Broadband Antenna Matching for a Multi-Band RF Transceiver

Four ISM bands, two antennas, one transceiver.

KickR Design asked MRS to deliver production-ready matching networks for a sub-GHz multi-band RF transceiver built around the TI CC1101RGPR. The board supports four ISM bands routed through an SP4T switch to two chip antennas: a narrowband Yageo ANT1204F002R0315A for 315 MHz, and a multi-band Abracon ACR1504I3 for 433, 868, and 915 MHz.

Each band needed a matching network such that |S₁₁| ≤ −10 dB across the required operational range, measured in-case with the display installed — the closest metallic object to the antennas and, therefore, the largest source of detuning. Deliverables were the four matching designs with component values keyed to the production schematic.

The signal chain — and why the antennas couple.

The CC1101's differential RF port is driven through an Anaren B0310J50100AHF balun into a SKY13414-485LF SP4T switch, which selects one of four per-band matching networks. Only one band is active at a time; the switch common port references 50 Ω post-balun. The photo shows the board in its production case with the display lifted on its hinge and two coax probes taped to the matching-network test points.

Both antennas sit in the same corner of the PCB, directly adjacent to the back of the display — which measurably couples them and detunes them: tuning one band shifts the response of another, and any match measured with the display removed is a lie. That coupling drove the retune order described in the next section.

Three calibration strategies. The third one worked.

Three approaches were attempted before the efficient one took hold:

1. Analytic PCB model. A simple delay + attenuation model of the trace from the balun to the matching network. Fast, but could not converge without capturing the balun and switch — the first-pass match for RF4 was the only one it produced cleanly.
2. Full Short / Open / Load calibration. Characterizing the PCB at both sides of each matching network. Highly accurate, but each band required six calibration solderings and three more components per tweak — sustainable for one band, not four, with the board's wear budget.
3. Three-measurement calibration (adopted). A variation of the first approach: measure the antenna alone, then in series with a known inductor, then with an additional shunt capacitor. Those three measurements extract delay and attenuation for the simple model, after which 2–3 component iterations center each band. Total: about six solderings per band, a 2× reduction over the full SOL approach.

The retune order was driven by coupling: RF3 (315 MHz) was tuned first because shifts there propagated to the other three bands through the shared antenna geometry. RF4, RF2, and RF1 were then remeasured, tweaked, and finalized in sequence.

Final results — all four bands.

Measurements taken in-case with the display installed. Each chart references the chip's target differential impedance (not 50 Ω), so the −10 dB threshold corresponds to the conjugate match CC1101 datasheets specify.

RF3

315 MHz Band

314–316 MHz · FCC Part 15 ✓ Pass

Final |S₁₁| measurement — gold band indicates the required frequency range.

Antenna

Yageo ANT1204F002R0315A

Required BW

2 MHz

Target impedance

122 + j31 Ω

Matching topology

51 nH series · 7 pF shunt

Board refdes

X1 = 51 nH, X4 = 7 pF

Result (in-band)

−14 to −15 dB

Note. Band was retuned first because 315 MHz affects the others through antenna coupling. In-case, in-spec.

RF4

433 MHz Band

432–434 MHz · Region 1 ISM ✓ Pass

Final |S₁₁| measurement — gold band indicates the required frequency range.

Antenna

Abracon ACR1504I3

Required BW

1.74 MHz

Target impedance

116 + j41 Ω

Matching topology

20 nH series · 0.5 pF shunt

Board refdes

X19 = 20 nH, X23 = 0.5 pF

Result (in-band)

−13.3 dB (min)

Note. Relative to 866 MHz (≈2×), the same topology scales cleanly. No final tweak required after the 315 MHz retune.

RF2

868 MHz Band

863–870 MHz · ETSI EN 300 220 ✓ Pass

Final |S₁₁| measurement — gold band indicates the required frequency range.

Antenna

Abracon ACR1504I3

Required BW

7 MHz

Target impedance

86 + j43 Ω

Matching topology

4 pF shunt · 15 nH series · 1.0 pF shunt

Board refdes

X10 = 4 pF, X7 = 15 nH, X11 = 1.0 pF

Result (in-band)

−15 dB or better

Note. Hardest match of the four. L-section wasn't enough; added a shunt cap to move the resonance into band.

RF1

915 MHz Band

900–928 MHz · FCC Part 15.247 ✓ Pass

Final |S₁₁| measurement — gold band indicates the required frequency range.

Antenna

Abracon ACR1504I3

Required BW

26 MHz

Target impedance

86 + j43 Ω

Matching topology

2 pF shunt · 15 nH series · 0.2 pF shunt

Board refdes

X16 = 2 pF, X13 = 15 nH, X17 = 0.2 pF

Result (in-band)

−17 dB at 911 MHz

Note. Widest fractional bandwidth. Centered the resonance on the band after adjusting the input shunt down to 2 pF.

Recommendations for production.

• Verify on a second board. A separate fabrication run confirms the solution transfers — PCB-to-PCB variation in the balun, switch, and trace geometry is small but non-zero.
• Avoid 5 %-tolerance components. The tighter matches (RF2, RF1) rely on component values within 1–2 % of the tuned set; 5 %-tolerance parts risk drifting the resonance out of band.