Fin Cant Angle & Roll Rate — theory

Overview

Fin warpage makes every fin-stabilised rocket roll at random and adds random pitch/yaw moments. Setting the fins deliberately serves two purposes:

• Dispersion. A well-chosen roll-rate profile averages out the impact-point errors caused by body-fixed misalignments, especially thrust misalignment.
• Loads. Fin misalignments add angle of attack, and with it structural load on the fins and the whole airframe.

The process is: measure the as-built cant angles, null each panel’s misalignment normal force, then set all panels to one common cant angle that produces the planned roll rate. This page derives the roll derivatives and that common angle.

Strip theory

Each fin panel is divided into streamwise strips of span dy, each with the same normal-force slope c_{N\alpha}. There is no fin–body interference and no spanwise variation of lift slope from tip losses — the simplest model that accounts for planform geometry. Thin-airfoil theory makes c_{N\alpha} depend only on Mach number; using a panel-averaged value captures root and tip losses approximately.

A strip at radius y from the body centre line, on a fin canted by \delta and rolling at rate p, sees the local angle of attack \delta-py/U. Its normal force and rolling moment are

For a straight-tapered fin running from the body radius R to the tip radius b, the chord is linear:

Integrating over the panel gives the rolling moment of one fin,

where the two planform integrals are

The spanwise location of the normal force due to cant is

Roll stability derivatives

With reference area S=\pi R^2 and reference length d=2R, the rolling-moment coefficient is C_L=L/(qSd)=L/(2\pi R^3q). The roll-driving derivative of a single fin is

and the roll-damping derivative of all N fins, per unit pR/U, is

M is negative, so C_{Lp} is negative: rolling always generates an opposing moment.

Fin adjustment and the cant angle for a roll rate

Many adjustments are possible, but the preferred one sets every panel to the same cant after nulling its normal force. For three or more panels this cancels the combined pitch and yaw moments and gives all panels the same cant-induced loads. In the field the as-built cant is measured with a laser level aligned with the body axis: the spot’s height on a screen at the nose tip shows the cant, which is then adjusted to the desired value.

In steady roll the driving and damping moments balance, L=0, so the common cant for a roll rate p at speed U is

This is the steady-state cant. The memo’s expression also includes the lag caused by roll inertia during the boost:

where h is the burnout altitude and L the launcher length. The memo considers this good enough for small rockets up to about 8 inches in diameter; for larger ones it suggests fitting the measured cant data, the simplest fit being a helix \delta(y)=(-\delta_1y_2+\delta_2y_1)/(y_1-y_2)+y(\delta_1-\delta_2)/(y_1-y_2) through two spanwise measurements.

Choosing the roll rate

For multistage rockets, set the last stage first and work down. For a single stage, the mean cant \bar\delta_F for a given roll wavenumber \lambda_R is

Whole roll cycles. To minimise dispersion from body-fixed perturbations, the booster should complete a whole number M of roll cycles between launcher exit and burnout:

which is solved for \bar\delta_F. M should be at least 2 or 3; more is better. This choice matters mainly for the booster.

Fraction of the pitch wavenumber. Alternatively, for any stage, choose \lambda_R as a fraction of the pitch wavenumber \lambda_P. Body-fixed perturbations produce an angle-of-attack response proportional to \lambda_P^2/(\lambda_P^2-\lambda_R^2), so roll erodes static stability; as long as \lambda_R stays below 30–40 % of \lambda_P the damage is small. A higher roll rate still reduces dispersion. The memo’s example sets the burnout roll rate to 0.3 × the burnout pitch natural frequency, about 5 roll cycles to burnout.

Roll through the boost

The roll does not start at its final rate. The rocket must first spin up its own roll inertia. Hoult’s “Small Sounding Rocket Boost Phase Roll Rate” (2013, eqs. 2b and 5b–5e) treats a boost with constant acceleration, U^2=2ah, and one representative air density. The steady roll wavenumber (roll angle per distance flown) of a common cant \delta is

Here x is the height flown and L the launcher length. J is negative because C_{Lp}<0, so the roll lags the steady value and catches up as the rocket climbs. The roll angle since launcher exit is \lambda_{R0}\left[x-L+k\ln\frac{x-L-k}{-k}\right]. The “Roll through the boost” plot shows both the lagging and the steady roll rate against speed, and the roll turns. The target cant includes this lag: it is solved with J at burnout.

How fast may it roll? The verdict

A rolling rocket turns body-fixed misalignments into a rotating disturbance. The angle-of-attack response grows by \lambda_P^2/(\lambda_P^2-\lambda_R^2) as the roll wavenumber approaches the pitch wavenumber (Fin Setting Theory rev. 3, p. 6). The memo advises that the damage is small while \lambda_R stays below 30–40 % of \lambda_P. The calculator reads that guidance as bands (DECISIONS D31):

• SAFE up to 30 % of the pitch frequency at burnout;
• MARGINAL up to 40 %;
• UNSAFE above 40 %. Past 100 % the roll locks onto the pitch frequency: see the Roll Lock-In calculator.

The memo also asks for at least 2–3 whole roll turns between launcher exit and burnout, to average out misalignments. That is shown on the turns card. The solved values are:

• the cant for your target;
• the largest cant that keeps the roll at 30 % of the pitch frequency;
• the smallest cant for 2 turns.

Assumptions and limits

• Linear strip theory: no fin–body interference, constant c_{N\alpha} along the span, small angles.
• Straight-tapered fins; the root is at the body surface.
• In practice, errors in measuring misalignment and cant tend to dominate everything else, so one usually only aims for a desired roll rate.
• The calculator warns above 3° of cant; that threshold is a rule of thumb.

Worked example (Gold Rush III)

R=3.125 in, b=5.5 in, c_R=10 in, c_T=2 in, 4 fins, c_{N\alpha}=3.5/rad:

1. A=20.526 in, C=3.368, L_\delta=24.292 in², M=-100.75 in³, Y_{CP}=4.049 in.
2. C_{L\delta}=3.5(2.375)(24.292)/(2\pi\,3.125^3)=1.0531 /rad (memo); C_{Lp}=-5.5906.
3. For 1.24 Hz at 1037.2 ft/s: p/U=6.26\times10^{-4} /in and \delta=0.1488° (steady state). With I_{xx}=0.22 slug·ft², \rho=0.0023, h=2000 ft and L=18 ft, J=-0.299 and the memo’s cant is 0.193°.
4. For 4 roll cycles to burnout: k=-592.1 ft, bracket 1111.9 ft, \delta=0.448°.

Equations ↔ code

Verification: tests/cant-angle.test.js reproduces the reference example to 10⁻¹², checks L_\delta, M and the memo’s C_{L\delta} against direct numerical strip integration, and checks that the lag form reduces to the steady state when J\to0.

Notes on the method

• Roll-driving derivative. C_{L\delta}=c_{N\alpha}(b-R)L_\delta/(2\pi R^3), as in the memo. Direct numerical strip integration confirms it (1.0531 /rad for Gold Rush III). The steady-state cant for a roll rate does not depend on C_{L\delta}.
• Roll-inertia lag. The calculator shows the steady-state cant and the memo’s form with the (1-J) factor side by side.

References

1. C. P. Hoult, “Strip Theory for Setting Fin Angles (rev. 3),” RS&T memo, 22 May 2014. View in library
2. C. P. Hoult, “Small Sounding Rocket Boost Phase Roll Rate,” RS&T memo, 28 September 2013. View in library
3. Source file: CANT ANGLE3.xls; slides: Fin Adjustment Tips.ppt (rsandt.com).