Surface RF Receive Coil Designer

Full circuit design for a single-turn surface receive coil for MRI or NMR. Enter the loop geometry, Larmor frequency, and sample properties to get every component value needed to build the coil: distributed tuning capacitors, the impedance-matching network, the preamplifier decoupling inductor, and the required λ/4 cable length — plus the minimum number of capacitor breaks to suppress standing-wave artefacts.

D	Loop diameter (mm). Outer dimension of the single-turn conducting loop.
d	Conductor outer diameter (mm). Use outer diameter for both round wire and copper tubing.
f₀	Larmor frequency (MHz). The resonant frequency of the target nucleus in your static field.
h	Sample distance (mm). Distance from the coil plane to the nearest surface of the sample.
Sample	Tissue type or phantom. Sets conductivity σ and permittivity ε_r via interpolated dielectric data (Gabriel et al. 1996 / IT'IS database).

Physical constants & reference values ▸

µ₀	4π×10⁻⁷ H/m
¹H γ/2π	42.577 MHz/T
¹H at 1.5 T	63.87 MHz
¹H at 3 T	127.74 MHz
¹H at 7 T	297.7 MHz
IEC SAR limit (WB)	2 W/kg

Inputs

D — Loop diameter

5 – 600 mm (outer loop dimension)

d — Conductor diameter

mm OD

Outer diameter (skin effect uses outer surface)

f₀ — Larmor frequency

MHz

h — Sample distance

Distance from coil plane to nearest sample surface

Sample type

Circuit Schematic

Feed gap at bottom · ⊣⊢ symbols are capacitor breaks · all values labelled in the component table below

Coil Parameters

Electromagnetic

Inductance L—

Skin depth δ (Cu)—

Wire resistance R_wire—

Sample loading R_s—

Total loss R_total—

Quality Factor

Unloaded Q_u—

Loaded Q_L—

Q ratio Q_u/Q_L—

Sample limited?—

Parallel R_p—

ℹ Q ratio: A ratio Q_u/Q_L > 5 indicates the coil is sample-noise dominated — the dominant noise source is thermal noise in the tissue, not the wire. This is the desirable regime for high-SNR receive coils. If Q_u/Q_L < 2, the coil conductor is contributing significantly to noise and conductor quality should be improved (larger wire, better material, silver plating).

Capacitor Breaks

Recommended

—

⚠ Standing-wave rule: Each wire segment between capacitors should subtend less than λ_eff/10 to prevent standing-wave resonances that distort the current distribution and degrade field homogeneity. The free-space λ/10 criterion is too lenient in practice because the sample's high dielectric constant shortens the effective wavelength seen by the coil. An eighth-space model is used: ε_r,eff = (1 + ε_r,tissue)/8 (capped at 9), giving λ_eff = λ₀/√ε_r,eff. The formula is N_breaks = max(4, ⌈πD / (λ_eff/10)⌉). A minimum of 4 is enforced: 1 for the output/matching port and at least 3 for distributed tuning. If a PIN diode active-detuning trap is also required, add one more break.

Component Values

Component	Value	Qty	Function

Preamplifier Decoupling

Decoupling Parameters

Series inductor L_m—

Shunt capacitor C_m—

λ/4 cable length—

Required Z_preamp≤ 5 Ω

Estimated decoupling—

Matching Network Q

Match Q_m—

Reactive impedance ωL_m—

Z_high via λ/4 (2 Ω preamp)1250 Ω

Bandwidth (−3 dB)—

Cable type (suggested)—

ℹ Preamplifier decoupling principle: The series inductor L_m and shunt capacitor C_m form an L-network that simultaneously achieves two goals: (1) it transforms the coil's high parallel resistance R_p down to 50 Ω for efficient signal transfer to the coaxial line, and (2) it creates the correct reactance environment for preamplifier decoupling. The low-input-impedance preamplifier (Z_in ≈ 1–4 Ω), connected via a λ/4 cable, presents a transformed high impedance Z_high = Z₀²/Z_in ≈ 600–2500 Ω in series with the coil loop. This suppresses currents induced by neighbouring coils or the transmit pulse without attenuating the receive signal, since the series LC resonates at ω₀ and is transparent. This is the Roemer–Reykowski preamplifier decoupling technique used in all modern phased-array MRI receivers.

Design Equations

Loop inductance (Neumann formula, single circular turn, round conductor, r = loop radius, a = wire radius):

\( L = \mu_0 \, r \!\left[\ln\!\left(\dfrac{8r}{a}\right) - 2\right] \)

Wire AC resistance (skin effect in copper, ρ_Cu = 1.72 × 10⁻⁸ Ω·m):

\( \delta = \sqrt{\dfrac{2\rho_{\rm Cu}}{\omega\mu_0}}, \qquad R_{\rm wire} = \dfrac{\rho_{\rm Cu}}{\delta}\cdot\dfrac{r}{a} \)

Sample loading resistance (quasi-static, conducting half-space, h = coil-to-surface distance):

\( R_s \approx \dfrac{\pi}{6}\,\mu_0^2\,\omega^2\,\sigma\,\dfrac{r^6}{(r+h)^3} \)

Valid in the quasi-static limit (wavelength in tissue ≫ loop diameter). At 7 T+ with large coils, waveguide effects and dielectric resonances become significant and this formula underestimates the true loading. Always trim on the bench.

L-network matching (shunt C_m at coil, series L_m at output; R_p = ω²L²/R_total):

\( Q_m = \sqrt{\dfrac{R_p}{Z_0}-1},\qquad C_m = \dfrac{Q_m}{\omega R_p},\qquad L_m = \dfrac{Q_m\,Z_0}{\omega} \)

Distributed tuning capacitors (N_d = N_breaks − 1 identical caps in series giving C_total = 1/ω²L):

\( C_{\rm dist} = N_d \cdot C_{\rm total} = \dfrac{N_d}{\omega^2 L} \)

Capacitor break count (λ_eff/10 standing-wave criterion with tissue dielectric loading, N_d = N_breaks − 1 distributed caps):

\( \varepsilon_{r,\text{eff}} = \min\!\left(9,\;\dfrac{1+\varepsilon_{r,\text{tissue}}}{8}\right),\qquad \lambda_\text{eff} = \dfrac{\lambda_0}{\sqrt{\varepsilon_{r,\text{eff}}}},\qquad N_\text{breaks} = \max\!\left(4,\;\left\lceil\dfrac{\pi D}{\lambda_\text{eff}/10}\right\rceil\right) \)

The eighth-space model (ε_r,eff = (1+ε_r)/8) approximates the partial dielectric loading experienced by a coil lying on the surface of a tissue half-space. The cap of 9 prevents over-counting with high-permittivity tissues (e.g. blood, gray matter) at ultra-high field where the full tissue ε_r is not fully seen by the loop fields. The free-space λ/10 rule alone is far too lenient at MRI frequencies and would always produce the minimum of 4 for clinically-sized coils.

λ/4 cable length (v_f = cable velocity factor, typically 0.66–0.84):

\( \ell = \dfrac{v_f\,c}{4f_0} \)