Deep Dive: Harmonic Form

Provenance (2026-09-29): read/checked in full for this pass — research/veridian/WMA13-verified-facts.md (grepped and read in full for "harmonic", "R cos/R sin", "2.3", the spec table in §1, the formula-booklet contents in §3, the general marking conventions in §4, and the harmonic-form.ts audit log in §9); content/veridian/curriculum/wma13/harmonic-form.ts (the full 993-line lesson, read start to end, to identify what is already taught and avoid duplicating it); content/veridian/technique-hub/wma13/codex-formula-and-method-bank.md, forge-exam-technique-guide.md and signal-frequency-and-boundaries.md (grepped for harmonic-form coverage); content/veridian/question-bank/wma13.ts (grepped — contains no harmonic-form items, so no overlap risk there); content/veridian/curriculum/wma13/compound-and-double-angle-formulae.ts (headings only, to confirm the double-angle derivation itself is out of this document's scope). All worked numeric micro-examples below were computed and independently checked with Python (exact math arithmetic, substituting the result back into the original expression) before being written in; none reuse the specific (a, b) pairs already used in harmonic-form.ts (3-4-5, 7-24-25, 8-15-17, 5-12-13, 2-5-√29) or its MCQs (5-12-13, 7-24-25, 2-5-√29 again).

#CODEX Deep-Dive: Harmonic Form — a cosθ + b sinθ ≡ R cos(θ∓α) / R sin(θ±α)

#What this document is for

harmonic-form.ts already derives the identity in full narrative form, drills it through a worked chain, a chain-drill and a marked solution, and carries seven real, evidence-grounded traps. codex-formula-and-method-bank.md already gives the mark-scheme skeleton. Neither lays out, in one place, the complete sign-case structure that decides which of the four target forms actually applies to a given (a, b) — the lesson teaches the two cases it needs for its own worked examples, not the full picture — and neither states, plainly and in one place, exactly how thin the real past-paper evidence for this specific sub-topic is. This document does both, adds one genuinely uncovered sign-case with its own micro-example, and closes with a precise synthesis of every real exam angle this topic has actually been tested from.

Spec 2.3's own wording (transcribed verbatim in the verified-facts bank, §1) is the frame for everything below: "use of formulae for sin(A ± B), cos(A ± B) and tan(A ± B) and of expressions for a cos θ + b sin θ in the equivalent forms of r cos(θ ± α) or r sin(θ ± α)", with guidance adding "to include application to half angles" and that students should be able to "solve equations such as a cos θ + b sin θ = c" in a given interval.

#1. The identity is a method, not a formula on the sheet

The compound-angle formulae cos(A∓B) and sin(A±B) are printed in the exam's Mathematical Formulae and Statistical Tables booklet (verified-facts §3, from the booklet PDF directly). The harmonic-form identity itself — R = √(a²+b²), tanα = b/a — is not printed anywhere in the P3 section of that booklet. codex-formula-and-method-bank.md already states this plainly: there is no boxed line to look up; R and α are rebuilt from the booklet's own compound-angle formula by expanding the target shape and matching coefficients, every time. That derivation — expand, match coefficients of cosθ and sinθ separately, square-and-add for R via cos²α+sin²α≡1, divide for tanα — is done in full in harmonic-form.ts's own mechanism block and is not repeated here.

#2. The sign-case structure in full — the part not written down anywhere else

Expanding each of the four target forms and matching coefficients gives four different pairs of conditions on Rcosα and Rsinα (or Rsinα and Rcosα for the sin-led forms), and for a first-quadrant α — the range every real question and every existing MCQ in this course uses (0° < α < 90°, or 0 < α < π/2 in Q7's own wording) — both matching values must come out positive. Working through all four:

harmonic-form.ts already builds and drills the first two rows (its worked chain uses a=7,b=−24 for the cos(θ+α) row) and the fourth row (its chain-drill uses a=8,b=15 for sin(θ+α)). The third row — a < 0, b > 0, giving sin(θ − α) — has no worked example anywhere in the existing lesson or its MCQs, checked by reading every prequestion, chain, drill, marked-solution and MCQ choice in the file. It is a genuinely uncovered corner of the same sign logic, not a new topic:

Micro-example (fresh numbers, self-verified). Take a = −3, b = 4. Row three needs Rcosα = b = 4 and Rsinα = −a = 3, both positive — so R = √(3²+4²) = 5, tanα = 3/4, α = 36.87° (2 d.p.). Check: 5sin(θ − 36.87°) expands to 5sinθcos36.87° − 5cosθsin36.87° = 5(0.8)sinθ − 5(0.6)cosθ = 4sinθ − 3cosθ = −3cosθ + 4sinθ. Substituting θ = 0°, 30°, 60°, 90°, 120° into both −3cosθ+4sinθ and 5sin(θ−36.87°) reproduces the same value in every case to five decimal places.

The fourth sign combination — a < 0 and b < 0 together — is not reachable by any of these four forms once α is restricted to the first quadrant, which is the restriction every real question located in this research actually imposes (Q7's own "0 < α < π/2"; every existing MCQ and worked example in the lesson uses 0° < α < 90°). Checking a = −3, b = −3 against all four rows above gives a negative value for at least one of the two matching quantities in every row (confirmed by direct substitution). No located WMA13 question tests this fourth combination, and this document does not construct one to fill the gap — per this course's standing rule, the evidence simply runs out here.

A second, genuinely useful observation from the table: the first and fourth rows share the same sign condition (a > 0, b > 0) but are different identities with different α. The same expression can legitimately be built in either family depending on which one the question (or a later step) needs. Micro-example: for a=1, b=2 (R=√5 in both cases), the cos-led form gives tanα₁ = b/a = 2, α₁ = 63.43°; the sin-led form gives tanα₂ = a/b = 1/2, α₂ = 26.57°. These are complementary (63.43° + 26.57° = 90°, exactly), and both √5cos(θ−63.43°) and √5sin(θ+26.57°) reproduce cosθ+2sinθ at every θ checked — the same wave, described two ways.

#3. The half-angle sub-case (spec guidance's own "application to half angles")

Spec 2.3's guidance names this explicitly, and the only real past-paper question located that tests harmonic form head-on happens to be exactly this sub-case: Q7's model is H = 24/(3 + cos(½t) + 4sin(½t)) (verified-facts §5, re-extracted from the primary mark scheme). Because the wave's argument is ½t rather than a bare θ, recovering t from a solved (½t ± α) needs an extra doubling step on top of the usual "convert the shifted angle back" — a fact verified-facts records as newly confirmed by this specific question and not previously documented elsewhere in this bank.

Micro-example (fresh numbers, self-verified, distinct from Q7's own 24/(3+cosx+4sinx)). Let H = 12/(5 + cos(½t) + 2sin(½t)) for 0 ≤ t ≤ 13, worked throughout in radians — the same unit the real Q7 mark scheme uses for α and for the final recovered t, with no degree conversion at any step (verified-facts §5's own transcription: "α = awrt 1.326" and "awrt t = 6.09", both radian-native). Here a=1, b=2, so R=√5 and α=arctan 2 = 1.1071 rad (matching row four's own R, reused deliberately to show the same construction feeding a different application). Solve H = 3.5: this needs 5 + wave = 12/3.5 = 3.4286, so the wave itself equals −1.5714, giving cos(½t − 1.1071) = −1.5714/√5 = −0.7028, so ½t − 1.1071 = ±2.3496 rad (2 s.f. on the last digit throughout). Doubling to recover t: t = 2(1.1071 + 2.3496) = 6.91 on the positive branch, and, adding a full period of 2π to the negative-branch value before halving back, t = 2(1.1071 − 2.3496 + 2π) = 10.08 on the second branch — both inside [0, 13], both independently confirmed by substituting back into the original denominator to recover H = 3.5 exactly. The mechanism is identical to Q7's own part (c); the numbers are not.

#4. The optimisation sub-case: harmonic form inside a fraction

Q7's part (b) is the only real evidence for this angle, and it records a genuine, confirmed point of confusion: minimising a fraction of the shape k/(c + wave) requires maximising the wave, not minimising it, because a larger denominator makes a positive fraction smaller. The examiner report's own words for this: "the denominator had to be a maximum."

Since Rcos(θ−α) ranges over exactly [−R, R], a denominator c + Rcos(θ−α) with c > R ranges over [c−R, c+R] and never reaches zero. Micro-example: for H = 20/(6 + cosθ + 2sinθ) (R=√5, c=6>R so the denominator stays positive throughout), the denominator's maximum is 6+√5 ≈ 8.236 at θ=α≈63.43°, giving H's minimum ≈ 2.428; its minimum is 6−√5 ≈ 3.764 at θ=α+180°, giving H's maximum ≈ 5.314 — both checked by direct substitution. No calculus is needed for either extreme, which is exactly the payoff spec 2.3's guidance is naming when it pairs harmonic form with maximum/minimum problems.

#5. The other named application: solving a cosθ + b sinθ = c on an interval

This is spec 2.3's other explicitly named payoff, and it is the one harmonic-form.ts already teaches in the most depth — its worked chain, chain-drill, marked solution and first three trap-taxonomy entries (missing-second-solution, degrees-radians-mismatch, extra-out-of-range-solutions) all sit here, each grounded in its own multi-series citation (verified-facts §5). None of that is repeated in this document. One fresh check, purely to confirm the mechanism generalises past the lesson's own five (a,b) pairs: for 20cosθ + 21sinθ = 15 on [0°, 360°), R=29 (a 20-21-29 triple), α=46.40°, giving cos(θ−46.40°)=15/29=0.5172, θ=105.25° or θ=347.54° (both independently confirmed to reproduce 15.000 on substitution) — the same two-solutions-per-period structure the lesson's own traps already document, on numbers the lesson never uses.

#6. Synthesis: every real exam angle this topic has actually been tested from

Stated plainly, because the hard rule against inventing likelihood is only honest if the underlying thinness is stated too: across the eight series read in full and the eleven series' mark schemes checked for this course's WMA13 research, exactly one real past-paper question — October 2020, Question 7 — was located that tests spec 2.3's r cos/r sin(θ±α) construction head-on, rather than as a buried step inside a differentiation or integration question (verified-facts §5, §9). That is not a gap in this pass's research; it is the verified-facts bank's own explicit finding, re-confirmed rather than newly discovered here.

From that one 9-mark question, three distinct, separately mark-scheme-confirmed exam angles exist:

1. Constructing R and α to a specified exact form and unit — part (a), 3 marks. The construction marks themselves, not just a later solved angle, can require an exact surd for R ("do not allow decimals for this mark") and the angle in the unit named ("the degree equivalent α = awrt 75.96° is A0") — both already built into the lesson's own trap taxonomy and its hf-mcq-7.
2. The half-angle embedding and its ×2 recovery step, plus the maximise-the-wave-to-minimise-the-fraction direction check — part (b), 2 marks — covered in §§3–4 above with fresh numbers.
3. Solving cos(βt±α)=k for the original variable, where β=½ — part (c), 4 marks — the same "solve on an interval" payoff as §5, but applied to a rescaled argument rather than a bare θ.

Beyond this one question, the cross-cutting marking conventions that verifiably apply whenever this topic is examined are well-evidenced independently of harmonic form specifically: "quote the formula before substituting" is corroborated across five or more series for learnt-formula method marks generally (verified-facts §4), and the paper's calculator-methods rubric ("solutions relying entirely on calculator technology are not acceptable") is confirmed on specific questions across multiple series and already embedded in the lesson's hf-mcq-5. These transfer to harmonic-form questions because they are general P3 marking principles, not because a second harmonic-form-specific question has been found to confirm them directly.

No second series-independent confirmation of a harmonic-form-head-on question exists in this research. A future pass that locates one should extend this section, not this document's confidence in the sign-case and half-angle mechanics above, which are self-verified mathematics rather than exam-frequency claims.

#What this document deliberately leaves to the existing files

The full narrative derivation, the 3-4-5 concept ladder, the diagram, the 7-24-25 worked chain, the 8-15-17 chain-drill, the 3-4-5 marked solution with its wrong-path exemplar, all seven trap-taxonomy entries, all seven MCQs and all eighteen flashcards live in content/veridian/curriculum/wma13/harmonic-form.ts and are not reproduced here. The mark-scheme skeleton for both the construction and the fraction-optimisation hence-part lives in codex-formula-and-method-bank.md. The frequency table (5 series, 8–9 marks) lives in signal-frequency-and-boundaries.md. This document's job was the sign-case structure, the one uncovered corner of it, the half-angle mechanism with fresh numbers, and an honest statement of how much real past-paper evidence this specific sub-topic actually has.