
What Is an Accelerator RF PCB?
An accelerator RF PCB is a printed circuit board engineered to drive, control or monitor the radio-frequency cavities and amplifiers inside a particle accelerator. These boards must survive continuous high-power RF signals (often hundreds of watts to kilowatts at frequencies from roughly 30 MHz to beyond 3 GHz), cumulative ionizing radiation doses that can exceed 100 krad over the board’s service life, and thermal loads far beyond those of commercial electronics. The combination of power density, radiation exposure and frequency precision makes accelerator RF PCBs one of the most demanding niches in RF PCB engineering.
Operating Environment: Why Standard Boards Fail

Inside a synchrotron or linac tunnel, electronics face three simultaneous stresses that ordinary FR-4 boards cannot tolerate for long:
- Ionizing radiation — gamma rays, neutrons and secondary particles degrade epoxy resin, shifting dielectric constant (Dk) and increasing dissipation factor (Df) over time. Standard FR-4 can show measurable Dk drift after only 10–20 krad of total ionizing dose (TID).
- High CW power — low-level RF (LLRF) control loops and solid-state power amplifier (SSPA) driver stages can dissipate 5–50 W per board, while final-stage PA carriers may handle hundreds of watts. Copper weight, via thermal resistance and laminate Tg all become critical.
- Frequency stability — cavity tuning loops demand phase noise and amplitude ripple well below 0.1° and 0.1 % respectively, so the board’s Dk must stay stable across temperature and aging.
Material Selection for Radiation and Power
Radiation-Hard Laminates
Polyimide-based laminates (Tg > 260 °C) tolerate TID levels above 1 Mrad with minimal Dk shift, making them the default choice for boards mounted inside the accelerator tunnel. PTFE-ceramic composites such as ceramic-filled PTFE laminates also perform well under radiation because the ceramic filler is inherently inert, and the PTFE binder resists chain scission better than standard epoxy.
When boards sit in shielded service galleries with lower dose rates (below roughly 10 krad lifetime), high-Tg thermoset laminates—Rogers RO4000 series or Isola Astra MT77—can be acceptable, provided their Dk tolerance is verified against the expected dose budget.
Thermal Laminates for High-Power Stages
Solid-state PA driver boards and bias-supply boards benefit from laminates with elevated thermal conductivity. Rogers 92ML StaCool laminate offers 2.0 W/m·K thermal conductivity, roughly four times that of standard FR-4, letting heat spread into the ground plane and out through a heatsink without excessive junction-to-board thermal resistance.
Material Comparison for Accelerator RF PCBs
| Property | Polyimide | Ceramic-Filled PTFE | Rogers 92ML | High-Tg FR-4 |
|---|---|---|---|---|
| Dk (10 GHz typical) | 3.5–4.0 | 3.0–10.2 | 3.6 | 4.2–4.5 |
| Df (10 GHz typical) | 0.008–0.012 | 0.001–0.004 | 0.004 | 0.018–0.025 |
| Thermal conductivity (W/m·K) | 0.35 | 0.5–0.7 | 2.0 | 0.3 |
| TID tolerance (approx.) | > 1 Mrad | > 500 krad | ~50 krad | 10–20 krad |
| Tg (°C) | > 260 | > 280 | > 280 | 170–180 |
Power-Plane Integrity and IR Drop
Accelerator LLRF boards often carry mixed high-current bias supplies (28 V / 10 A for GaN PAs, ±15 V for analog front-ends) alongside sensitive phase-detector rails that cannot tolerate more than a few millivolts of DC drop. Heavy copper planes—2 oz (70 µm) to 4 oz (140 µm)—reduce resistive loss, but the designer still needs to verify that the voltage at the load pin stays within spec under worst-case current draw.
Use the calculator below to check the IR drop across a power plane for your specific copper weight, plane dimensions and load current. Enter the plane length and width in millimeters, the copper thickness in ounces, and the DC current in amps.
For example, a 2 oz copper plane that is 80 mm long and 25 mm wide carrying 8 A shows an IR drop of roughly 18 mV—acceptable for a 28 V PA supply but potentially significant for a 3.3 V FPGA core rail. If the result exceeds your budget, widening the plane, increasing copper weight or adding parallel via stitching between split planes are the standard remedies. Boards destined for heavy copper and thermal RF construction routinely use 3–4 oz inner layers specifically to keep IR drop under control in high-current accelerator service.
Stackup and Transmission-Line Design
Most accelerator LLRF boards are 6- to 12-layer stackups combining controlled-impedance signal layers with dedicated power and ground planes. Stripline is preferred for RF distribution (cavity drive signals, local-oscillator feeds) because the enclosed structure reduces radiation and coupling—critical when traces run beside high-current bias planes that radiate switching noise.
Microstrip is used mainly for the top-layer interface to SMA or N-type RF connectors. Coplanar waveguide with ground (CPWG) transitions help maintain 50 Ω impedance right up to the pad, minimizing VSWR at the board edge. A detailed treatment of multi-layer RF stackup considerations is available in our guide to multilayer RF PCB fabrication.
Radiation-Hardening Techniques at Board Level

Beyond choosing a radiation-tolerant laminate, several board-level practices extend service life in the tunnel environment:
- Guard rings around sensitive analog nodes — a grounded copper pour ring limits leakage current induced by surface ionization of the solder mask.
- Redundant via stitching — radiation-induced delamination at the drill wall is mitigated by tighter via-to-via spacing (≤ 2 mm pitch) around critical ground transitions.
- Conformal coating — acrylic or silicone coatings rated per IPC-CC-830C protect exposed copper and solder joints from corrosion accelerated by ozone generated in the beam environment.
- Component derating — capacitors and resistors are derated to ≤ 50 % of rated voltage and power, because radiation degrades dielectric strength over time.
Thermal Management Strategies
Even a well-chosen laminate needs a complete thermal path. Common approaches for accelerator RF PCBs include:
- Thermal vias under power devices — arrays of 0.3 mm filled and capped vias on 1.0 mm pitch beneath GaN or LDMOS pads, connecting the top pad to a bottom-side copper heatsink pad.
- Embedded copper coins — press-fit copper slugs (typically C110 alloy, ~390 W/m·K) placed directly beneath the PA die for a near-zero thermal interface.
- Forced-air or liquid-cooled cold plates — the PCB bolts to a machined aluminum cold plate with thermal interface material; board flatness (≤ 0.1 mm/25 mm) is specified to ensure full contact.
Designers working on PA carrier boards will find additional thermal layout detail in our RF power amplifier PCB guide.
Testing and Qualification
Accelerator facilities such as CERN and Fermilab typically require board-level qualification that goes beyond standard IPC Class 3:
- TID testing — sample boards are exposed to a calibrated Co-60 gamma source and measured for Dk shift, leakage current and delamination at dose intervals up to the design lifetime dose.
- Thermal cycling — –55 °C to +125 °C, 500 cycles minimum per JEDEC JESD22-A104, with cross-section and microsection inspection afterward.
- Controlled-impedance verification — every production panel is TDR-tested; tolerance is typically ±5 % on 50 Ω lines, tightened to ±3 % for phase-critical cavity-drive paths.
FAQ
Can FR-4 be used for accelerator RF PCBs?
Only in shielded service galleries with very low radiation dose rates (under ~5 krad lifetime) and at frequencies below roughly 1 GHz where its higher Df is tolerable. Inside the tunnel, polyimide or ceramic-filled PTFE is strongly preferred.
What frequency range do accelerator RF systems cover?
Most proton synchrotrons operate cavities between 30 MHz and 800 MHz. Electron storage rings and free-electron lasers use superconducting cavities at 1.3 GHz (TESLA/ILC standard) or 3 GHz (S-band). LLRF control boards must handle these fundamentals plus harmonics and diagnostics up to several GHz.
How thick should the copper be on an accelerator RF power board?
2 oz (70 µm) is a practical minimum for bias-supply planes. PA carrier boards often use 3–4 oz (105–140 µm) outer layers and 2 oz inner layers to manage both IR drop and thermal spreading.
What is the typical impedance tolerance for accelerator LLRF boards?
±5 % on general 50 Ω lines, tightened to ±3 % or better on cavity-drive and phase-reference distribution traces where phase stability directly affects beam quality.
Ready for a quotation?
Send your requirements for a material, stackup and manufacturing review. Price and schedule are confirmed for your project.
