Spaceclam,

I think you got the gist of what I was saying. Looking at my diagram, it seems that making phi negative, i.e. placing the cg slightly above the fans, would have a stabilizing effect. Note, however, that I don't mean above the thrust line, as the thrust line is not what contributes to the stability problem. I looked at your craft in a cross section, and broke it's mass into three parts. The largest, and central part is where you keep the battery and radio and whatnot, and then two smaller outboard masses that approximate the weight of whatever is outboard, i.e. fans, motors, and structure. If you look at the moments around the cg, which is a good way to do it, you can see that the two fans cancel eachother as long as their thrust is the same. If we assume that the main mass, i.e. battery and radio are at the cg, which is not a bad assumption as they are probably much heavier than the rest of the craft, then they contribute no moment. If we assume that the main mass is slightly below the cg, which it would be reality, then it contributes a slight righting moment when the craft is tilted. This will have a stabilizing effect. Looking at the outboard masses, when the craft is tilted, they destabilize it. The easiest way to see this is to draw a tilted T representing a cross section of your craft that has tilted. Make the vertical portion much longer than reality in order to better see the effect. Draw two gravity force vectors at the outer tips, they will be the same magnitude because hopefully you made your craft equally weighted. A trick to see their moment is to extend those lines downward past the cg at the bottom of the T. You can then measure the moment arm of each force as the distance between the downward extending line and the cg where the line you measure is perpendicular to the downward extending line and goes to the cg. It is obvious that the moment arm from the downward tilting side is larger than the upper, and since the force magnitude is the same, the downward portion contributes a larger tilting moment, which augments the problem. However, as I mentioned before, this effect is probably small. The outboard weight will be small compared to the rest, and since in reality your main weight is probably a little below the cg, it may cancel the effect. Further, phi is very small on your craft, I think about 10 degrees or so. It will have some effect though, and moving the cg up towards neutral may help.

I think your main problem is that you have a diverging system. Around neutral it's quite unstable, and needs to tilt a good bit before it's slight stability kicks in to correct. Then it will overcorrect because of its lack of stability around the neutral point, and then this process repeats until, plop. The thrust made by each fan will have a huge effect on the moments of the craft. When these oscillations happen, are you trying to control it? It may be you overcorrecting because it responds too slowly. Maybe give a try and try to be very gentle, letting it correct very slowly, and as soon as it begins to correct taking out your correction.

I saw a demonstration of a similar craft once. They incorporated micro angular sensors and rate gyros and accelerometers in theirs, and I think a small computer chip that kept it stable. When they flew it it was with a RC transmitter, so I'm not sure how much of their electronics was doing the work, if any. The presentation was in German so I have no clue what stage they were in with their development! I guess just keep playing with it, and try to figure out a way to control the fans together.

Good luck,
Jamie