First RF board spin: the five measurements that catch mistakes before the second fab

A lot of early-career RF spins fail the same way: the schematic is "matched," the layout looks clean, and the first bring-up burns a week because nobody decided **which five measurements are the gate** before ordering more boards.

This is the gate list I wish I had on spin #1 (sub-6 GHz radio / front-end style boards). Not a career essay — a bring-up contract.

**1. DC / bias map before any RF**
Rail voltages, current per rail, sequencer order, and "enable then RF" vs "RF then enable." Half of "dead LNA" tickets are bias or sequencing.

**2. Unpowered continuity / shorts on RF nets**
DC ohmmeter across RF paths that should be blocked by series C; shorts to ground on bias tees; wrong footprint bridges. Cheap, saves a fried first sample.

**3. S11 / S22 at the connector with a known cal**
Not "looks matched in sim" — measured at the launch you will ship. Log cal kit, torque, and cable ID next to the screenshot (Chinese Corner thread on daily lab checklist applies here too).

**4. One absolute level path**
CW in → CW out (or noise figure / gain) at a single frequency with pads and SA/VNA settings written down. Relative "it makes a tone" is not a gate.

**5. Spur / supply sanity under modulation or keyed TX**
Even a crude ACLR / harmonic check at one channel catches PA instability, missing decoupling, and LO leakage before you optimize EVM.

**What I deliberately leave out of spin-1 gates**
Full OTA pattern, full temp suite, and every corner case — those are spin-2. Spin-1 is "does the chain behave like the block diagram on the bench."

**Discussion**
On your first (or last) RF spin, which single measurement would you add to this five — and which one do juniors most often skip?

— rfforum 编辑部 / Editorial Team

Editorial addendum / open checklist for first-spin RF boards: the five measurements only help if they are taken with the production stack (connectors, saver policy, bias network) and archived with settings. A clean S11 on a blue-wire fixture that will never ship is not a spin gate.

Readers: which of the five do you treat as a hard stop before tape-out of spin 2 — and have you ever passed all five on the bench then failed EMC or thermal in the enclosure?

— rfforum 编辑部 / Editorial Team