60 GHz In-Cabin Radar PCB: Carrier Boards for AiP Occupant Sensing
60 GHz in-cabin radar PCB guide: AiP carrier boards for occupant and child-presence sensing, 60 GHz dielectric loss, via stub control, warpage and VNA tests.
Table of Contents
In-cabin sensing is becoming a standard automotive requirement — child-presence detection, seat occupancy, vital signs and intrusion detection — and the sensor class that carries it has collapsed into a single chip: the newly published 60 GHz automotive AiP radar sensor integrates the RF front end, PLL, ADC, state machine and the antennas into an 8 × 8 × 0.86 mm package qualified to AEC-Q100 Grade 2, sweeping 57.7–62.4 GHz with a 4.7 GHz FMCW bandwidth.
The tempting conclusion — “the antennas are in the package, so the PCB got simpler” — is exactly backwards. What the integration does is move the RF performance boundary onto the carrier board: the package-to-board launch, the reference continuity under the antenna region, the via stub budget, the laminate loss across the sweep and the flatness of the mounting surface are now the manufacturing conversation. This page computes that conversation: the 60 GHz board arithmetic, the range-resolution story, the laminate loss estimate, the stub budget, and the flatness budget against a package that is less than a millimetre tall. Every figure is computed; the assumptions are stated where they are used.
The forward-radar side of the car — 77 GHz long-range modules and ADAS boards — is treated in our automotive radar PCB overview, and radar fabrication across bands in our radar PCB manufacturer review. This page is about the cabin: the carrier boards behind an AiP occupant-sensing radar.
1. What an AiP Radar Module Asks of the PCB
The module is one chip, but the shipping unit is a board group with three distinct disciplines. The table maps it.
| Board in the in-cabin radar module | What it carries | What the fabricator must confirm |
|---|---|---|
| AiP sensor carrier board | 8 × 8 × 0.86 mm sensor footprint (reported AiP class), antenna keep-out above the array, laser-via fan-out | Blind/laser stackup, keep-out discipline, low-profile copper |
| Baseband & interface board | MCU, PMIC, CAN/LIN or private vehicle bus, power sequencing | Isolation between digital returns and the RF carrier |
| Module carrier / bracket interface | Connector to the vehicle harness, shielding can, mounting to the plastic trim | Mounting flatness and shield coplanarity carry the RF budget |
| Coupon set & launch structures | 60 GHz launch de-embedding, loss-trend and phase coupons for the carrier | Coupon set agreed before production, not after first article |
Integrating the RF front end, PLL, ADC, state machine and antennas into one AiP package does not remove the board — it moves the RF performance boundary onto it. The package-to-board transition, the reference continuity under the antenna region, the coplanarity of the mounting surface and the coupon that proves all of this lot by lot are now the manufacturing conversation. Quoting a "digital-class" carrier for a 60 GHz sensor is the most common way this module fails its bench test.
The row that changes quotation behaviour is the last one: because the antenna lives in the package, the only place performance can be proven is the carrier and its coupons. A fabricator who treats the carrier as a two-layer digital board has already lost the module’s bench test; a fabricator who prices the coupon set into the first article is the one whose units pass the vehicle-level detection requirements on lot 20 as well as lot 1.
2. The 60 GHz Arithmetic: 134 Degrees per Millimetre
Everything unusual about this carrier follows from one table.
| Geometry at 60 GHz (εr 3.48 class carrier) | Value | Design consequence |
|---|---|---|
| Free-space wavelength λ0 | 5.00 mm | The whole sensor fits inside one free-space wavelength |
| In-medium wavelength λg | 2.68 mm | Every trace is a distributed structure |
| Phase per millimetre | 134°/mm | A 0.5 mm layout slip is 67° of unplanned phase |
| Quarter wavelength λg/4 | 0.67 mm | The stub budget for every via on the RF chain |
λg = λ0/√εr. On a PTFE-ceramic class carrier (εr ≈ 2.95) λg is 2.91 mm and the phase density is 124°/mm — the arithmetic changes with the laminate, which is why the stackup is reviewed before layout, not after. At 77 GHz the front radar world already lives with sub-3-mm wavelengths — the method is set out in our 77 GHz radar PCB design guide — and the in-cabin band applies the same discipline two octaves of layout care earlier than most digital teams expect. Impedance and phase verification flow through our controlled impedance RF PCB and 50 ohm trace guides.
Two rows discipline the layout. The phase density means every trace on the RF chain is a distributed structure — the fan-out, the ground stitching and the launch geometry are dimensioned to tens of microns, and a 0.5 mm detour is 67 degrees of phase the sensor did not ask for. The quarter-wave row is the manufacturing budget: with λg/4 at 0.67 mm, the via strategy has one honest answer, which is section 5. The layout-level method for mounting RFIC-class packages is the same discipline as our QFN/MMIC RF PCB layout guide, tightened to the 60 GHz scale.
3. Range Resolution: What the 4.7 GHz Sweep Buys
The sensor’s application stack — child-presence detection first among it — rests on range resolution, and range resolution is a board-level conversation about sweep integrity.
| FMCW range resolution (reported 4.7 GHz sweep bandwidth) | Value | What it means for the board |
|---|---|---|
| Ideal resolution ΔR = c/(2B) | 3.2 cm | Occupant vs seat vs child-scale separation in the cabin |
| Effective resolution (70% sweep utilization) | 4.6 cm | Windowing and sweep flatness cost resolution, not the board |
| Vital-signs sensing | mm-level chest displacement, 1–2 Hz | Handled in baseband processing; the board’s job is phase stability |
| Phase stability demand | Carrier and RX chain phase noise floor set by laminate and copper | Loss-flatness across 57.7–62.4 GHz, not just insertion loss at one point |
ΔR = c/(2B) with the reported 4.7 GHz FMCW bandwidth; the 70% row is a stated calculation assumption for windowing loss, not a device figure. The cabin application explains the spec stack: centimetre-class range resolution separates an out-of-position occupant from the seat, and millimetre-level vital-signs detection rides on phase stability across the sweep. That is a laminate loss-tangent and copper-roughness conversation — which is the next table.
Centimetre-class resolution is what separates an out-of-position occupant from a seat, and millimetre-level vital-signs detection rides on phase stability across the sweep rather than raw sensitivity. The PCB contribution is unglamorous and decisive: loss that is flat across 57.7–62.4 GHz, a phase-stable reference structure, and coupons that measure exactly that. A spot-check insertion-loss number at one frequency proves none of it.
4. Laminate Choice: the Loss Budget at 60 GHz
Laminate selection at 60 GHz is a numbers conversation, and the estimate is blunt.
| Laminate class at 60 GHz (dielectric-loss estimate) | εr / tanδ | αd estimate @ 30 mm trace |
|---|---|---|
| Hydrocarbon RO4350B-class | 3.48 / 0.0037 | 1.13 dB — workable only on the shortest runs |
| PTFE-ceramic RO3003-class | 2.95 / 0.0010 | 0.28 dB — loss budget survives the sweep |
| PTFE ceramic-filled mid-class | 3.0 / 0.0013 | 0.37 dB |
| Conductor side: skin depth at 60 GHz | 0.27 µm | Standard copper (Rz ~8 µm) puts roughness inside the conduction band — low-profile copper is a spec line |
αd = 27.3·√εr·tanδ/λ0 — a dielectric-loss estimate for ordering laminate classes, not a substitute for the measured sweep; tanδ values are datasheet-class figures at 10 GHz and rise with frequency, which is why the acceptance test is a swept measurement, not a spot check. The skin-depth row is the other half of the budget: at 60 GHz the current crowds into 0.27 µm, so copper roughness stops being a cosmetic grade and becomes an insertion-loss line item. Material families are profiled in our Rogers RO3003 PCB review and compared across classes in the PCB material comparison chart; the materials are selected by loss budget and stackup fit — each family carries its own application envelope, and the RFQ decides.
The fourfold difference between the laminate classes over a 30 mm run is the whole story of section 2’s phase discipline applied to the bill of materials: on a carrier whose RF runs are short by design, the class with four times the loss can still work — if the sweep, the budget and the coupon data say so. The point is the method, not the brand: swept loss across the band, copper roughness as a specified grade, and the decision taken at stackup review with the link budget in hand. The measurement side — swept verification on production launch geometry — is set out in our mmWave VNA testing guide.
5. Via Stubs: the 0.67 mm Quarter-Wave Budget
If section 2 gave the budget, this table spends it.
| Via residual stub (εr 3.48 carrier) | Quarter-wave resonance | Verdict at 57.7–62.4 GHz |
|---|---|---|
| Laser stub 0.3 mm | 133.9 GHz | More than 2× above band — clean |
| Laser/controlled-depth stub 0.5 mm | 80.4 GHz | Above band, acceptable |
| Stub 0.8 mm | 50.2 GHz | Sitting on the band skirt — marginal by construction |
| Through via 1.5 mm (un-backdrilled) | 26.8 GHz fundamental — reactive across the whole band | Out by design: the RF chain is blind/laser or backdrilled with residual stub as a drawing dimension |
f = c/(4·L·√εr). The quarter-wave number from table 2 is the rule here: with λg/4 at 0.67 mm, any via longer than ~0.7 mm is a band-edge resonator, and a 1.6 mm carrier built with through-hole fan-out puts a 1.5 mm stub in every pad. In-cabin carriers are therefore thin, laser-via boards by design — not as a premium, but because the arithmetic leaves no alternative. Drilling capability and depth control are set out in our drilling capabilities page, via-fence and ground discipline in the via fence PCB guide.
The consequence row is not a preference: on a carrier that still uses through-hole fan-out, every pad on the RF chain carries a stub that is two quarter-waves long at band edge — a reactive mess across 57.7–62.4 GHz. In-cabin radar carriers are therefore thin, laser-via boards by construction, with the residual stub written on the drawing as a dimension and verified by microsection per panel. This is the same discipline the 77 GHz front-radar world already applies — our drilling capabilities page covers depth control, and the via fence PCB guide covers the ground-side stitching that keeps the reference intact around the keep-out.
6. Flatness: Warpage Budget Against a 0.86 mm Package
The last physical budget is mechanical, and the reference stick is less than a millimetre tall.
| Flatness class (bow & twist) | Deflection on a 100 × 100 mm panel | Versus the 0.86 mm package height |
|---|---|---|
| IPC-conventional 0.75% | 0.75 mm | Consumes the entire package-height budget |
| Tightened SMT class 0.5% | 0.50 mm | Leaves 0.36 mm — thin margin after reflow |
| Carrier-class 0.3% | 0.30 mm | Workable with the shield can and bracket in the budget |
| 0.3% on a 200 × 200 mm panel | 0.60 mm | Panelization size is part of the flatness spec |
Bow-and-twist percentages scale with panel size, and the reflow step on the carrier adds warp of its own — the assembly house pays for whatever the fab drawing did not specify. The 0.86 mm package height (reported AiP class) is the reference stick: a conventional 0.75% flatness class on a 100 mm panel already deflects 0.75 mm before the shield can, connector and bracket tolerances take their share. Flatness, warpage class and panelization are therefore spec lines on the fabrication drawing, and reflow survivability of the laminate belongs in the stackup review. Package-to-board launch mechanics are treated in our mmWave package-to-PCB transition guide and the QFN/MMIC RF PCB layout guide.
Flatness is where digital-class carriers quietly fail automotive radar modules: the fab ships boards inside a conventional 0.75% class, the reflow step adds its own warp, and the assembly house discovers that the shield can no longer seats coplanar over the sensor. On this product the flatness class, the panelization size and the reflow survivability of the laminate are one tolerance chain, agreed at stackup review and written on the drawing. Package-side mounting mechanics and launch tolerances are treated in our mmWave package-to-PCB transition guide; module-level assembly and qualification flow in the antenna module PCB assembly guide.
7. The DFM Checklist Before the In-Cabin Radar RFQ
New on this subject
- 60 GHz dual-beam communication module PCB — element pitch, beam-to-beam isolation and the 57–70 GHz loss budget
Nothing here is exotic; each line traces to a number computed above, and each is cheap on a drawing and expensive to reconstruct after a first article fails in the cabin.
| # | Drawing or PO line item | Why it is on the list |
|---|---|---|
| 1 | Antenna keep-out: no copper, no solder mask irregularities above the AiP array region | 134°/mm — anything in the near field is a radiator (section 2) |
| 2 | Laminate selected by swept loss budget across 57.7–62.4 GHz, not one-point insertion loss | 0.28 vs 1.13 dB per 30 mm trace (section 4) |
| 3 | Low-profile copper specified for the RF layers | 0.27 µm skin depth (section 4) |
| 4 | RF chain fan-out blind/laser: residual stub ≤ 0.5 mm as a drawing dimension, microsection-verified | 0.67 mm quarter-wave budget (section 5) |
| 5 | Flatness class ≤ 0.3% with panelization size written on the drawing | 0.75 mm conventional deflection vs 0.86 mm package (section 6) |
| 6 | AEC-Q100 Grade 2 sensor environment carried into board spec: wide-temp lamination and surface finish | The module is automotive-qualified; the carrier must not be the weak link |
| 7 | Launch and coupon structures agreed before production: de-embedding, loss-trend, phase | Swept verification, not spot checks (sections 3–4) |
| 8 | Shield-can land pattern and bracket interface flatness in the same tolerance chain | Coplanarity carries the RF budget (sections 1, 6) |
Eight lines, each tracing to a number computed above. Items 1–4 decide whether the 60 GHz signal survives the board; 5–6 decide whether it survives the car; 7–8 decide whether any of it can be proven lot by lot. An in-cabin radar carrier RFQ that carries these eight answers is quotable on engineering.
The pattern across the whole page: AiP integration moved the RF boundary onto the carrier, and the carrier is now quoted on swept loss, stub arithmetic and flatness — not on layer count. For the adjacent decisions — the forward 77 GHz radar boards sharing the vehicle, the module assembly flow that qualifies the finished sensor, or the material families behind the loss budget — the 77 GHz radar PCB review, the 77 GHz design guide, the antenna module PCB assembly guide and the PCB material comparison chart are the right starting points, and our RF PCB manufacturer overview covers the fabrication capabilities behind all of them.
60 GHz In-Cabin Radar PCB — Frequently Asked Questions
What in-cabin radar is, whether AiP removes the board, laminate choice at 60 GHz, why flatness is a spec line, and what the RFQ should contain.
What is in-cabin radar, and why is it moving to 60 GHz?
In-cabin radar is a small FMCW sensor inside the passenger compartment that detects occupants, child presence, seat occupancy, vital signs, gestures and intrusion. The 57.7–62.4 GHz band (60 GHz class) is used because the short wavelength allows a compact antenna-in-package sensor, the sweep bandwidth gives centimetre-class range resolution, and the band is harmonized for such sensing. The newly published 60 GHz automotive AiP sensor class covers this band with a 2-transmit/4-receive array and AEC-Q100 Grade 2 qualification.
The antennas are in the package — does the PCB still matter?
Yes, and the arithmetic explains why. At 60 GHz the wavelength in a typical carrier laminate is about 2.7 mm and the phase density is 134° per millimetre. The package-to-board transition, the reference and ground continuity under and around the array keep-out, the via fan-out stub budget and the coplanarity of the mounting surface are all inside the RF budget now. An AiP removes antenna design from the customer, not the carrier board’s RF discipline from the fabricator.
Which laminate should a 60 GHz in-cabin carrier use?
The one whose swept loss budget fits the link, decided at stackup review with the sweep in hand. A dielectric-loss estimate at 60 GHz puts a hydrocarbon RO4350B-class laminate near 1.1 dB over a 30 mm run and PTFE-ceramic RO3003-class near 0.3 dB, and tanδ rises with frequency, so the acceptance test is a swept measurement across 57.7–62.4 GHz. Each material family carries its own application envelope; the RFQ, the sweep and the stackup decide — not brand names.
Why is board flatness a specification line for a radar carrier?
Because the sensor package is only 0.86 mm tall and reflow warpage is compared against that stick. A conventional 0.75% bow-and-twist class deflects a 100 mm panel by 0.75 mm — the entire height budget — before the shield can, connector and bracket tolerances take their share. In-cabin carriers therefore specify a tightened flatness class with the panelization size written on the drawing, so the assembly house inherits a defined surface, not a surprise.
What should an in-cabin radar carrier RFQ include?
The conditions that change the board: the sensor package footprint and its transition style, the antenna keep-out definition, the operating band and sweep (57.7–62.4 GHz class), the flatness and warpage class with panel size, the automotive environment (AEC-Q100 Grade 2 module class, wide-temperature lamination, surface finish), and the acceptance evidence — swept loss-trend and phase coupons, launch de-embedding, microsection proof of laser-via residual stub. Suppliers who ask for these before quoting are pricing engineering.
Building a 60 GHz in-cabin radar carrier?
Send us the sensor footprint and transition style, the antenna keep-out, the band and sweep, and the flatness class with panel size. We will review the laminate loss budget, the laser-via stackup, low-profile copper and the coupon plan at stackup review, and quote the carrier, baseband interface board and coupons to IPC Class 3 with automotive-qualified processing.
Contact Us → WhatsApp: +86 137 6047 3650Request a PCB Quote
Upload your Gerber ZIP file and project requirements. Our engineering team will review your PCB material, stackup, impedance needs, surface finish, and production quantity before quoting.
Please prepare:
- Gerber files in ZIP format
- PCB material or stackup requirements
- Controlled impedance notes if available
- Prototype or batch production quantity