RLC_43.pdf — technical review / szakmai értékelés

Technical review of Váltakozó áram R, L, C elemekkel (RLC_43.pdf, Csepi Kálmán, 2026-06-02)

Scope. The full 39-page PDF was read. Every numerical example was independently recomputed in Python (all impedances, phase angles, the t₁ = 3 ms instantaneous-value checks, the resonance/bandwidth/Q tables, and every filter design task), every derivation and appendix identity was checked step by step, and each finding below was then re-verified against the exact wording (and, where needed, a 200 dpi render) of the cited page. Page numbers refer to the PDF pages; the document carries no printed page numbers.

This review reports mathematical content only — errors in derivations, identities, results, physical reasoning, conventions that affect correctness, and figures that contradict their own parameters. Spelling, grammar, punctuation, decimal-separator and unit-casing typos are deliberately not listed, at the author's request.

Standard parameters. Time domain (chapters A2–A8): ug0 = 15 V, f = 50 Hz, R = 100 Ω, L = 200 mH, C = 15 µF, t₁ = 3 ms. Filters (A9, B): L = 50 µH, C = 2 nF, R = 5/10/20 Ω (series) or 500/1000/2000 Ω (parallel), f0 ≈ 503 kHz. The document uses the sine reference u(t) = U0·sin(ωt) throughout.

Bottom line. The geometric phasor/triangle method is sound and most results are correct (audit trail at the end). There is one flat error in a headline result — the appendix derives sin′(x) = −cos(x) (§ME‑1) — plus a cluster of genuine mathematical errors in the filter algebra (A9/1–A9/4, B3) and in the time-domain power/phase discussion (A4–A8). None of the filter errors change the final design numbers, which are independently correct; they are localized slips and unjustified steps in the derivations and figures.


1. Major mathematical errors

ME‑1. sin′(x) = −cos(x) (p. 34, D7/4)

As printed (bold, underlined), at the end of the sine-derivative derivation:

…=sin(x)*1*0-cos(x)*1=-cos(x) . Eszerint sin′(x)=-cos(x)

The derivative of sine is +cos(x). The result has the wrong sign and contradicts the document itself in three places:

Root cause: the line "cos(Δx)−1 = cos(Δx)−cos(0) = 2·sin(Δx/2)·sin(Δx/2)" drops the leading minus (correct: −2·sin²(Δx/2)), and a spurious minus is then carried onto the cos(x)·sin(Δx)/Δx term. Correct derivation:

sin(x+Δx)−sin(x) = sin(x)[cos(Δx)−1] + cos(x)·sin(Δx); divide by Δx and let Δx→0: sin′(x) = sin(x)·0 + cos(x)·1 = +cos(x), since [cos(Δx)−1]/Δx → 0 and sin(Δx)/Δx → 1.

ME‑2. Current placed on a cosine reference; uR(t₁) applies the projection twice (p. 12, A8)

As printed (series-RLC summary and its numeric evaluation):

i(t)=I0*cos(ωt) … i(t1)=I0*cos(ωt1)=83,45*cos(2π·50·3e-3)mA = 49,05 mA uR(t1)=i(t1)*R*sin(ω·t1)=49,05 mA*100Ω*sin(2π·50·3e-3)=6,75059V

Two errors:

  1. Wrong reference. The document declares and uses the sine reference i(t) = I0·sin(ωt) everywhere else (p. 3; A5 "iR(t)=i0·sin(ωt)"; A6 "i(t)=I0·sin(ωt)=70,7mA·sin(54°)=57,19mA"). Under it, i(t₁) = 83,45·sin(0,9425) = 67,5 mA, not 49,05 mA. The line is also self-contradictory: if i(t)=I0·cos(ωt), then uL(t)=uL0·sin(ωt+π/2)=uL0·cos(ωt) is in phase with the current — impossible for an inductor.
  2. Double projection. uR(t₁)=i(t₁)·R·sin(ωt₁) applies sin(ωt₁) a second time. Evaluated literally: 49,05 mA·100 Ω·sin(0,9425) = 3,97 V, not the printed 6,75 V. The printed 6,75 V is in fact I0·R·sin(ωt₁) = 83,45·100·0,809 = 6,75 V, i.e. the correct sine-convention value — so formula, operand and result are mutually inconsistent.

Correct: uR(t₁) = I0·R·sin(ωt₁) = 6,75 V with i(t₁) = 67,5 mA. Everything downstream (uL, uC, ug, the KVL check) is otherwise consistent.

ME‑3. The band-stop "plottable" dB formula drops the sin (p. 20, A9/4)

As printed — first the correct definition, then the "plottable" version:

A dB-beli erősítés-csillapítás: G′dB = 20*log{sin(arctg[1/A*(x−1/x)]} Ezzel a G′dB=20*log(abs{arctg[1/A*(x−1/x)))} már ábrázolható.

The second form is 20·log|arctg(y)|, a different function from 20·log|sin(arctg(y))|. Numerically, with y = (1/A)(x−1/x):

pointcorrect 20·log|sin(arctg y)|printed 20·log|arctg y|
band edge (y = 1)−3,01 dB20·log(π/4) = −2,10 dB
far from f0 (y → ∞)→ 0 dB20·log(π/2) = +3,92 dB

The plotted figure dips to −∞ at resonance and returns to 0 dB — i.e. it was drawn with the correct sin version; the printed plottable formula is wrong.

ME‑4. x − 1/x = ±√B instead of ±B (p. 24, B3), with a mis-placed square

As printed (parallel band-pass, −3 dB band edges):

1 + [1/B*(x−1/x)]² = 2 ; (1/B)(x−1/x)² = 1 ; x − 1/x = ±√B

From 1 + [(1/B)(x−1/x)]² = 2 one gets [(1/B)(x−1/x)]² = 1 → (1/B)(x−1/x) = ±1 → x − 1/x = ±B. The middle step wrongly moved the square onto (x−1/x) alone and then took a spurious square root. The very next line uses the correct x₁−1/x₁ = −B and x₂−1/x₂ = +B, so the ±√B line is internally contradicted. (Numeric check for R = 500 Ω, B = 0,3162: x−1/x = ±B gives x₁ = 0,854, x₂ = 1,171 with x₁·x₂ = 1 and G(x₁) = 1/√2; x−1/x = +√B = 0,562 gives x = 1,32 where G = −6,2 dB, not the −3 dB point.)

The same square-misplacement appears one page earlier (p. 23): "GdB = −10·log{1 + [1/B·(x−1/x)²]}", where the square belongs on the whole bracket [1/B·(x−1/x)]² (the G expression printed just above it has it correctly).

ME‑5. Sign error in the pure-inductance instantaneous power (p. 8, A4)

As printed:

Esetünkben. P(t)= uL0*iL0*1/2*[cos(π/2)sin(2ωt)] = ueff*ieff*[sin(2ωt)]

With x = ωt+π/2, y = ωt, the identity sin(x)sin(y) = ½[cos(x−y) − cos(x+y)] gives ½[cos(π/2) − cos(2ωt+π/2)]. Since cos(2ωt+π/2) = −sin(2ωt), the bracket is [cos(π/2) + sin(2ωt)] = +sin(2ωt), so

P(t) = +ueff·ieff·sin(2ωt).

Both the bracket and the final result have the wrong sign. (Verified: sin(ωt+π/2)·sin(ωt) = +0,2129 at t = 0,7 ms = +½·sin(2ωt); the printed −½·sin(2ωt) = −0,2129. Physically P > 0 just after t = 0, when the inductor absorbs energy.) The conclusion "no average power on the inductance" is unaffected (mean of sin 2ωt = 0). Note the analogous series-RL power on p. 9, P = ug0·i0/2·[cos φ − cos(2ωt+φ)], is done correctly.


2. Further mathematical errors and unjustified steps

ME‑6. "At ω = ω0 the phase is zero and the attenuation is 0 dB" (pp. 13–14, A9/1)

ω=ω0 esetére a fázisszög nulla, a csillapítás 20*log(1)=0.

At the corner frequency x = 1: φ = arctg(1) = 45° and G = 1/√2 = −3,01 dB, not 0° and 0 dB. The statement is true only for ω ≪ ω0. It directly contradicts the same chapter's own "ω=ω0-nál ez +45°" and "x=1 esetén … −3,01 dB."

ME‑7. −20·log(0,7071) = −3,01 dB (p. 14, A9/1) — sign

vagyis −20*log(0.7071)=−3,01 dB

Evaluated, −20·log₁₀(0,7071) = +3,01 dB. The correct expression is +20·log₁₀(0,7071) = −3,01 dB (equivalently −20·log₁₀(√2) or −10·log₁₀(2)). Related: the same passage calls G(1) = √2/2 the "hiba" (error); √2/2 = 0,7071 is the transmitted level (the gain at cutoff), and the loss relative to unity is 1 − 1/√2 ≈ 29,3 %.

ME‑8. Asymptote through (0;0) and "second quadrant" (p. 14, A9/1)

Pontjai: (0;0), (10;−20), (100;−40), (1000;−60) stb. … Ez egy −20dB/dekád meredekségű egyenes a második síknegyedben!

The line GdB = −20·log(x) passes through (1;0), not (0;0) — x = 0 does not exist on a logarithmic axis. And for x > 1 with GdB < 0 the branch lies in the fourth quadrant, not the second. (The other three listed points are correct.)

ME‑9. Missing arcsin (p. 15, A9/2)

És φ = arcsin(−XC/Z) = −1/√(1+x²)

The right side is sin(φ), not φ: the equation equates the angle with its own sine. At x = 1 the RHS is −0,707 (a pure number) whereas φ = −45°. Correct: φ = arcsin(−1/√(1+x²)). The chain sin(φ) = −XC/Z = −1/√(1+x²) itself is right.

ME‑10. The parity proof establishes the wrong symmetry (p. 17, A9/3)

Bizonyítandó a párosság, azaz h. G(x)=G(−x) … = (−1)²·(x−1/x)² = (x−1/x)², tehát a párosság igaz.

The algebra is correct but proves evenness about x = 0, which cannot justify the mirror symmetry of the log-x plot about x = 1: negative x does not exist on a logarithmic axis (log(−x) is undefined). The symmetry actually present is G(x) = G(1/x) (reciprocal points): (1/x − x)² = (x − 1/x)² gives G(1/x) = G(x). This is also the correct basis for the geometric-mean statement on p. 18.

Related (p. 18): "a középfrekvencia a két határfrekvencia logaritmusának számtani közepe" — a frequency is not a mean of logarithms. Correctly: log(f_center) = ½(log f₁ + log f₂), i.e. f_center = √(f₁·f₂), the geometric mean (the clause "azonos a mértani közepével" is the correct one; here f_center = f0 exactly because x₁·x₂ = 1). The same page also labels its figure both "x-ben lineáris" (top) and "logaritmikusan skálázott" (later) — the mirror-symmetry / geometric-mean argument requires the logarithmic-x reading.

ME‑11. Wrong reason for choosing the + root (p. 19, A9/3)

x²+Ax−1=0 → x₁,₂ = −A/2 ± √((A/2)²+1). Ebből 1-nél nagyobb x-re a + előjel lesz érvényes.

This is the lower band edge x₁, which by the document's own setup satisfies 0 < x₁ < 1, so "for x greater than 1" is false. The + sign must be chosen because x₁ must be positive (the other root −A/2 − √(…) < 0 is a meaningless negative frequency ratio). The value x₁ = −A/2 + √((A/2)²+1) = 0,939 (R = 20 Ω) is itself correct; the parallel passage on p. 24 gives the correct reason: "…mivel így teljesül a 0 < x₁ < 1."

ME‑12. Wrong general quadratic formula (p. 24, B3)

[ x²+px+q=0! és x₁,₂ = p/2 ± √((p/2)²+q) ]

Both signs are wrong: x₁,₂ = −p/2 ± √((p/2)² − q). Checked against the document's own equation x²+Bx−1=0 (p = B, q = −1): the correct formula gives −B/2 ± √((B/2)²+1), matching the root printed beside it; the printed formula would give +B/2 ± √((B/2)²−1), which for B = 0,316 is complex.

ME‑13. Product of roots written as a sum (p. 24, B3)

x₁·x₂ = (a+b)(a−b) = a²−b² = (B/2)² + [√((B/2)²+1)]² = 1

With a = √((B/2)²+1), b = B/2, the identity is a² − b² = [√((B/2)²+1)]² − (B/2)² = 1. As printed it is a sum, equal to 2(B/2)² + 1 ≠ 1. (The result x₁·x₂ = 1 is correct; only the substituted expression is mis-ordered/mis-signed.)

ME‑14. Series form factor A used in the parallel band-edge equation (p. 24, B3)

x²+Ax−1=0 x₁,₂ = −B/2 ± √((B/2)²+1)

The quadratic should read x²+Bx−1=0: A = √(R²C/L) belongs to the series chapters A9/3–A9/4, whereas the parallel chapter uses B = √(L/(R²C)); the solution printed beside it already uses B.

ME‑15. Inductor P(t) loses its time dependence (p. 6, A4)

P(t)= UL*I(t)= L*(I1/t1)t(I1/t1) = L*I1²/t1

The middle expression is L·(I1²/t1²)·t — a ramp in t. The printed L·I1²/t1 is constant and equals it only at t = t1, i.e. it is P(t1), not P(t). It also contradicts the adjacent figure, which shows P growing linearly (the OAB triangle). The next line, E = ½·t1·P(t1) = ½·L·I1², is correct.

ME‑16. "Infinite inductive voltage" (p. 6, A4)

Az áramváltozás sebessége Δi/Δt > 0 korlátlan ideig … nem tartható, mert ez végtelen induktív feszültségre vezetne.

The preceding sentence states a fixed-voltage source keeps Δi/Δt constant; the inductive voltage then stays finite (equal to the source voltage). What grows without bound is the current i(t) = (Δi/Δt)·t. The claim should read "…would lead to unbounded current."

ME‑17. The 1,786 ms lead attributed to the inductive voltage (p. 9, A5)

Δt=−T*φ/(2π)=…=−1,786 ms NB. Ennyi idővel siet az induktív feszültség kezdőértéke.

φ = arctg(ωL/R) = 0,561 rad is the phase of the generator voltage relative to the current (and to uR), so 1,786 ms = φ/ω is the generator's lead. The inductive voltage leads the current by 90°, i.e. T/4 = 5 ms, not 1,786 ms.

ME‑18. φ = +45° at ω0 where the document's own tangent gives −45° (p. 21, B1)

A sarokfrekvenciára, ahol XL=XR, azaz ωL=R, és a φ=45°-hoz tartozó körfrekvencia: ω0=R/L

The same page derives tg(φ) = −YL/YR = −R/(ωL) = −1/x (with the explicit note "a negatív előjel"), so at ω0 (x = 1) φ = arctg(−1) = −45°; the p. 22 phase plot itself marks −45°. (The B2 case, tg(φ) = +ωRC → +45°, is the one where +45° is correct.)


3. Convention issues that affect correctness

CV‑1. Phase-sign convention flips at A8, and arccos(R/Z) is sign-blind (p. 12 vs pp. 8–11)

A5 and A7 write ug(t) = ug0·sin(ωt + φ) with a signed φ (A5: φ = +0,561; A7: φ = −1,1304 from sin φ = −XC/Z). A8 writes ug(t) = ug0·sin(ωt − φ) with φ = arccos(R/Z) = 0,9808 forced positive; its numeric line even prints the symbolic "+φ" while substituting −0,9808. Beyond the bookkeeping, arccos(R/Z) cannot distinguish an inductive from a capacitive circuit — it is sign-blind. A8's recipe sin(ωt − arccos(R/Z)) gives the right phase here only because the circuit is capacitive (XC > XL, φ < 0); for an inductive series RLC it would produce the wrong sign. Fix: use the signed form φ = arctg((XL−XC)/R) throughout.

CV‑2. Δt sign flips between A5 and A7 (p. 9 vs p. 11)

A5 prints Δt = T·φ/(2π) (leading minus; φ = +0,561 → −1,786 ms); A7 prints Δt = +T·φ/(2π) (no minus; φ = −1,1304 → −3,598 ms). With one fixed definition and signed φ, the two physically opposite cases (generator leads in A5, lags in A7) should give opposite-sign Δt; as printed both come out negative, so the sign of Δt carries no consistent meaning. Under Δt = −φ/ω, A7 gives +3,598 ms.


4. Figures inconsistent with their own parameters

FIG‑1. Phase curves far too steep for the labelled R (p. 25, B3)

The lower figure plots φ(x) for R = 500/1000/2000 Ω on an x-axis ticked 0,95 … 1,05. For R = 2000 Ω (green, Q = 12,65) the formula φ = arctg[(1/B)(x−1/x)] gives −52° at x = 0,95 and reaches −45° only at x ≈ 0,96; the plotted green curve is already at ≈ −85° at x = 0,95 and crosses −45° near x ≈ 0,99. The R = 500 Ω (blue) curve reaches ≈ ±75° at x = 0,95/1,05 where the formula gives ±18°. The plotted curves match an argument roughly 10× larger (equivalently an x-range ~10× wider than the tick labels). The amplitude figure directly above is consistent with its BW/Q values; only the phase figure is off.

FIG‑2. Reactance-crossing label (p. 13, A9)

The impedance-vs-frequency figure labels the XL = XC crossing "115,7 OHM." With L = 200 mH, C = 15 µF the crossing reactance is √(L/C) = 115,47 Ω → 115,5 Ω, consistent with the figure's own frequency label f0 ≈ 91,9 Hz (XL = 2π·91,89·0,2 = 115,47 Ω). The label digit is wrong.


5. Numerical / dimensional (genuine, not rounding)

PageAs printedCorrectNote
p. 26–27R = Q·√(L/C) = 12,6·√(5E−5/2E−9) = 2000 Ω1992 Ω12,6·158,11 = 1992,2 Ω. The input Q = 12,6 is inconsistent with the document's own table value Q = 12,65 (which gives 2000 Ω). State Q = 12,65 or report R ≈ 1992 Ω.
p. 27 (C1 table)BWf = 2π·f0·A and BWf = 2π·f0·BBWf = f0·A (resp. f0·B)Spurious 2π. The tabulated BWf values equal f0·A (R = 5 Ω: 503,3·0,0316 = 15,9 kHz); 2π·f0·A = ω0·A ≈ 100 krad/s is an angular bandwidth, not the value in the kHz column.

(The many further last-digit differences in the tables come from rounding chains — e.g. f0 taken as 503 kHz in one place and 503,3 kHz in another, so series Q reads 31,65/15,82/7,911 in the p. 27 table vs the correct 31,62/15,81/7,91 in the p. 18 figure. These do not change any conclusion and are not itemized here.)


6. Document-integrity notes (not mathematical)


Appendix — independently recomputed and confirmed correct (audit trail)

All values below were derived from scratch in Python and reproduce the document's numbers.

Time domain (A4–A8). A4: XL = 62,83 Ω, iL0 = 238,7 mA, ωt₁ = 0,9425 rad = 54,0°, uL(t₁) = 15·cos 54° = 8,817 V, iL(t₁) = 193,1 mA. A5: Z = 118,10 Ω, sin φ = 0,5320, φ = 0,561 rad = 32,14°, I0 = 127,0 mA; KVL at t₁: 14,966 V = 4,691 + 10,275 ✓. A6: XC = 212,2 Ω, iC0 = 70,7 mA, uC(t₁) = −8,817 V, iC(t₁) = 57,19 mA. A7: Z = 234,6 Ω, I0 = 63,94 mA, sin φ = −0,9045, φ = −1,1304 rad = −64,77°; KVL: −2,803 V = −7,976 + 5,173 ✓. A8: Z = 179,76 Ω, I0 = 83,45 mA, XL − XC = −149,4 Ω; the KVL check uR + uL + uC = ug(t₁) ≈ −0,576 V balances to full precision (the printed digit strings are internally rounded, but the identity holds).

Filters (A9, B). G = cos φ for all six transfer functions. Series band-pass: A = √(R²C/L) = 0,0316 / 0,0632 / 0,1265 with Q = 1/A = 31,6 / 15,8 / 7,91 (R = 5/10/20 Ω). Parallel: B = √(L/(R²C)) with Q = 1/B = 3,16 / 6,33 / 12,65 (R = 500/1000/2000 Ω). Band edges x₁ = 0,939, x₂ = 1,065 with x₁·x₂ = 1 and BW = x₂ − x₁ = A (resp. B); reciprocity A·B = 1. f0 = 1/(2π√(LC)) = 503,3 kHz. Design tasks: L = 1/(C(2πf0)²) ≈ 50 µH; series R = (1/Q)·√(L/C) = 10 Ω; parallel R = Q·√(L/C) ≈ 2000 Ω (see § 5, first row). RL low-pass L = R/(2πf0) = 1,6 mH; RC high-pass C = 1/(2πf0R) = 16 pF.

Appendix identities (D1–D7). D2 angle-addition construction; D3 product formulas (cos·cos, sin·sin); D4 sin²/cos² identities correct in this version — sin²(x) = ½[1 − cos 2x], cos²(x) = ½[1 + cos 2x]; D5 half-angle; D6 geometric difference-quotients correct — Δsin/Δx = cos, Δcos/Δx = −sin (the boxes are right; only the D7/4 re-derivation on p. 34 concludes the wrong sign, § ME‑1); D7/3 product / reciprocal / quotient / chain rules; D7/4 power rule and arcsin / ln / exp derivatives; and the e = lim(1 + 1/n)ⁿ table (all nine tabulated values recomputed correct; e = 2,718281828…).