- Two or more canards ahead of the CG: This is the most well-known/well-proven approach. I used to do a lot of RC slope soaring with flying wings, so I'm familiar with the potential for 2 control surfaces to control pointing in all 3 directions. Control roll to align the canards with the desired pitch axis and then pitch up. The canards can be directly connected to the servos, which is simple and avoids backlash, but is not compatible with narrower airframes unless the canards are offset along the length of the airframe (which might be o.k.). Linkages to allow the servos to be stacked vertically are possible, but add complexity and maybe some backlash. If the canards are offset from the fins to prevent control reversals during the flight, they are in the way of tower rails, and so rail guides (possibly detachable) would be required. Canards add drag throughout the flight, and aren't effective during the earliest part of the boost while the rocket is moving slowly.
- Full-airframe thrust vector control (FATVeC): This is what I have in mind for steering in a 38mm airframe. Connect the upper and lower halves of an airframe with a central pivot that takes all the axial loads from the boost, and then use 2 servos to drive pitch/yaw lead screws to tweak the alignment between top and bottom parts of the airframe. This is for pitch/yaw only; something else is needed for roll control. The advantages here are that there is almost no additional drag, it produces control torques at any velocity as long as there is an acceleration, and also aerodynamic steering when the rocket is moving. No interference with the tower. The downsides are a more complicated and heavier mechanism, and it's unproven whether the servos could be fast enough and strong enough. Here's a photo of some hardware I bought to play with. A 6mm lead screw, and two each of servo clamping couplers and lead screw nuts from Servo City, and two of McMaster's smallest ball rod ends. Pen for scale.
The lead screw and nuts are very smooth, low-friction, and free of backlash. Impressive value since all 5 parts combined for about $26 plus shipping.
Some mini servos that have a chance of fitting into a 38mm airframe are on their way. This idea also requires low roll rates or the servos will have no hope of keeping up. Which brings me to the half system: - Roll control with one servo for 2 canards: @Finicky inspired this idea with his retractable canards. Similar to Finicky's system, the angle of attack of the control surface is fixed but the amount of the control surface that is exposed is variable. The idea here is to put one flat, angled control surface that is directly connected to the servo output and entirely inside the airframe when the servo is centered. When the servo is active it sticks the surface out a slot on one side or the other to provide roll torque (and some pitch as a side effect). To minimize undesired pitch moments, this would work best near the CG of the rocket, which is convenient anyway because that's where the FATVec system would be. If the servo is small enough, then the control surface can be centered in the tube and overlapping the servo. Otherwise, it would need to be ahead of or behind the servo. Sorry I don't have a sketch for this yet.
- Active 2-stage coupler: I'm in the early stages of planning for a 54mm/38mm 2-stage rocket for BALLS next year that would be designed to go over 100k before I get my L3 cert. The 38mm/54mm coupler presents an opportunity for a different variation on the FATVeC idea. The 38mm rocket's motor is the inner part of the coupling. It's surrounded by 38mm airframe tube connected to the end of the 54mm airframe for the stage coupling. If that 38mm airframe can pivot at the front end, then there would be some nice leverage at the back for 2 servos to provide pitch and yaw actuation. I think I could get two servos side-by-side in the airframe and move the end of the tube with simple ball end links. This would only add about 2" to the length of the airframe, not counting the roll control that you'd also need. I'm not sure about the pivot at the front end of the coupler yet, but I'm thinking about something like a 2" derlin ball bored out for the 1.5" airframe and riding inside a 3d printed cup.
- If the FATVeC system turns out to be too fragile or weak, a more robust approach that would work while the rocket is accelerating and decelerating is to move around some ballast weight inside the airframe rather than the whole airframe. I suspect the mass to get decent control authority would be substatial, and you wouldn't have the aerodynamic control that you would with the FATVeC approach, so I have put this one on the back burner for now.
Last edited: