of these 1'x4' pieces of posterboard-coated foam from a copy shop's dumpster. Not really appropriate for a fuz, but we've been wanting to make a flying wing for a while. Once we got real drunk and gave one very inexact polyhedral and put a weight in the nose (jammed my 'helping hand's alligator clips straight through it), and it was a grand old time to toss around the yard.

So this is what we came up with... We never decided if it was meant to fly "inverted" or not, so... Let's call this the bottom:

And let's call this the top:

Note the tiny tiny elevons on this version. We'd never flown any sort of elevon or flying wing airplane before, so this was a big experiment.

When I hooked up an elevon mixer to the receiver, I found the same bizarre problem that had haunted us earlier...when I plugged it in, suddenly the receiver stopped working, and even when I unplugged it the receiver wouldn't drive a servo directly. I was very puzzled by this, but eventually figured it out: every time I hook up the elevon mixer I stupidly put it on wrong the receiver channel (basically I turn the thing upside down), and from then I'm totally confused. Way dumb, eh?

So after a couple flights with absolutely no pitch stability, even after we hung a 2oz pocket knife (because it had a convenient belt clip) on the front of it, we decided something had to change. We noticed the elevons could actually fight some of the poor pitch behaviour, so bigger elevons was our first decision:

I weighed a glue stick and it's only about half an ounce so we sunk about three into this aircraft. It made very strong joints for those vertical stabilizers. We also found that the cloth medical tape we were using for elevon hinges really soaks up hot melt glue well and sticks to the posterboard great, so Zack put it all over the tape -- these may be our strongest control surfaces ever. You can really see how it's piled on in the above picture.

We also had had luck earlier taking just a normal reasonably-sized pair of wings and adding just a long strip of balsa hanging off the back of them. Just throwing the wing around, it seemed pretty stable. So we also did the same thing here, giving it a bit of a tail:

The end result wasn't really flyable. Part of the trouble, though, is that my tiny battery pack never got a chance to charge, so it gave out shortly after we went outside. The bigger batteries, though the weight wasn't really a problem, didn't get the real tape treatment they required.

Overall our experiences with this flying wing reminds me of a couple months ago when we were first putting wings on sticks, before we really started caring too much about center of gravity and stuff. I don't know any rules of thumb for flying wings, and it's not quite as simple as with the smaller wings we've used. They are as likely to spin up as to spin down, it's kind of obnoxious. We'll have to try again tomorrow.

December 19, 2004

Today Zack said, "do we have any more of those foam gliders?" So I just put some control surfaces on one.

Really, we've done exactly this before, only the mindset has changed. The intention is different, this one isn't intended to have a motor. Which doesn't really change anything because we threw around all of the motor-destined ones without a motor first just to see what we were doing. The real change is the factors that we are now aware of.

So here I have collected some specs about the airplane. Pretend it's a cool physics class project. Its got wings with "aggressive" dihedral (I don't have a protractor) and aggressive delta or back sweep (which has basically the same effect), so it is very stable. The center of gravity when it is assembled unmodified is about one centimeter from the back of the wings where they touch the airplane. The wings have a "root chord" (from the leading edge to the trailing edge of the wing where it touches the fuselage) of 7 inches, and are tapered to a tip chord of 3.5 inches. The wings are about 26" from fuselage to tip, for a total area of about 136 sq in per wing, or about 1.9 sq ft overall. They are cambered, but not as severely as the wings I have been using (those are actually concave on the bottom, this is just flat).

I plugged in the root and tip chords into a formula and found the "mean aerodynamic chord" to be about 5.4 inches, which I think means that the wing acts overall about the same as a rectangular 5.4"x26" wing (plus whatever advantages tapering gives us), which brings us up to 1.95 sq ft. Not a big difference.


4.5ozfoam glider
0.5ozhitec 4-channel feather receiver
0.5oz x 2HS-55 servos (rudder/elevator)
1oz180mAh 4.8V NiCad battery
0.5ozleftover (tape, balsa, wires, superglue)

Not quite as light as the balsa toothpick we broke last weekend, but it has bigger wings, so it is still only about 4oz/sq ft, which is close enough to the target 3oz/sq ft. To maintain the existing center of gravity, I have about 1.5oz of electronics at the tail of the plane, so I moved the 1oz battery forward until it balanced about where it should.

The greatest part about building this was we already learned from all the old mistakes with these foam planes. We already know the best way to mount servos on the tail and stuff like that, so this is all pretty much "on the first try." There's no residue from a dozen different things being taped to this plane and taken off, no huge glue puddles or useless rubber bands scattered all over it.

So I put it all together, trimmed it in my livingroom, and gave it a toss:

(yeah it can meet the wall without damage)

It has one weird characteristic I can feel just playing with it. When the elevator is full down, if I release the stick and allow it to return to center immediately, it makes the whole plane vibrate just a little bit. A lot more than any other control surface I've ever seen, but probably not nearly enough to matter. I guess something bends a little bit when the servo is at full throw?

I took it out to the church parking lot and gave it about a dozen tosses. It's really fun, as gliders go. On the third toss (just a hand toss from ground level, with one hand on the radio) I was able to get it pretty much all the way across the parking lot, rivalling our previous best that was achieved with one guy standing on top of a big concrete sign about 12 feet in the air. Due to the innate stability of these foam planes (and perhaps due to some reflexes I'm slowly building), every single landing was a perfect level belly landing, even after I started just tossing it twenty feet straight up to give it a little more range, it never felt out of control. Very predictable and controllable.

So tomorrow I think I'm gonna launch it at a nearby park with a huge hill, see if I can catch some thermals (hah right). As I see it, in terms of most specifications, this plane is about the same as a lot of the professionally produced gliders in this class, so I expect that it should be capable of some fun flight characteristics. I don't know much about aerodynamics, though, it might have excessive drag or something. I do know that the "aspect ratio" (span / chord) of the wings is about 10, while good glider wings have an aspect ratio more like 20, which results in a better lift/drag ratio. On the other hand, the aspect ratio effects are not linear, so 20 is a lot closer to 10 than 10 is to 3 (it makes sense if you look at some graphs).

Well we launched it, and we found that this foam gets brittle when it's cold outside:

March 7, 2005

It's finally warm enough to think about going outside, so yesterday we put back together our basic stick plane and gave it a couple tosses. But we forgot the real reason that we don't fly much during winter: wind. The plane was perhaps a bit underpowered after the addition of a fuel tank, but the wind was fatal. So without further adieu, enjoy a video (8MB).

March 12, 2005

Today Zack reapproached making a straight wing out of the foam gliders (which, in their natural state, have really severe dihedral and back sweep). The one we built in the past worked alright but was a little bit too flexible for the ailerons we had and dramatically ripped themselves apart in midair. The foam is very strong and the wood and epoxy that we used in the center part to hold the two halves together were also strong, but the joint where the inflexible wood ment the very flexible foam went the way of the dodo. So this time he ran a balsa spar almost the full length of the wing, and we used the less brittle hot melt glue instead of epoxy.

He also used a curious wing mounting technique for another one of our stick planes (this time, a glider, but we'll probably add a .049 to it tomorrow):

I was skeptical of the tiny control surfaces, but the elevator actually proved plenty powerful (probably because it had such a long lever arm). The rudder was almost but not quite useless though -- it probably didn't have as clear a 'pivot point' to take advantage of the lever arm.

We threw that around the parking lot a bit and even though there was a light wind, it was obviously a pretty capable aircraft. Finally an inverted landing took off the rudder.

In the process of putting this together, I finally broke a servo gear. One of the tiny HS-55 servos was sounding a little louder than usual last weekend, and this weekend when I was trying to gently move it to a different position so I could get access to something, and it jammed. Moments later the servo arm was spinning free of the motor. We took it apart and determined exactly which gear it was -- only one tooth was gone, but it didn't matter, the moment the servo got to that gear, we lost control of the arm. Even if we didn't need control in that range, every visit to that spot on the gear destroyed the trim. I just ordered a replacement set of gears, it's only $5.

I also ordered a Magnum XL-15 engine, which is about 3x as big as what we've been using (but only about 6oz, still). All together, this will probably more than double the weight of our basic stick plane, but it should provide enough extra power that I'm frankly not too worried about that. But it will probably fly a lot faster (unless we start building bigger wings), so it should be exciting.

When I was cleaning up I realized what I'd turned the corner of the bedroom into:

I clearly have the bestest girlfriend ever.

March 13, 2005

Today we took that same plane from yesterday and added servos inside of the wings (just cut a hole straight through the foam) to control ailerons. We also left off the rudder, figuring ailerons would do us. On the bench it worked great and trimmed out well, but when we put it to work in glide tests we ran into a few problems.

The first obvious problem is that I used the battery extension cable to make the aileron Y cable, so we had to put the battery under the wing, and almost an ounce of ballast in the front to get the center of gravity about right. This plus the extra servo added enough weight that the thing didn't glide long enough to get a satisfactory understanding of its flight characteristics.

The second obvious problem is a little more obvious in hindsight but was quite astonishing in person. The thing had a tendency towards extreme instability. Even moreso than our other no-dihedral projects. It would go into a flat spin at the slightest provocation. Duh, when we decided to leave off the rudder nobody ever bothered to reattach the fixed vertical stabilizer! We ran back inside and glued on some sheet balsa real quick, and it improved, but the whole thing was still a little unstable, and the wind, and etc., and I did a knife edge landing and managed to break off the last 9 inches of one of the wings.

A little discouraged that we'd almost demolished our plane during simple glide tests, we switched for the moment back to old faithful: walmart-style foam gliders (we actually buy ours from The Treasure Chest, a wonderful family-owned hobby shop downtown). I had one sitting around pretty much built (not that that takes much), so we just slapped some servos on it and it was ready to fly in about 15 minutes. Instead of the typical plastic U thing to hold the wing on, we used just two rubber bands, and that actually worked out wonderfully.

So we took it to the local park with the big hill and had a great time with it. The winds were unpredictable but mild enough that it mostly didn't matter. The only control issue was that the rudder didn't have quite the assertiveness that I wanted when messing with a strong wind. A couple times we caught the wind just right and actually rode it for maybe 5-10 feet of lift, but not enough to convert that into speed to do a 180 and ride it more. You sure get a lot of exercise running up and down a hill chasing a glider!

The process of learning is quite pleasant. Zack was throwing the glider up the hill towards me, and right before he tossed it I tried to work it out in my head "it's coming towards you, your turning is reversed" but it actually messed me up, already my idea in my head was naturally reversed because I was thinking "in the pilot seat." Now that I've spent some time with an r/c simulator, even after years of normal simulators, r/c perspective feels much more natural, especially when you are playing with ground effects and tight turns and slope drafts. Much easier to keep your relative-to-ground orientation from outside! Another great learning experience was knowing when we're in lift. I'd read from some thermaling pages that you very quickly gain an idea for whether or not your plane is in lift. Even from hours of sims I hadn't really gotten that reflexive feel, but watching the plane on the hill, you could really tell when you got uncommanded lift.

One thing that was a bit of a disappointment is the human eye. I always have a lot of depth perception difficulty in the sims, which is natural for a low quality 2-D projection of 3-D reality. But we were trying to fly the plane into a shelter at the park (some distance away from us) and both Zack and I thought that it had pretty much made it, so I stalled the airplane to avoid ramming the picnic tables. When the plane hit ground, we were a good ten feet away from the shelter. I guess when the plane is 190ft away and the shelter is 200ft away, it's hard to tell the difference.

Another interesting thing I noticed is how different flying personalities are. I mean there are some obvious differences from experiences, but I get the impression that I like the elevator trimmed substantially more "stick back" than most people (at least on gliders) -- definitely moreso than Zack, and moreso than the default on the sims I play. My reasoning is that I want it to go pretty much level even as it loses speed, so that the decision to trade altitude for speed is a conscious decision, so I get speed in bursts for maneuvering and stuff. I think the status quo opinion, though, is that a glider should be trimmed to have a glide path that pretty much maintains speed and slowly loses altitude. *shrug*

We had enough fun that the plan is to put together a second foam glider (we have all the parts here), slap some LEDs on it, and fly some duels tonight maybe with a camcorder.

Click here for some video (QuickTime, 2.5MB) of some relatively successful glides down the hill.

I had some really weird servo behaviour on the second glider. Click here for video of it (QuickTime, 1.6MB). I think the problem was that there is a break in the signal wire (verified with voltmeter). Because the servo control signal is, in effect, a waveform (a square wave with variable duty cycle), the signal could still sort of make it through using capacitive coupling, but not very effectively. Specifically, releasing the servo from either extreme tended to cause it to carry on going back and forth uncommanded for a while. I just put the gears from this servo with the electronics of the one with the broken gears.

We also did a feasibility test on rocket power for takeoff (to get altitude). Click here for video (QuickTime, 1.5MB). I think it could work but we haven't figured out a good way to trigger the rocket in midair, and nobody has volunteered to hold it for a launch. Launching it along a guide post has been considered, but I bet it would be more efficient if we didn't use the rocket for the initial start.

Finally, the guy who did most of the taping (Kava) turned out to be an amusing cinematographer. Click here for video (QuickTime, 8.1MB).

Here's the two assembled gliders with lights:

March 18, 2005

Last night I was describing to Matt the problems that Zack and I were having with this foam glider. See, the hobby store now carries 3 varieties of foam gliders: swept back super-dihedral, straight mini-dihedral, and totally flat. On Monday we went up to the park with the hill with a swept back super-dihedral and a totally flat one. We found that the flat one had GREAT glide properties, but the slightest perturbation from exactly level flight yielded a dramatic instabilities. I'm talking about things like tail first dives. Zack finally got it trimmed out to where it was flyable, but even so (in the wind we were experiencing) there were frequent inverted landings and spectacular cartwheels. On the other hand it was the only glider to be able to do a full loop from a hand toss.

Now, I'm sure anyone who knows anything about planes, or perhaps anyone who was paying attention to what we were doing back in October could recognize right off what was going on here. And it became really obvious to me as I was explaining it to Matt. A quick check confirmed that indeed, the CG was maybe half an inch behind the center of the wing. Moving the battery forward to the very tip moved it about an inch ahead of the wing center. We took that out in the parking lot and gave it some tosses, and *boy* was that thing stable!

I guess that's just life, being reminded again and again of the lessons you apparently didn't really learn the first time.

Matt also suggested a direction for these foam gliders that I think will really make them into serious airplanes. All we have to do is coat one in fiberglass. 3 or 4 coats of some fiberglass and the foam planes will be stiff and probably just as resillient. Probably something to do once it gets warmer outside, because I don't really want curing resin inside and it prefers to cure warm.

March 19, 2005

My brushless motor finally arrived from lensrc. I'm not sure about the details, but I think Leonard basically takes the stator and can from a CD ROM motor, adds a high quality bearing assembly, and rewinds the coils for maximum performance. The end result is a low-cost low-weight efficient little motor.

Yesterday we did a first trial of fiberglassing a foam glider, using 1.5oz/yd cloth, and some 20 minute epoxy. It gave us impressive stiffness and strength, but it added some weight too. I think it allows these foam toys to act like a 'real airplane' but I think that also demands some 'real power' to haul it through the sky.

So today after the fiberglass cured, we put on some control surfaces and the new brushless. Total weight is about 15.5oz. Plane:

You can see some ripples in the epoxy on one wing:

And here is the motor (with a somewhat dubius prop attachment):

We flew it despite the wind (when in doubt you can always substitute optimism for experience when making decisions such as these). There were a few disadvantageous things going on. One was that the peak cruising speed of this plane was probably substantially below wind speed, and the wind was a little gusty too. Second was that we didn't seem to be getting quite as much power as expected out of the motor, but I think this could be because I'm using an undersized nicad pack (the li-polies are still in the mail). Also it's hard to judge power when your plane is being tossed about willy nilly. I think it would have been able to keep us in the air on a calm day. But the problem I'm actually more concerned about (and which we won't really be able to even definitively diagnose until a calm day), is that it seemed to be weak at the controls. My suspicion is that the wind was too much for these HS-55 servos and it might be time to start using bigger ones (maybe the HS-81s?).

End of the story is that we managed to break the fuselage in two. I did a few more wraps of fiber around the break, and I think it will be fine tomorrow.

I'm starting to develop an opinion that the number one goal should be weight, more than even strength. The majority of our breakups happen on impact, where the weight causes the destruction. It sucks to have your wings flapping around like a toy, but for the type of flying I want to do for the moment, it shouldn't matter (I'm not in search of a sport aircraft). I'm inclined to next fly essentially the same plane, but without any covering.

I'm also getting to have more of a 'no-wind policy' towards flying. I've been building and flying aircraft that are simply too small for heavy wind. I like the 'park flyer' aspect, so I want to keep my aircraft small, so I guess that leads to this no-wind compromise. Even the essentially indestructible super-dihedral gliders start to sustain inverted landing- and cartwheel-related damage on windy days. In general it seems that it's less windy in the mornings here, probably because the sun is responsible for most wind. I'm curious about what parts of the night may also be good.

Also, I was stupid to only buy three props. If I had a bunch more of these props, I would try again tomorrow without hesitation, but as it is, I think I will wait until I get the better battery. I should perform research about which prop alternatives would work. The ones I have are super-light (and also delicate). I think I read something suggesting this motor doesn't like heavy props, but maybe only when the braking feature is enabled in the controll.r

March 20, 2005

Today we put together and flew the stick plane with the .15 motor we just got. One of the major improvements was a good quality stiff foam wing with ailerons. It has (among other features) a carbon fiber spar of some sort (arrow shaft?). This was loaned to the cause by Eric.

So all together we actually used all 4 channels with servos on each one. Our most laiden beast. Though I swore not to touch the rudder so as not to confuse myself. Picture:

Other angle:

We'd flown almost this same plane in the past with a .049 engine and found it to be relatively tame (though not capable of handling much of a breeze), which is how we wanted it. But the .049s aren't throttled so the upgrade was necessitated. The .15. Yow! Being nowhere near underpowered, combined with ailerons and a freakish calm part of the day made this a really satisfying flying experience, even though I did eventually run it into the ground. It was going too fast, so I pretty much pulled the throttle all the way back, and then it wasn't going fast enough, so I slammed it all the way forward (see the problem). From then on, it was just bouncing from one near crisis to another. A very assertive flyer! The elevator that was pretty decent at low speeds was downright terrifying at high. I had to line up every turn precisely before pulling back, because the moment I pulled back, it was like an immediate snap turn.

I am to this moment a little mystified how I managed to get it into the ground. The video (QuickTime, 10.5MB) isn't terribly clear on the subject, but there is reason to believe that it was upside down at some point, and that may have confused me. But the way Kava described it really kind of sums it up: I was going too fast, and I was going too close to the ground. Together those have an obvious extreme endpoint, especially with anything less than perfect reflexes. I think I turned into a turn that I would have wanted to straighten up, and might have done so even enough to actually make it so pulling back sent it further down.

The crash was not so bad. The plane came apart into battery, motor, wing, and fuselage (with tail still attached). The joy of thickish balsa is that it doesn't break, the CA joints break. The aileron servo is a loss (might be able to just get new gears again), and our cheap AAA receiver pack got all the wires ripped off. The prop is of course history, and you can probably tell how fast it hit the ground by how much longer one blade is than the other (presumably, it hit first, then 180 degrees later, the other blade hit and got shortened even more). Would rather have not crashed it and maybe flown again, but overall a very good experience. The plane performed perfectly.

We threw together a glider with some lights for some late night fun. It turns out that the line where it is too cold for EPP gliders (makes them way too brittle) is very sudden. We've flown these things lots of times when it was cold, but last night was just a little bit colder, and the thing was breaking all over. Ran out of tape so we just rubber-banded the wings onto the wing stubs after they broke off. Until the fuselage snapped in two, it actually flew sort of alright, even when only one wing was broken. Final state:

March 25, 2005

I received those li-poly batteries a few days ago and finally I had the time to play with them at the same time as the weather was accomodating (not raining *too hard* and not *too much* wind).

So I put the 1200mAh, 1.8oz, 2S1P (7.4V) battery on the foam-plus-fiberglass that we flew the first test flights with the brushless with. It was too tail-heavy from all the fiberglass I used to patch the tail boom that we had to add 3 lighters on the front (universal dead weight from the coffee table). We gave it a couple tosses and found that it went to the right almost uncontrollably and overall wasn't terribly side-to-side stable (it rolled easily, and didn't tend to right itself). When we put the motor on, we had it pointing to the right (to counteract prop twist), so we just ran new sticks in pointing dead ahead.

I flew it like that for a while and it was kind of fun, but it was totally roll-unstable.. Getting it upright was hard, but once it was there, you could pick which way you wanted to roll. But I noticed that getting out of left-hand turns was harder than getting out of right-hand ones and eventually got it doing about a 20 foot horizontal inside loop (knife edge), turning left. But I was giving it full right rudder. It wasn't losing any altitude, but I was pretty mystified by how stable it was in this one particular configuration. I eventually just pulled full back figuring it would destabilize the loop, which it did, but it turned it into more of a death spiral, so I cut the motor to save the prop. It immediately rolled back to level (right before hitting the ground *sigh*) once I cut the power. What a graphic demonstration of prop twist! In hindsight, if I'd cut the throttle or even probably tried to go left or down, it probably would have come out of it.

The fiberglass plane was a little heavy, and didn't have much dihedral, so we decided to use a stock super-dihedral swept-wing foam glider with the motor+elevator+rudder that we've gotten so good at slamming together. Those are always stable no matter what you do, we figured. Their wings just *flap* in vigorous maneuvers.

It worked out pretty much as we wanted, it was totally flyable. The plane was about 10oz. I was expecting a little bit more power out of my motor+prop, but it turned out to be a totally tame climber and we had to keep it at or near full throttle most of the time to not lose altitude. I suspect I should be using a bigger prop (I was using an 8x4, and then switched to a 7x3.5). But a tame climber is really what I wanted to fly, just something that would comfortably keep the thing in the air without going crazy, so I was happy with it. We didn't totally drain the battery (it's still at 8.1V...peak is 8.4V, cutoff is at 6V, and I think it has a relatively linear discharge curve), so all of the high-throttle stuff doesn't seem to have cost too much. I'm going to have to put an ammeter on the battery with various props to see if I'm drawing anywhere close to the 6A that lensrc says this motor can draw. I think I may need to get a 3-cell li-po pack.

No damage to the plane or the motor, but we are going through props like crazy (the 8x4 and 7x3.5 from GWS are very light-weight brittle delicate props). We threw on a Cox 6x3 that we'd used with a .049 for a while, and it actually worked pretty alright. It wasn't a real climber, but it held altitude almost.

I think the big lesson is that it is time to invest the time into some ailerons. Turning with the rudder basically sucks, especially if you're trying to stay in a small field like this neighborhood park.

I have my fiberglass plane sitting behind my chair, with a wing poking into my arm space. And I accidentally bumped my arm into it, and it didn't budge, it didn't bend, and it hurt my arm. Fiberglass rules!

March 26, 2005

I tried to get the rubber-power-oriented props on my LensRC motor again. I think I managed to balance two of them well enough, but they still wouldn't go over half throttle. They seem to work smoothly up to about half throttle, but then the motor starts missing and the motor controller turns off. Maybe I should fiddle with the advance settings on my controller. Probably they're just way too pitchy for high-RPM. They don't seem to generate a tremendous amount of pull at low RPMs, though, so I think I may abandon that idea. I should try the middle-sized props I guess.

I flew the rudder-elevator foam plane for a while at the park with a Cox nylon 6x3 (because it's the only prop the motor seems to like that I happen to have not broken yet). I glued on a prop saver and it seemed to mostly do its job (though in the end I let a kid fly it into some concrete and that broke the prop saver...the prop is fine though). With the 6x3 it is pretty much at the bare minimum to be able to gain altitude at all, but it does work. Really hard to recover from a near miss with the ground or a tree and regain altitude, especially because I was flying in such a small park on such a steep hill, I usually would have to turn and start gaining altitude right away. Again, this rudder-elevator thing sucks for tooling around in tight spaces.

It's amazing how much of a difference small details like tape can make. We'd been using medical tape for everything, but then I got some good electrical tape which really makes a very strong attachment if you can get a wrap all the way around something. So I electrical taped the motor bearing onto the motor mount stick, and the result was that every nose-down landing (*shrug*) it would break the stick, rather than the tape. We would just put in another set of sticks each flight. When I pulled the sticks out of the foam this morning, I found quite a few broken off inside it:
, so I decided to switch back to the medical tape, which has an anti-rip-stop property which makes it rip whenever the motor experiences trauma, rather than breaking anything else.

Anyways, the plane is spending the night in a tree. So I have performed some research (read a few threads on, and it appears that there are a few popular approaches for getting planes out of high spots in trees (defined as, where a ladder gets useless):

I really don't see how we as a species have allowed these hazards to navigation to dominate the surface of the planet so thoroughly!

March 27, 2005

I decided that I couldn't allow the loss of a plane to slow me down, at least not until I lose too many parts. So I decided to build a light-weight lazy-person's version of the Wally wing. The idea is to take just the wings from one of the foam gliders and put them together in a way not dissimilar from the Zagi or CombatWings aircraft. Ken uses spruce spars and a speed 400 motor and lots of care. I decided this needed to be Gregified.


Total flying weight of 4.5oz-5oz (depending on tape, quarters, etc.). Pretty nice for being about 1.5 square feet of wing area.



With the equipment cover removed:

With creator:

I'm still having a lot of trouble getting plane-like performance out of flying wings. I worked out the center of gravity for level flight by taping quarters onto the wing until a hand toss in my livingroom went level. Then once I got the controls on it, I tried that spot +/- about half an inch (by taping quarters onto it, again). I found that some spots were better than others, but I didn't find any spot where it would perform well at all at low speeds. It seemed like the flight characteristics and the control reaction were all totally crapulent unless it was going pretty fast. In other words, it always seemed to be flying just past stalled, at least from a controls perspective.

I think part of the problem is that I didn't make big bold to-the-wingtips elevons. I'm not sure how the wing works at varying speeds, but I bet the wingtips matter more at lower speeds or similar, and that would explain why controls were only reactive at high speeds (when they were reactive, they were pretty dramatic).

I did get to try some new construction techniques. This is another satisfying use for hot melt glue. To hold the wings together, I decided to forego everything complex/heavy/strong. At the center I just had a single wrap of packing tape. Then after I broke the wing (it has a habit of landing hard on a fact it seems to fly through the air as well like a javelin as a wing) I added a single strip along the top and bottom of each wing, towards the front of the wing. The wings are getting all sorts of warped (to no apparent effect, heh), but they have survived lots of impacts with just this little amount of tape. This is also the first plane I've built with a fuselage around the equipment rather than simply taping stuff on. I pretty much just cut up random scraps of foam to make walls for the equipment area (mostly because I wanted to protect the new li-poly battery). I have to say it worked out quite well -- usually I have a battery go flying around at least once in a night, but this prevented all the nuissance behaviour while also protecting the battery and wire.

I think the biggest deviation from Wally Wing that probably impacted stability a lot were my elevons that were all alone on the trailing edge and covering only a small percent of the back of the wing. On most other flying wings I've seen the elevons covered most of the back of the wing (or at least were pretty far out along the wing). The wally wing in particular has a much cleaner trailing edge all around. I would just tape on elevon extensions, except that this wing flexes a good amount, and that makes ailerons and elevons weird. In fact that may be the entire problem -- can't wait until I stiffen it up with fiberglass.

In the future I'll want to add a brushless motor (1-2oz) and some thin fiberglass (1-3oz) and maybe cut the wing down a bit in size. The success of a small amount of tape at providing strength leads me to believe that even a thin layer of fiberglass will make a light wing virtually indestructible.

I just got back from throwing it around with extensions taped onto the elevons to bring them almost to the tips of the wings. Much improved! It behaves much more plane-like. But the flex of the wings is a real problem, as anticipated. The elevon tips pretty much ripped off their tape as the wing pounded into the ground. When it got back inside it was an ounce heavier from all the water and mud *sigh*.

March 29, 2005

The prodigal plane! There were a series of storms over the weekend, and I went up to check that the plane was still in the tree in the middle of the storms. But it apparently fell out at the end of the storms, and the next day someone found it in the street and put a sign on the nearby telephone pole. Thanks Kyle!

So I got it home and checked the damage. The voltmeter says the Li-poly is toast -- 1.5 V. But, being a little adventuresome, I just took my spare battery and threw it in place. Everything works just as before! This thing had a LensRC motor, Castle Creations Phoenix-10, Hitec Feather rx, and Hitec HS-55 servos. Kudos! Surviving the rain with the battery still plugged in, with no damage at all (knock on wood)! I sure wish the Phoenix-10 had a second low-voltage cutoff that triggered like .5V below the first one, to save the battery in the not-rare-enough event of long-term treeing. Or maybe the Kokam batteries should have the cut-off, but then they'd have to worry about how to bypass it when recharging.

The really surprising part is that the balsa pieces (the rudder and elevator) are both still fine. I think the wood feels a little different, and the elevator has some non-critical cracks in it, but considering how much water they probably absorbed, it is nothing. I guess the bamboo cross-grain supports that Zack added really did their job.

I can't find any information about recovery of li-poly from deep discharge, so I think I will recharge it at a slow rate and see if it recovers at all.

So far so good -- I am charging it with a fixed-current charger I use to trickle charge NiCads -- something like 25mA (this is a 1200mAh pack). I remember reading somewhere that if you ever get a Li-poly deep discharged, you should charge it at no more than .5C until it reaches a sane voltage, so this is kind of following the same plan. :) It is increasing the voltage at about .01V/6 seconds (1V every 10 minutes). From about 1.5V unplugged, 2.6V when I first put it on the charger, now it's at 3.4V and it has slowed considerably. I am sure it won't have a very linear charge curve, but when it reaches a more sane voltage (say, 6V), I'll put it on the li-poly charger at like .2C and see how long it takes.

I suppose I really should be charging both cells in parallel -- they may not have sustained equivalent damage from the deep-discharge and may need levelling. It is not difficult to level a pack, but i do not really want to cut this pack open. So I guess I will only do it if it doesn't hold a very good charge after this.

It has slowed down its charging substantially, so far averaging about 1 volt an hour. It has reached 5.3V and so far I have only put about 75mAh into it. My hypothesis is that it stores much more energy (exponentially so) in the higher voltage part of the charge than the low-voltage part.

(the next day) It seems to have charged, after experiencing about 9 hours of 25mA to bring it up to 7.2V, then about 2 hours of 110mA to bring it up to 7.6V, then about 3 hours of 250mA, for a total of about 1300mAh, of which about 1000mAh was delivered after it reached 7.4V. It was originally supposed to be a 1200mAh pack, so if I can draw a sizable fraction of that 1000mAh in real use, it will be an "acceptable loss" and I will keep on using the pack, with pleasure. Straight off the charger it was at 8.45V, and I put it on a plane and just ran the motor for about 20 seconds and it was down to 8.38V after that. Not bad considering how badly it was discharged.

In sadder news, when I actually tossed around the plane, it turns out that the motor has seen better days. The repeated pounding into concrete has definitely jarred some of the magnets loose, and sometimes it won't even spin (but once you get it spinning it is happy). My dad speculated that the magnets were being pulled towards the stator, and I was like "that's crazy", but then I looked really closely, and that's what's happening. Once it's up to speed they stay out of the way. But I'm going to have to add some glue to the system at some point. Considering the nose-first approach to landing on concrete, I am not terribly surprised.

Also, my dumb ass dropped the little shaft clip in the grass. *sigh*

May 13, 2005

So I've been building this computerized mixing thing to add to my radio. The idea being that then I'll get easy access to features such as mixing (like for elevons), exponential control throws, and aileron differential. I'm really satisfied with the design, but the implementation isn't there yet, so it is kind of frustrating waiting for it.

We built an aileron ship a while back, and I actually have a lot of video of it behaving weirdly. The trouble is that we put little ailerons at the tips of the wings with a tiny vertical stabilizer and no rudder and were experiencing strong adverse yaw tendencies that, coupled with a slightly back center of gravity, resulted in a bizarre frisbee-like aircraft. We actually had all of the clues necessary to figure it out, including a couple long (for a craft with one aileron) flights with -- essentially the perfect experiment ground for playing with simulated differential by isolating the control to only one wing. Anyways, we never did figure it out. I had one crazy idea, but I didn't really put any faith in it and it really didn't seem like it could be causing this dramatic an effect.

But then I did a little bit of websearching and there's this adverse yaw phenomenon that is quite common and well understood and particularly affects craft like this one. It's caused by an effect that my intuition says should be minor, but turns out to be huge in some situations. What happens when you turn left with ailerons is the aileron on the left side of the ship goes up (producing less lift) and the aileron on the right side of the ship goes down (producing more lift). When the aileron is down, it is acting kind of like flaps -- it is increasing the amount of lift but also dramatically increasing the amount of drag. The opposite effect is occuring on the other side of the plane -- the upward-ailerons allow the wing to fall through the air in a reduced drag fashion and gain speed. The resulting yaw effect goes in the opposite direction of your turn. You can solve this problem to some extent with coordinated rudder usage (or mixing), or some variation on aileron differential.

The idea of aileron differential is to do something to make the up aileron have about as much drag as the down aileron. The typical approach is to make the control have much less travel in the lower half of its movement, so that the down aileron does not produce as much lift or as much drag. There are other cool approaches but I figured this differential idea fit well in my head and is something I would enjoy playing with. It also fit with my experimental data: flying with only the right aileron worked quite well so long as I only made right turns (thus using only up aileron control).

So anyways, that all happened a few weeks ago. We entered into this night of experimentation in the context of poor aileron behaviour and an inclination to wait for the computer to get differential right. So we were not going to bother with ailerons. But we're also recently enamoured of flat wings, I guess, so rudders also seemed out. So we came up with a lot of not-aileron ideas for roll control.

The most ridiculous was the shark idea. We put a huge vertical control surface over the center of gravity of the aircraft -- basically an oversized rudder in the middle of the plane. The hope was that by providing this side-to-side force with no fore-aft moment arm we would neutralize the rudder effect and instead it would roll the aircraft. It didn't work. Are you surprised? It had just a very very weak rudder effect. Forget we ever mentioned it. Really it would have been alright if we had put another fin on the bottom to provide a counter. Since they weren't generally lifting surfaces, I think they would have acted like ailerons without any need for differential.

Then we kind of had this idea that started out as big dramatic airbrakes on the wingtips but in reality turned out to be undersized funny-shaped ailerons with only up movement (i.e., a pull-only string system... down aileron movement just introduced slack into the line). They provided good predictable and uselessly tiny roll control. I guess they qualify as a proof of some concept, but the church parking lot was simply not ammenable to their turning radius.

I noticed a kind of neurotic flying habit I have. I try constantly to keep the plane within a relatively narrow range of altitude and pitch. It gets worse when I'm in limited urban spaces, where I wind up doing silly stuff like keeping the plane below 20 feet constantly and trying to do tight turns with unpredictable control surfaces within 10 feet of the ground. This habit has not made our planes into terribly long-lived instruments of mayhem. I think I picked it up due to my overwhelming awareness of the urban environment I'm flying in and my inexpertise in climbing rooves. So I guess I'll hold onto it until next time I'm at the farm or something.

The little LensRC motor I got has been one of the really outstanding aspects of these planes. The first thing I notice when I throw this motor on one of our thrown-together stick-foam constructions is that it is really easy to mount. Then I notice that Zack seems to have a little difficulty hanging onto the plane if I throttle it up when he's not expecting it. Then I get the plane into some crazy attitude in midair and it starts climbing violently, or hovering, or drifting slowly towards the ground. And never am I stuck in level flight with insufficient forward velocity. Finally, an electric power system that simply provides an easy to use harness thrust that hardly weighs down the plane!

But the real thing I'm learning is how durable it is. A long time ago I broke some of the magnets loose with some concrete. They would sometimes jam when the motor was first starting, so I decided to reglue them with some CA. It should surprise no one that I did a very poor job of this, and before you know it, the magnets were all loose again. I was watching a video from a couple weeks ago showing the LensRC motor making some bizarre rattling noise and you can see two guys standing around it poking at it, and in the next clip the plane is in the air. I mean pretty much nothing will make the motor unflightworthy. Tonight we finally cracked a magnet, and magnet dust completely jammed the thing. It took a few tries, but I finally got all the magnet dust out, and it runs just fine, even missing about a quarter of one of the magnets.

"Now, Greg," you might say, "why not simply stop running the motor into stationary objects?" And basically, I am starting to see the wisdom of our haphazard approach. We're learning basic lessons of aerodynamics, construction, and piloting all through the basic experience of trial and error. Certainly, we read a lot and talk to some people I'd even consider experts. But ultimately we just do what seems like a good idea at the time, and then later we learn why it wasn't such a good idea. I can think of a few lessons and the planes that taught them to me (some lessons had several planes...I'm a slow learner I guess). The fact that I can get motor technology that supports this approach is sweet, sweet icing on the cake!

December 30, 2005

Shortly after my last entry, I bought a house and I've been too broke and overwhelmed to play much since. But I *do* have a huge workbench in this house, so...

A few months ago (Feb 2005 Model Aviation) I saw plans for Bob Aberle's Scratch-65 (a 65% scaled down extra light version of the Scratch-One, published there before I subscribed to MA). It is designed to introduce us new timers to the art of balsa stick construction, so it features an interesting shortcut. Instead of cutting out a zillion little ribs, there were just bent sticks of balsa over the top of the spars. I decided I would build one.

The Scratch-65 was too small (Bob was using super-micro servos and receiver that I do not have), and I couldn't conveniently get ahold of Scratch-One plans, so I went ahead and built one a little bit smaller, say a Scratch-90. Hopefully its flight characteristics will not suffer too much from the number of decisions I had to make in this approach.

The wing features polyhedral (a flat center panel with two raised tip panels). Here is the center panel:

If you look closely, you can see that some of the balsa sticks broke at the apex of the bend (over the main spar). I eventually determined that this was actually occuring after I lifted my fingers off the sticks -- I think that some glue-related process was causing an increase in tension. Probably a particular pattern of glue application would have solved this problem, but I didn't figure it out. I solved it by fiat, declaring that since they were still strong, they were still airworthy.

Here are the tip panels fresh from assembly:

Here is the complete wing (still without covering).

The plans called for an angled cut of 1/64th inch plywood to hold the tip panels on, but that didn't seem sufficient to me (perhaps I didn't understand something). So I also super-glued on tiny bits of balsa to fill the gaps at the LE/TE of the join and wrapped the whole join with a thin strip of scrap fiberglass (1/2oz, I believe) and epoxy:

Tomorrow (epoxy willing) I will cover it with MonoKote.

September 11, 2010

Yay! I broke a prop yesterday!

Various lifestyle elements conspired to keep me away from flying. I would always have a strong desire in the winter (when it is not really a good time for it), but in the summer I aws always too overwhelmed with the rest of my life. Which is a bummer because I never flew that plane I made in 2005! Not long after my previous entry, I covered it in monokote and cut out control surfaces for it, and then it's sat ever since.

So yesterday Zack came over and we threw together one of those foamies, with control surfaces. And since I didn't have a battery eliminator that wasn't firmly connected to a brushless motor, we threw a motor on it.

The "micro 10g" motor from gobrushless is a real treat. The only thing I don't like about it is that it is actually so small that it is a minor ordeal to mount it. We wound up lashing it to a little balsa stick which was then rubber-banded to some heavy 12 gauge (house wiring) copper wire that was stuck through the foam.

There is one problem with the motor. The Castle Creations Phoenix 10A ESC will get caught stuttering if you increase the throttle too quickly. Basically, full power mode in the ESC creates a failure mode if the motor is not spinning. But in practice, it is repeatable -- if you bring the throttle up slowly, the motor always comes on and delivers power as you would expect. Anyways, I'm sure if I bothered to learn how to configure the ESC, I would surely be able to fix it somehow.

The foamie crash characteristics are really top notch. The whole plane is around 8-9oz, but on impact the wings (2oz) and the battery (also 2oz) fly off immediately because they are held on only by rubberbands. So the impact force on the prop/motor is only about half the weight of the airplane. Plus the "prop saver" uses a rubberband as well. So when we flew it into grass about a dozen times, we didn't break the motor and we only broke one prop. Yay!

Anyways, I decided to finish my balsa plane. Which is to say I have vowed to crash it. Or you could say that I have promised to fly it until the wings fall off (or whatever).

I am considering the directions I should go in the future and whether this plane is really any better than a foamie.

Weight is a major game. The avionics package (2x HS-55 servos, 1200mAh 2s1p li-ion, receiver, ESC, and micro 10g motor) weighs about 3.5oz all together. The foamie weighs about 4oz (2oz wing, 2oz fuz). That puts a lower bound of 7.5oz for the assembled plane, but I know they weigh a little more. Combined planes tend to weigh more (1-2oz more) than the parts that go into them, presumably because of things like extension cables, rubber bands, tape, control wires and horns (and even control surfaces) and stuff like that. You know, stuff that never makes it onto the scale separately.

I can replace the fuz with a 1/2"x1/2" balsa stick weighing about 0.7oz (compared to 2.0oz). My carefully-constructed balsa box fuz is also about 0.7oz, and much more rigid. But a real downside to the balsa box is that I did a poor job -- the wings and the elevator are not exactly aligned, for example! But the balsa stick comes square from the factory. Also it is relatively undurable and once it breaks it is pretty much garbage.

Anyways, this balsa plane is about 7.5oz.

Another place I can save an ounce is the battery -- I could get a new smaller li-ion and reduce the avionics package.

The motor is only 0.3oz but allows me to use a much less than 1oz battery to power just the servos, giving me a net savings of 1.3oz.

So I think the glider version of this plane would weigh as little as 4.5oz.

But really, anything under 10oz is fine. It is my opinion that this tiny motor is plenty sufficient for models up to 10oz. It can't quite pull it straight up, but yesterday our flying conditions included a little bit of wind and when I was at the stick I was able to fight into the wind and maintain assertive control, which is pretty spectacular for a light foamie in the breeze.

I looked up the motor and it is supposed to do 5oz thrust while swinging a 6x3 prop with 2 cells, so that is just enough to lift the avionics package straight up, and therefore plenty to pull a 10oz plane sideways.

Anyways, any way you slice it the foamie wings are a steal. They're the same size and weight (2oz) as my hand-built wings, but they're cheap and quite durable. I've got a big pile of broken foamies in the basement and for the most part the wings are intact or at least salvagable. It is the fuz that breaks, and the nice fuz-wing connection point. But if you abandon the foamie fuz you can just take the two wings, cut off the connection point, and put a single wrap of packing tape around where they meet and you get a wonderful wing (and depending on how you can cut it, you can decide dihedral and sweep-back).

I think I want to finish one of the flying wings.

September 23, 2010

I wound up flying right after I wrote the last entry. I found a good field, though its hills made a weird play with the sunset winds. I flew it until the battery died, rather than until the wings fell off. Which is something new for me.

The neatest thing about the airplane is it handled just like the polyhedral rudder-elevator ship in the sim (Slope Soaring Sim, SSS). Whenever you hit full rudder, it would yaw and then roll and then when you released it it would jerk back in a very awful way. Which I kind of assumed was simulation error resulting from the fact that the model in question wasn't programmed correctly into SSS. But nooo, that's what happens when you combine intense polyhedral with rudder control. It wags its tail.

I went shopping. I got a ton of cheap stuff from hobbypartz, some props from gobrushless, and then I discovered radicalrc, for the stuff I forgot from the others (mostly Deans connectors).

Speaking of Deans connectors, hardly anyone specifies an amperage for them! I think that is awful. Anyways, I found consensus on this point from disparate sources: the JST connector is good up to 5A (it will do 10A but will show some resistance). The micro deans is good up to 10A. The ultra deans is good 50A or beyond. Which, now that I've seen it, I don't doubt. I really respect that they're the first little connector I've seen that uses the spring on the non-contact side of their connector (i.e., pushing the contact instead of bridging the gap).

I'm wasting it on a piddly 18A ESC.

So I ripped out all my JST connectors and put micro-deans on everything except the big batteries that get ultra-deans.

Also in connector world, I have abandoned those stupid gold-plated "bullet" motor connectors. I was never very good at disconnecting those damn things. I tried a couple days ago with a pair that I had soldered (because we were going to use the motor control as a BEC without the motor) and it was damn-near impossible. I built it myself so the heatshrink wasn't as tight as professional and it just slid off the connectors. But then once I had bare connectors I still couldn't get enough of a grip on them to pull them. The only plausible way to disconnect them is to yank on the wires and hope my solder joins hold, which is just pathetic.

So I have joined that chorus of people who says "motor controllers are cheap these days anyways" and directly solders the motor on. I mean, the only time I really need to disconnect it is when I total a motor anyways, and I'm actually only on my second motor ever.

Anyways after all that busy work on futzy little connectors, I'm a little bit better at free-wire soldering and heatshrinking than I used to be. I can't believe I had to see some dude do it on youtube before I figured out that a butane lighter makes a better heatshrinker than a soldering iron does.

I am going to just weigh some stuff and post the numbers so I can have them for my own reference:
2.7oz 18A ESC + my biggest motor + 8x4 prop
0.7oz 10A ESC + 8g motor + 4x3 prop
0.5oz 6-channel rx
2.9oz 2s 1300mAh
1oz 2s 500mAh
0.9oz 3x micro servos

So it seems like my smallest electronics package will be on the order of 2.8oz (and my biggest will be 7oz). 2.8oz is only marginally less than its thrust, which supposedly will barely reach 5oz with a 5x3 propeller. But I have this idea of an indoor high-alpha toy. And the indoors I have in mind is a very small and cluttered basement. Which means I need to be able to surround the prop with foam, which is kind of a lot of weight maybe?

An appropriate-sized sheet of this paper-laminated foam-board that I have is about 0.5oz, which is about the heaviest that I can realistically expect to go 3-D with. Or I could use this thin spruce piece and cover it. One way or the other I think a less-than 1oz 3-D "airframe" isn't too far out of the question.

I was thinking briefly that I could build a somewhat bigger 3-D ship and fly it indoors too, but I realized that is exactly what is in my plans with the IFO.

So this weekend I think I'll get to fly my IFO (first time). I'm pretty excited. I think the prop+motor combination should be pretty close to the capability of hovering, but I'm gonna have a big field to play with it in so I can fly fast and level instead.

September 25, 2010

Back in like 2004 I bought this "Indoor Flying Object" from uh, whoever makes them. I immediately assembled the kit, but I never bothered to put electronics on it because I didn't really have a big enough motor. But in this last shopping trip I decided to get a slightly larger motor than the usual "micro" size that I usually buy (uh, this one is 37g). And I had this idea it could swing the prop hard enough to fly the IFO. I bought a full-size IFO instead of the mini IFO because I'm dumb, or something.

So today I flew it! I have lots of skepticism about this model. For example, the servo-elevon linkage seems designed to tear the teeth of the gears. Anyways, when I got it out it had all sorts of disappointments and I was about ready to declare the thing crap. Need I point out that all these problems were my own fault? I mean, you could tell, right? Anyways, eventually I got a bigger prop on it, shifted the weight around a little (though really it was pretty resilient to the battery moving about), and got it far enough from the ground to have assertive control (I settled on a kind of light frisbee launch).

This thing flies like a dream! It is the best. I've only gotten to fly an aileron ship once, and it very briefly (it was waaaay too hot for me to handle with a little 0.15 glow engine...drove it right into the ground). Other than that, I've always flown rudder-elevator ships, UNTIL TODAY.

In the SSS flight simulator there are a pair of nearly-identical flying wings called uh the Weasel and the Moth (I think based off of real models, actually). And they are outrageously good, and this thing handled basically like them. The only thing that wasn't cool is that as such a large floaty ship, it was extremely susceptible to the winds.

Oh, did I mention that I flew it outdoors on a relatively windy day? The wind wasn't very intense (I was always able to make some minor headway into it), but it was very gusty and constantly switching direction. Gee I wonder why I had so much trouble keeping it off of the hillside.

But once I got it going, it went for ages. I flew the battery almost entirely down in two or three flights. And the landings were only necessary because I'd get it into an awkward configuration with regards to a wind gust while low to ground. But if I caught a wind gust high up, I could recover. But man, I should paint one side of this thing a different color! It's got a tendency to flutter every which way in the breeze if you get it sideways (gee, it's not like it's shaped like a frisbee or anything) and I'd just have to wait for it to establish an orientation and then play "spot the tail" to guess which way up was.

Anyways, lots of fun.

To my surprise, the motor only got hot once and I don't know why that was because generally it wasn't hot at all even after extended flights (though I did start to use it closer to 60-70% throttle after I got used to it). Even though it was spinning a 10x6 prop! And I didn't break the prop once, even though the prop rubbed on the carbon fiber front spar from time to time (I mounted the motor too far forwards, I guess).

I think I kind of want a slightly faster/more powerful version of my rudder-elevator toy next... I think I'll just put a bigger prop on it and see how that goes.

I discharged the battery all the way to the low-voltage cut off. When I measured it at home, it was 6.5V and the cells were "balanced" at 3.33V vs 3.17V. Which I guess will do because I don't really have a leveler, but gee it looks like if it gets any worse then I'll have to buy or build a leveler. *sigh* I think I can build one out of a comparator (or two?) and two transistors and some resistors. But then if I ever get a 3S pack I'll wish I'd bought one...

Talking about sighs, I had a little radio charlie-foxtrot. I bought this new fancy 6 channel 2.6GHz piece of crap from Exceed. It's supposed to be the same as the FlySky FS-CT6A etc. Anyways, it mixes all the channels together (crazy defaults, maybe heli mode?). When I hook it to the PC, it doesn't quite talk the protocol that was published by Hans Brand for this sort af radio. It's probably slightly newer. But to top it off, it hardly even talks what it does talk, intermittently spitting out gibberish or dropping characters or who-knows-what. So until I sort that out, I can't fly with it.

I say, "no problem, I built a computer mixer for my analog radio." No sooner do I reorient myself to its UI (it is really pretty slick) than I determine that it sucks. When I have it in the loop, all of the servos chatter like crazy and not in an insignificant way either. Which is a real bummer because I want exponential and a little bit more configurability for my elevon settings.

But oh well, so I have the little hardware elevon mixer that I mount on the airplane, and it works great. Yay.

So I need to fix my homebuilt computer mixer (though god knows what the problem could really be); and/or reprogram the Exceed, and/or gut the Exceed and put a new computer in it. Which is kind of tempting, but really I don't have that great a track record for building computer radios, apparently. I'd probably do no better than the Chinese, and perhaps even worse.

Anyways I have two absolutely delightful airplanes at the moment and neither of them is completely crashed. Sweet!

Though I will confess the IFO is slightly more delightful than my Scratch-90.

September 26, 2010

HobbyPartz had some ridiculously cheap 1oz camera so I had to buy it. And put it on my Scratch-90.

I also upgraded the plane from a 6x3 prop to a 7x3.5 prop, which is kind of ridiculous with this little gobrushless 10g motor on it, but the motor doesn't seem to mind. The battery minds a great deal, as is highlit in the first video. I can barely get half a minute of flight at a time because the battery presumably overheats and then stops giving enough power. The battery is years old and has been abused so it is not too much of a surprise.

So I put a newer smaller battery in it, and it managed to get like 6 minutes on one flight and with much better power. So yay new stuff!

I did a little more testing on the "not so good" battery. It was 7.8V when I got home, which means it didn't fully discharge before it started being unable to power the motor, which is not too surprising. The surprising part is I put it on the charger and after only a few seconds, it showed red light and green light, which I think is an error condition. I assumed the charger somehow detected that the pack was mad out of balance (it is very old and was deep deep discharged in a tree many moons ago). So I ripped off the Kokam shrinkwrap to get at the half-voltage tap. It is 4.17V vs 4.24V. The first surprising part is that it's not very out of balance at all. The second surprising part is it is already up to a 2S voltage of about 8.4V, even though the charger didn't hardly get a chance to do anything to it. I'm not sure what exactly this means. I didn't realize that the discharge amperage was so limited when a battery aged. Anyways, the battery is probably as good as dead because all it would be useful for now would be models that don't draw too many amps, but such models would have to be fairly small and the small ones don't need such a bulky battery. Right?

Anyways, next time I think I'm gonna try to fly this plane with the much bigger new battery (1300mAh 2S). I think the plane has enough lift for another 1oz of weight, and the bigger battery ought to be able to run it strong for 10 minutes or so.

September 29, 2010

Today I decided to fly my yellow polyhedral+rudder plane again. I decided not to bother with the camera (too much weight). Instead I was going to try a heavier (2.9oz 1300mAh 2S) battery, and bring the lighter (1oz 500mAh 2S) battery as well. Both batteries are brand new, and the light one was a delight last time, so I figured I'd probably get away with it.

It's slowly sinking in that using the 10g gobrushless motor is a mistake. I _love_ light things, but I'm not a talented enough pilot or builder to work in the "underpowered" range. Apart from the fact that I'm constantly flying at full throttle even though I somewhat overpropped it (7x3.5), there is the fact that it gets burningly hot now. But it hasn't melted, which is probably just a testament to my decision not to buy any 3S packs.

Plus there's the writing on the wall -- looking up models that are designed for motors this size, they are all approx 3oz AUW. Which is about the weight of my electronics package. And I don't seem to be able to make an airframe which is itself less than 4oz. I figure eventually I'll make a little undersized flying wing and try to make it under 1.5oz airframe, and these little micro brushless motors will be acceptable.

Anyways, the big pack was obviously too heavy and had the center of gravity a little too far forward even if I jammed it as far back as it would go. But it was flyable.

But my other mission in going out was to get a little better at close-in flying. I wanted to do more tight buzzes of my air traffic control tower (uh, my body), and many more in-low strafing run sort of flights, exploring the hollows along the hillside and so on.

So it is only natural that I was cocky enough to fly the thing nose-first into the ground 3 times today, after I thought I had grown out of that phase as of last weekend. Which happened to really blow open a problem with the motor.

Since when I first bought it, if I run the throttle all the way to maximum then it would stall the motor and kind of stutter around for a while before some ESC cutoff killed it. But a nice slow startup works. I put "mess with the ESC's soft start setting" on my todo list, but never bothered. I would just start it very smoothly. Anyways, the more times I did this the more apparent it became that it would always stutter briefly as the throttle passed a certian threshhold. And over time, the percent of times that I would attempt to start the motor where I would actually get a running motor decreased. Today it finally descended past the 50% mark, though I was always able to ultimately start the motor. I think maybe prop balance could be slightly improved, but once it is up to speed it doesn't have a problem so I don't really think that's it. From the sound, I think it is rubbing the magnets on the stator.

I played with the motor and the bearing that has always been "kind of loose" is now downright floppy in a pathetic fashion. I have definitely broken the motor. *sigh* But it also makes me feel a little better about my ESC, as a loose bearing cannot very well be its fault. I think it is necessary to accept a couple limitations when dealing with motors this small. When it stutters, what it is doing is the magnets are experiencing enough force that if the motor vibrates at all then it will pull out of tolerance and rub (???). At full throttle if the blade is spinning fast enough then it tends to hold itself close enough to centered that it doesn't rub (uh, usually). That's certainly what it sounds like, anyways.

Oh, that little battery pack? I forgot to charge it. I extracted one very brief flight from it, probably to its detriment (discharge to 6.25V *sigh*). Someday maybe I'll buy a new battery for my multimeter and stop being such a dunce.

Anyways, again due to poor depth perception I again flew my plane directly into the metal post again. Luckily the balsa wings seem to be immune to destruction, somehow!

I did have some fun flying. My plane was just enough out of trim that the TX trim was insufficient, so I had a lot of practice holding the stick just barely away from centered, and I did a few really nice smooth turns. And I became a little more proficient at making the turns precisely how much I wanted. I could fly some sort of four-sided pattern in that space. I still haven't lost a plane to the trees (surely it is about to happen, though, given my poor depth perception and my increasing cockiness -- *sigh*). I even increased the servo throws (moved the linkage to an outer hole) and continued to improve my finesse with the aircraft. It was relatively windless so I was lucky.

I made a couple attempts to fly inverted, and did not crash the airplane during these ridiculous attempts. Keep in mind that my wings are not only asymmetric but have huge polyhedral. If I held full down elevator it would get about 10% past straight-down and then flutter in a preposterous way (uh, tail stall?) until it was facing in some horizontal direction more rightside-up than upside-down. It has great stall behavior, at least. Anyways, once after doing this the plane would not climb anymore. I thought perhaps my battery was dead or my motor had finally stopped spinning. But I was able to power up the motor and it seemed fine so I threw it up again and in order to be level it needed so much up elevator that it did a tail stall the moment I added rudder, and had a fairly graceless belly-flop (at least it wasn't a nose-dive). The battery had shifted forward while it was "futtering." The battery compartment of the plane is very crappy.

The whole front of the plane is falling apart now from the nose-strikes, but it doesn't seem to matter because there is a thick dowel running back to the wings which is what the motor ultimately rests on, and what the rubberband catches attach to.

Anyways, another busted prop, and probably a busted motor.

A good day at the park.

Oh and I figured out a disadvantage to soldering the motors onto the controllers. If I ever decide to make a pusher I think I'd need to reverse the direction of the motor, and cut some wires. Not that that's any harder than disconnecting these bullet connectors. And I do think I may want to make a pusher, so I can avoid destroying motors with nose impacts. Oh well.

Or maybe a mid-engine aircraft.

October 2, 2010

I seem to keep looking these up. I have 3 brushless motors.

gobrushless 10g micro RTR
2200Kv, recommend 5x3-7x3.5,
plus phoenix 10A ESC and 6x3 prop = 0.8oz?
aeo-rc C10 micro 8g
2900Kv, recommend 4x2.5-5x3,
plus volcano 10A ESC and 4x3 prop = 0.7oz
"37g brushless motor"
1300Kv, recommend 8x4,
plus volcano 18A ESC and 8x4 prop = 2.7oz

Since I destroyed the 10g micro, I decided to build a smaller lighter plane to try to get some use out of the 8g micro. I was afraid I would not be able to build a plane small enough to use these micro motors.

I was psyching up to build a balsa frame flying wing when I ran into a mass-produced paper-covered foam wing. All the wings I've used thus far are about 2oz and about 2sqft, but this one was about 1.5sqft and 3oz. In other words, it was a bum deal destined to be unused for all eternity. So I ripped the paper off of it and now it is about 1.5oz. And since it's a smaller wing than the ones I've been using, maybe it won't have so much drag and these little motors will scream a little bit more on it.

So I built a basic stick plane around the wing, using the micro 8g and a 6x3 prop:

It is only 4.7oz AUW.

It flies amazing!!! As is often the case when I walk out to the field at sunset, when I got there it was suddenly preposterously windy. However, I flew anyways. It went faster than the wind, no problem. And it could climb near vertical, and could do vertical loops without any effort whatsoever. And it did that hillarious fluttering, flapping, and bending that is characteristic of unreinforced foam wings. But to no obvious detriment to the wings or necessarily even to its flight characteristics.

And man, when you took the power off, it was a total floater. It had a totally respectable glide path and I kind of had the feeling that if I'd had a rudder I would have been able to slope-soar for a little bit (the wind was actually almost cooperating for that).

I think the real gain was in the Kv of the motor. I didn't pay attention to the motor Kv when I was shopping, but the low Kv motors really aren't going to peak out on a 2S pack. I need to either get a 3S pack, higher Kv motors, or pitchier props.

Anyways there was one tiny little catch with the airplane. My rudder linkage was way sloppy, to the point that I basically did not have rudder control. I would hold full rudder deflection for a second and barely get tilted enough to effect a turn.

There were a few reasons for this. One is the linkage rod was too flimsy (long, thin, and pre-bent for trim). Also, to save weight, I had just single 1/32 inch (I think) sheet of balsa for all the tail feathers. So there's really nothing rigid between the servo and the rudder. Also, even if it had worked, I'm not sure the wing was still stiff enough for its polyhedral to effectively roll the aircraft (though that wasn't a problem in practice -- I wasn't even getting yaw).

I'm kind of surprised this isn't a lesson I learned earlier -- I probably should have but missed it (because I've done flimsy tail feathers like this, I know -- I used to even use rubber bands to hold the vertical stab to the stick!). Anyways, my control linkages have always been maximum sloppy. Oh well.

Anyways, the power was so fantastic that I really am excited about flying it again. But I'm kind of tempted to abandon these rudder planes. I think if I'm gonna bother to build ailerons then I'm gonna go all out and build elevons (flying wing) so I'm not tempted to add an extra servo or a sketchy dual-linkaged servo. Might even get the weight a smidge lower. I mean, this airframe was apparently 1.7oz. I figure it should be pretty easy to come under 2oz for a flying wing - though maybe not because I only have the thick monokote.

October 7, 2010

I wanted to build and fly a little flying wing today, but I got at it too late and decided to just fly. Which meant existing models. So I took the 4.7oz model with the flimsy unpapered foam wing and unresponsive rudder. I "fixed" the rudder by using a thicker (stiffer) control wire which was exactly the right length (i.e., straight) and passed through a fixed point to prevent bending. I added another piece of tape holding on the rudder, and I observed that if I held the rudder with my fingers, the servo could move the entire airplane. Seemed good enough for me.

But it wasn't, I think the floppy wing is ultimately to blame. It has all sort of hillarious bizarre modes. For example sometimes a strong control input will cause the wing to flutter violently instead of rolling the airplane. But the good news is it would always eventually roll (after more than a second if you had to roll to the left from having been rolled to the right). The good news is that once you did get it leaned over in the right direction, the elevator is plenty potent and it has a nice tight turning radius.

But if you are expecting to be able to turn and avoid the tree in the last 50 feet if you aren't already leaning in the right direction, you will tree the thing.

Which is to say that twice I put it in the weeds at the base of a tree because I just dove when I had the panic moment. Then I got the hang of anticipating my turns. Really makes you think more like a pilot.

Playing with it in the basement, I was pretty sure it could prop hang. But I never really found out because the moment it would get below a certain airspeed, it would completely lose control authority (tail stall) and ultimately take a nosedive. Now I know why these 3D planes all have huge ridiculous oversized control surfaces.

Anyways that plane flew great with the new 500mAh 20C 2S pack until it was run down. Then I threw on the 340mAh 20C 2S pack, which is much older, and it could barely keep in the air. Bummer, the older packs are pretty much useless.

I brought the IFO too because I knew the first one might have issues. It flew pretty alright, but it pulls hard to the left and it is very crappy holding the stick a little off center all the time. There are two causes. One is that my Hitec transmitter has almost no trim control on channel 1 (aileron channel), probably due to a manufacturing defect. Also, while the motor is more or less powerful enough, it doesn't feel like it because it hasn't nearly got the "pitch speed" (RPM * prop pitch) for decently fast flight. This really hit home today because I noticed how long it took to cross the field. It actually inspired impatience, it was so slow.

So I think I will buy a 3S battery pack and maybe also a pitchier propeller.

Anyways, I've really gotta make my computer mixer thingy work so I can fly elevon planes without needing them to be 100% trimmed from the start. And you know, I ought to fix the channel 1 trim pot on my transmitter. And build a little flying wing. A todo list!

October 21, 2010

The 4.7oz plane was pretty peppy on a fresh pack of batteries (loops without a preceding dive, for example). I finally figured out (it should have been obvious) why I never actually found my planes to be that performing. It's because the pack is only fresh for about 30 seconds.

Thinking the floppy wing was to blame for poor rudder response, I replaced it with the slightly larger and about 1oz heavier rigid balsa/spruce polyhedral yellow wing that I had built. I was wrong. The plane still turned abysmally. I now think that the relatively small moment arm between the wing and the tail is to blame. Which is pretty obvious, really.

Anyways, the extra weight didn't help it at all. The motor is picking up a bit of a vibration (perhaps the prop has too many chips in it). Plus I'm ambitiously trying the much heavier (1300mAh vs 500mAh) battery pack on after the small one runs down. The motor gets searingly hot and has always barely enough power, all factors combined.

I tried the 4x4 prop and the 7x3.5 prop. The 4x4 didn't have enough thrust to make it move hardly at all. The 7x3.5 prop was obviously stressing out the motor and also barely sufficient. Back to the 6x3.

I started out with the center of mass too far back. I was able to fly it, because I was determined, but every time that I would use the rudder, the nose would dip tremendously. I diagnosed it in midair, and flew it into the tree beside me so I could adjust it.

By "into the tree" I mean the front edge of the wing hit the trunk of the tree at speed, blammo. It is completely indestructible, I guess.

October 24, 2010

I took the Scratch 90 wing I have and scaled it down to 85%, or roughly "Scratch 75."

I drew designs on foam board (which I could push pins into):

I glued the balsa together (old wing for size reference):

And covered it in the same monokote:

The bigger wing is approximately 42 inch span, 1.5 square foot area, and 2oz weight. The new wing is approximately 36" span, 1.1 square foot area, and 1.5oz weight.

I wish the weight had been a little lower. The wood parts weighed only 0.8oz, so nearly half of the weight is the covering. It's because I'm using regular monokote instead of one of the new lighter variants (which are apparently not as advanced as I had assumed).

Speaking of monokote, I have yet to form a satisfactory wrinkle-free finish.

The plan is to use this wing with the smaller brushless motors, so hopefully it will have less drag. My hope is it should be able to lift up to 10 oz and still be "light" (i.e., less than 10oz/sqft). But since it'll be a smaller wing, it'll be more prone to going faster. You never know. It is really a bummer that it is not substantially lighter though. But at least it gives the floppy foam wing a run for its money -- only slightly smaller for the same weight, and substantially stiffer and more durable.

The next day that it is not so windy, I will take it out and find out how it works in practice.

Switching gears to TX issues.

First there is something which has been bothering me for a long time. Channel 1 on my Hitec Focus TX has no trim. Moving the tab has no effect. Taking it apart revealed a rather interesting design. To their credit, the trim setting is entirely mechanical. Instead of having a second potentiometer (introducing all manner of possible failure modes), they simply have a little levered system whereby the trim tab rotates the main potentiometer independently of the stick input. Way neato.

However (for some reason unknown to me) two of the pieces in the trim linkage are separate pieces even though they literally rest on top of eachother with complete overlap. They were joined at the factory by melting a groove across them. For whatever reason, this approach failed (it has always seemed like a half-assed approach to me when I have used it). So the tab could move without moving the rest of the linkage. I didn't think CA would work between too pieces of plastic, and my plastic cement has apparently aged out (nevermind that I don't hardly know how to use it anymore). So I decided to go ahead and run with their stupid approach. But instead of making a single groove between the pieces, I made a dozen "spot welds" between these two pieces, all along their common perimeter. I simply tapped the tip of my soldering iron at their intersection, which created a nice little shared divot and appeared to do a pretty thorough job of melting the plastic together. Anyways, it's holding for the moment.

This is a little bit reassuring to me. I bought an extraordinarily inexpensive "ExceedRC" 6-channel 2.4GHz TX/RX. It has a number of issues of its own that I'm working through. But the single issue that I identified as "most Chinese" is that one of the solder joins was so sloppy that it defied belief -- they literally bridged a quarter inch gap with free-form solder. If even a Hitec radio has the same sort of really unacceptable quality control, then I probably didn't make a mistake buying the crappy ExceedRC thing.

So that's one long-standing TX problem. The other one is that my "PIC TX" module is broken. I should describe it -- I built it a few years ago but never actually tested it with a real airplane.

The Hitec Focus has a trainer port on the back which has a pulse chain input and output. I modified the Focus so that the trainer switch is a regular persistent switch, rather than a momentary contact. This way, I could hook up a device to the trainer port which processes the pulse chain and produces its own pulse chain, which the radio will then transmit to the plane as though it came from a trainer connection.

The device that processes the pulse chain is built out of a Microchip PIC18F1320 (a tiny and inexpensive 8-bit embedded microcontroller), a 16x4 text LCD module, 5 buttons (accessing soft menus), a 7805 voltage regulator, and a transistor (to drive the output pulse chain at 9V). It runs a little piece of software I wrote in "GF", a little FORTH-like language that I wrote a compiler for. I also wrote the "pic18asm" assembler that was used, because the regular PIC syntax is stupid.

The software implements input centering/scaling (to convert the stick position into positive/negative numbers within a range), exponential (well, it's on my todo list), arbitrary mixing (allows specifying a 4x4 matrix of mixing coefficients), differential (have greater range above or below the center point; for example for ailerons to counter adverse yaw), and output scaling/centering (to convert it back to a duration in microseconds for the pulse chain). It has 5 configuration banks, so it can store configuration for 5 independent models and switch between them at will. More than sufficient for the occasional elevon/Y-tail that I need to control (and I'm not into helicopters anyways).

So basically I exchanged a great deal of time spent in my area of expertise (uh, compiler design, *sigh*) for a working computer radio.

It has a bug when dealing with negative channel values that produced all sorts of issues. I found it just now by inspecting the behavior and then reading the source code. But I haven't uploaded the fix yet, because I know if I do then I will be up all night working on it.

There are pictures and additional information available at:

October 30, 2010

I decided to fly the IFO today. I can now trim it since I fixed the TX, and I have some new 3S batteries.

I'm honestly not super crazy about its design. The plus side is that the carbon fiber elements are indestructible, and the rip-stop covering is pretty nice too. But the whole thing is based on the idea that you can make a rigid join using CA and carbon fiber. The first time I picked it up, one of the landing skids broke. The first time I threw it, the main landing gear separated. Today one of the battery sticks and one of the elevon control arms separated. It just seems like these little CA joins that are the heart of the design are simply insufficient. I'm especially bummed about the elevon control arm, as these oversized elevons with the permanently-attached servos are just destined to eat servo gears. The fact that they're already eating their own control arms is a bad omen.

The good news is it can kinda-sorta fly with only one meaningful elevon!

But overall I'm just not overly impressed with its flight characteristics. It does roll roughly like it's on rails (at least, moreso than any other plane I've owned). But it was not stable at all. Especially it lacked pitch stability.

I think there are two problems. One is that it is windy. But it always seems to be windy, and my more airplany-planes handle this alright. The other is that it isn't very airplany. Forward velocity is not a substantial asset to this airplane, and when it does have forward velocity (such as when I'm throttling up into the wind), it becomes extra super whimsical.

I wonder if maybe the CG is a little far back? Too bad it's not convenient to move on this design.

Anyways, I think I like slightly more airplany things. I should probably focus soon on building a wing with ailerons, or a flying wing.

November 20, 2010

I gave the IFO another try. I used the heavier 3S battery so it couldn't quite do unlimited vertical but it was quite a bit of fun once I got it tuned!

The tuning was exciting. My "picTX" computer radio is working fine now so I have a lot of flexibility. I mixed throttle into the elevons , so that it flies mostly level at any throttle setting! It only controls pitch, but maybe soon I'll bother to compensate for the (managable, by comparison) prop twist as well.

The IFO is really a dramatically unstable flyer when given -- uh -- erratic control inputs. So when it wasn't tuned for level flight at all throttle settings, the situation would very rapidly get to the point where I'd need to input a severe course correction, and at that point I would often lose orientation sense.

And once it was flying straight, it was really easy to do snap rolls! So I practiced for a while, doing split-S (roll then half loop) and uh a "roll off the bottom" (half loop then roll). It was a treat to finally have a plane that could do these maneuvers effortlessly.

In other news, I finally painted it white on the top so I can tell what it's doing when I do lose orientation.

The only thing disappointing about the IFO, though it's not really a shock, is that it can't glide for shit. Even in its heaviest configuration it only weighs about 10oz but it has a ridiculously large wing area coupled with a ridiculously low aspect ratio (drag central). So at low speeds it kind of flutters to the ground in a fairly managable way, and at high speeds it ... slows down to low speeds. It just won't hang on to energy to save its life. You can't get going fast enough in a dive to pull off a loop (though with the lighter batteries it will loop indefinitely without a dive).

The propsaver's rubberband has broken twice this month, and twice for my whole experience. So I think it doesn't like coldness. *sigh* If that's the only component I have to devise a replacement for in order to fly through some of the winter, I'm satisfied.

I've been watching birds. Especially the soaring kinds. A common raven is about 22 inches long, 55 inches wingspan, a chord around 7 inches, and about 24oz. Wing loading around 8oz/sqft. An aspect ratio around 1:8.

I'm really impressed by their ability to work to gain altitude and then turn off their throttle and glide around.

November 27, 2010

Yesterday it was windy and I flew my little (scratch-75) wing on my yellow fuz (with good rudder-elevator control and a 10g motor). Since I gave it the 3S battery, it flew fast enough to handle the wind but I felt like all of the configurations I flew had to be just a bit too much nose-up in order to fly level. Maybe that's just because I was flying on the leeward side of a hill during a substantial wind event (consistent down-drafts). Anyways, I enjoyed it, but I totaled the fuz nosing in, again trying to fly inverted.

Today it was a little less windy so I went back to my IFO. In addition to the 3S battery and mixing the throttle into the elevons, I have found a few new innovations that make the thing a really super performer: no wind, painting one side white, and exponential (50% stick = 25% control throw). Painting the one side white is the most important, because I can now send it tumbling and immediately discern its state. But the low wind was also a real bonus. When the wind did pick up while I was flying (Murphy's law), the airplane became much more unpredictable and had many more annoying tendencies that you were constantly fighting. But the exponential made it possible to fight the annoying tendencies without always bouncing from one extreme to the other.

I flew it with a 3S 0.8Ah pack and it had nearly unlimited vertical and could do loops from level flight. With a 3S 1.8Ah pack (heavier) it could climb on command but it was very timid at looping. With a 2S 0.8Ah pack it could fly pretty assertively but it could only marginally climb over the downdraft (the wind picked up by then), and looping was right out. I was impressed by how well the 2S performed, even after the 3S. It wasn't quite as night-and-day as I expected.

I definitely need a bigger motor, and it is just a question of what kind of plane I want to design for. I'm still pretty set on flying wing, so I really just need to pick a size. I guess now that the 10g micro motor is homeless, I should just build a tiny plane for it.

December 30, 2010


At the beginning of December I built a new wing. There are two new features in this wing. It has ailerons. It is built out of balsa ribs. I just drew an "airfoil shape" on my foam board and then traced it onto a piece of balsa with an x-acto, then I cut a bunch with that as a model. I did a very poor job but I knew I would -- the real experiment was to see if you could make a wing using this approach even if you don't cut very carefully (i.e., instead of buying a laser-cut kit).

Glue together the frame:

Add servos (they are inside the covering in this design):


Add ailerons and linkages:


The wing is very well trimmed which is a first for me in bending servo linkages. The gap at the aileron hinges is too large due to poorness of measuring, or something. The biggest nuissance factor is that it is not really very flat in the center so the traditional "hold it to a stick using rubber bands" approach is a little more precarious than desired. This is exacerbated by the fact that only two of the 4 wing spars actually touch the bottom of the wing, so it rests on those two points, moderated by the unevenly-sized ribs. Really, I should have picked the two "best" ribs and put them in the center of the wing.

Anyways, I built a stick plane around it. The major innovation on this one is finally a decent motor linkage. I bolted the motor to a piece of 3/32" inch plywood, then I glued that plywood to the main balsa stick. I made notches in the plywood and glued other pieces of plywood along the sides of the stick and then they had a wide part at the end, a very strong physical construction:

And here is the whole plane:

Today is the first day warm enough to fly it. It was too windy. The elevator is (barely) too small. The ailerons seem to be too small but probably at higher speed (or with decreased differential) they would be sufficient. Also, perhaps because I had strong differential configured, there was no adverse yaw that I noticed. The motor (a 10g micro motor from gobrushless) is too small for how much the plane weighs. The wind was coming from my back and I was throwing it from the top of a hill, so there was a substantial downdraft that really overloaded the small motor (a constant struggle to maintain altitude), whereas on a non-windy day I know that this motor can totally hack it even with such a big plane.

But I think it is totally workable, and I look forward to getting a bigger motor for it and throwing it again on a slightly less windy day.


Also, there are some noteworthy flights with the IFO that happened previously... At the end of November I flew the IFO with a camera, doing lots of low passes over my head. Which is, by the way, good practice to catch it. So at the end of the flight I caught it, so you can skip to the end for the dramatic catch:

On the first day with a lot of snow in December, I put a tiny set of LEDs on the IFO and flew it at night. It was so bright (due to ambient light reflecting between the snow and the low clouds) that I didn't really have too much trouble seeing my plane. But it was doing an odd thing where under heavy throttle it would pick up a vibration and then literally eject its rudder (held in only by a little carbon fiber tension). Without its rudder, the IFO is a really graphic demonstration of adverse yaw. Anyways, I kept on flying it, trying to diagnose the problem. Eventually on a landing one of the glue/aramid joins between two carbon fiber rods came undone, and it ripped its covering hugely. I think the vibration was from the join working itself free (it is one of the two main joins between the motor mount and the main curved outer frame). Anyays, I credit that to the superglue being brittle on all of the cold days that I've flown recently. Then once the superglue started to crack, it formed a worst-case abrasion hazard for the aramid. A bit of a disappointment, but that's what I get for flying sub-freezing. I hope the IFO doesn't mind having a huge rip in it.

Motor shopping

And I want to take a moment to warn about They are a curious outfit. Their prices are out-of-this-world fantastic, and they ship fast. They're very very bad at keeping things in stock, though. Tired of using micro-scale motors with mini-scale planes, I just ordered a new motor+ESC from Subsonic Planes, a Motrofly 2205-1800 with an FM-10 ESC. The combo cost $50 there. This was a hard decision at first for me because a similar-sized motor+ESC combo from hobbypartz would cost as little as $15 -- for the $60 I wound up spending at Subsonic I could have ordered a whole plane's worth of parts from hobbypartz.

When looking at brands like Motrofly and GoBrushless I see a ton of text like "high-temperature wire" and brags about the strength of the magnets. I'm the kind of guy who ignores such pitches, though. I mean, if you have 5g of magnet plus 5g of wire plus 5g of aluminum bell, the remaining distinctions will be pretty minor, right?

I know it's hard to directly compare motor specs from a website. But I noticed that when I was comparing GoBrushless specs to the hobbypartz components, often the hobbypartz component would weigh more while having lower thrust specs. At first I shrugged this off but I happened to notice something while glancing at a photo on The motor I was looking at that day had curved magnets in it, which caught my eye. And I'm glad they did, because it had 3 curved magnets in it. If you look at some little motor from GoBrushless, you'll see on the order of a dozen little square magnets packed in there, filling the space so tightly you wonder that the thing can spin at all. But a lot of the motors on simply have HUGE gaps between the magnets, huge regions in which the stator's flux has nothing to push on.

So that is the answer to the riddle: How can a 1.7oz motor from hobbypartz have the same on-paper specs as a 1oz motor from Motrofly? Why does that 1.7oz motor only feel twice as powerful as the 0.3oz motor from GoBrushless? Because the hobbypartz motor has air where magnets belong.

They're crap, just garbage. I'm not saying they don't work, or that I'm particularly unhappy with the two hobbypartz motors I have bought so far, but if you want to stay on the happy side of the weight/thrust curve, don't waste your time. Accept that you'll pay 3x as much and go to a reputable vendor.

February 19, 2011

It is finally warm and the wind is coming from the right direction, all on the same day. Pretty good for February!

I can't say enough good things about this Motrofly motor+ESC. I was a little skeptical about buying a new ESC, since I already have more ESCs than operating motors...but I wanted the manufacturer's assurance that it was appropriately matched to the motor. The ESC has great fine throttle control, much better than the "Exceed" stuff. And it seems to have a great low-voltage warning feature. It just pulsed the motor off for less than a second. Enough time that there was no way I would miss it, but not significant enough to affect the flight any -- I came in for a landing and the battery was at exactly 7.4V. Which was important because the motor was so powerful that it seemed like I could have kept flying even if it had lost 80% of its power.

With the small 3S pack, at 1/3rd throttle it was flying fine, at 2/3rd throttle I had unlimited vertical, and at full throttle it picked up a weird and awful vibration. With the small 2S pack, it was plenty powerful (but not excessively so) for mild aerobatics. But it didn't weigh enough, the wind kept flopping it around. I put the heavy 3S pack on it, and suddenly it could hold level cross-wind flight without rolling this way or that. It was too light!!

My aileron model has lots of infelicities:

It has one great charm -- it doesn't have too much adverse yaw even though I set the differential fairly low.

Anyways, I flew it into the trees. I'm not entirely sure why. The motor mount was broken. So maybe it came off in midair from the odd vibrating mode at full throttle? But the motor wires were not twisted together (PFEW!!) so that doesn't really seem likely. I remember thinking "gee that's awful close to the trees and it's moving towards them as if pulled by a magnet!" and I commanded what I thought was a pretty expert request for a right turn (i.e., hard but not so hard that I thought I would defeat my control surfaces). I remember watching it roll right and then suddenly roll back to level!

I know the wind was a contributing factor and it's possible it just picked up the wrong kind of wind gust at just the wrong moment. It's also possible that when I pulled back to complete the turn, the fuz was so twisted relative to the wing that the elevator caused it to roll back to level. It really, to me, seems as if the flight experience that put it in the tree was exactly the same experience as I had when the motor was penduluming on the end of its wire. I was on very low throttle at the time, so maybe it really was just a case of the motor mount disintegrating in midair and the wires didn't twist just because they're relatively heavy compared to the throttle setting at the time.

This idea of the motor mount disintegrating is really concerning to me. I used a pretty indestructible design (a bigger version of the one I photographed last December). The only way I could see it breaking was if the 1/8" plywood itself were to snap. Of course, the devil is only too happy to oblige a sinner, so that's what happened! Considering this is both the strongest and the stiffest motor mount design I've ever used (by a fairly large margin), I'm disappointed that it still has an apparent vibratory failure mode.

I was very lucky, it hit about 50 feet up in the tree but lodged at 25 feet up. I went home and grabbed 20 foot of 3/4" PVC pipe. No luck!! It was just a foot or two beyond my reach. The tree in question was shimmyable, but all of its low branches were dead and the only shoes I have that seem to be able to grab a tree are crocs which are not so good at grabbing my feet. So I switched to tossing a 10' segment of PVC at it (like a javelin). After a few dozen tosses, I got pretty good. It only took about a dozen strikes against the airplane before it was on the ground.

I am not sure if the wing is totaled. It needs a lot of repair to its covering (I even managed to punch a hole in the covering with a failed hand-catch before treeing it). There are a few broken ribs and so on but it is not so bad. I think I can probably just add a couple balsa splints and then packing tape the hell out of it. I guess the deciding factor will probably be the servo lifetime. Especially if I keep throwing javelins at it. Now that I know the ailerons are not really built right (and that it needs more weight!), I do not consider it a big shame to do haphazard repairs until it is completely totaled.

I think when it gets a little warmer I might build a sheet foam plane. I think mostly I like foam because you can very easily experiment with design parameters (such as control surface area) at the flying field simply using a knife or some tape.