I built a two-sheet foam board flying wing from plans I found online, and it refused to fly. Four attempts. Endless CG changes. A dead motor and ESC. I seriously considered quitting the hobby. The real problem turned out to be something I'd stared at the whole time — and the fix took one evening. If your plane keeps nose-diving no matter where you put the battery, this write-up is for you.
| Attempt | What happened | What I blamed |
|---|---|---|
| 1 | Nose-dived straight into the asphalt driveway | CG + loose battery |
| 2 | Didn't crash immediately. Still nose-diving hard | CG (nose heavy?) |
| 3 | Technically a flight — super pitchy, barely controllable, hard crash | CG, again |
| 4 | Briefly flew, then unstable crash; electronics ripped off | Honestly? No idea anymore |
| 5 | Flew. Smooth, trimmed, low passes, even a (very slow) roll | The real fix: removed the under-cambered trailing edge + added reflex |
The plans looked simple: a two-sheet foam board flying wing. Cut here, glue there, wooden spar for strength, electronics in, done. The build genuinely did go smoothly — wings cut and joined, fins on, motor pod, servos wired for the elevons, control rods hooked up. It went so well I decided to spend a few extra hours on a paint job before it had ever flown. Blue base coat, white streaks, black tips. Because everything was going perfectly and it was obviously going to be the best-flying plane ever, right?
Then I launched it, and it nose-dived straight into the asphalt.
Every RC builder knows the reflex: it nose-dived, so it must be nose heavy — move the CG. I strapped the battery down better and moved the CG. Then moved it again. Flight three "kind of flew," which gave me just enough hope to keep digging, but it was so pitchy it was barely controllable.
I spent more time than I'd like to admit on the CG question, including several long conversations with our favorite virtual helper, Chad. First we decided the CG was way off and moved it forward. Then, after I described the wing shape in more detail, we decided it should go way back instead — at one point my battery was 7½ inches behind the leading edge, about halfway back on the chord, which the math said was crazy. Every hand launch ended the same way: release, nose-over, catch it or eat it. By the end I'd moved the CG in a full circle back to roughly where I'd started, and I didn't trust anything anymore.
The lesson hiding in there: when moving the CG in both directions doesn't fix a pitch problem, the CG probably isn't your problem. CG errors respond predictably — too far forward flies stable but dives and needs constant up-elevator; too far back is twitchy and stall-prone. A plane that pitches down hard at every CG position is telling you something aerodynamic is forcing the nose down. Mine was. I just couldn't see it yet.
Right when I was ready for another attempt, the motor and ESC quit entirely. The only spare ESC in my bin had no built-in BEC — meaning nothing to power the receiver and servos — so I soldered in a separate external BEC and put new connectors on the motor. It was a whole thing, but it worked, and honestly it's a skill worth having: an external BEC is a totally normal setup on bigger planes anyway. If you've never wired one, my free Wiring Designer has external-BEC hookups built in — it'll draw you the exact diagram, opto-ESC cases included.
After flight four ripped the electronics off the airframe, I stepped away for the weekend. Ran the numbers through another CG calculator — which said my original CG had been about right. So if the CG was right, what was forcing the nose down?
Then I remembered a clue: earlier, adding more reflex to the elevons had helped a little. I looked closely at the wing and finally saw it — the plans had built in an under-cambered section about 5 inches deep along the trailing edge. The rear of the wing curved downward. In effect, my wing was flying with flaps permanently deployed: all that downward curvature at the back of the wing produces a strong nose-down pitching moment, and on a flying wing there's no tail to fight back.
So I tore the trailing edge apart, spread it open, and removed the under-camber to make the back of the wing flat. Now air flowing off the wing runs straight onto the elevons instead of being deflected downward first. I re-rigged the control surfaces with a bit of reflex, did a quick repaint, and headed to the field. Flight five: after one pushrod tweak (my initial reflex was too much — it ballooned), it flew. Really flew. Three mistakes high, trimmed out, smooth low passes, and the slowest aileron roll I have ever seen in my life. But it worked.
Sight down the wing from the tip. On a conventional plane, some under-camber (a concave lower surface) is fine — a tail is there to balance the pitching moment it creates, and slow flyers use it on purpose. On a flying wing, downward curvature near the trailing edge is poison. Symptoms: the plane noses over on every launch regardless of CG, it gets slightly better when you add up-elevon trim, and it feels like it's fighting you nose-down at speed. If moving the CG doesn't change the character of the problem, check the airfoil, not the battery position.
Reflex is the opposite of camber at the trailing edge: the last bit of the airfoil (or the elevons at neutral) angles slightly upward. A cambered wing naturally wants to pitch nose-down; a conventional plane cancels that with its horizontal tail pushing down way back on a long lever arm. A flying wing has no tail — so the back of the wing itself has to do the tail's job. That little upward curl is the "tail," built into the wing. This is why every flying wing you'll ever see flies with elevons resting a few degrees above neutral, and why my under-cambered trailing edge — reflex's exact opposite — made the plane unflyable.
On a conventional plane, the tail sits on a long moment arm, so pitch stability is generous — the "CG one-third back" rule of thumb has plenty of slop in it. A flying wing's entire pitch authority lives in a few inches of elevon at the back of the same surface generating the lift. The CG typically needs to sit around 15–20% of the mean aerodynamic chord, and the flyable window is a fraction of an inch on a plane this size. That's why the CG rabbit hole was so believable — flying wings really are CG-sensitive. Mine just happened to have a second, bigger problem stacked on top. When you design your next wing, get the CG computed rather than guessed — the Airplane Designer does it from your dimensions in about a minute.
Rock bottom on a project is real — you pour hours in and it just won't work. But every one of those failed flights taught me something I now can't un-know: read the airfoil before you blame the CG, test-glide before you paint, and keep a spare ESC with a BEC in the bin. The next flying wing might even fly on the first try. Might.
