At 77 GHz the wavelength in the dielectric is about 2 mm, and every loss mechanism matters: dielectric loss, copper loss, surface roughness, and the via transitions between layers.
Material choice is not only about Dk. A low-Dk laminate with high Df can lose more than a slightly higher-Dk material with low Df. Roughness adds copper loss that no Dk number captures.
This article covers what to specify for 77 GHz radar, and where the stackup decisions actually change performance.
What matters at 77 GHz
- Dk (permittivity). Sets the trace width for a target impedance and the wavelength. Lower Dk means wider traces and lower dielectric loss.
- Df (loss tangent). Converts directly into insertion loss. At 77 GHz, small Df differences become large dB differences.
- Copper roughness. Current crowds near the copper surface; rough foil increases the path length and the loss.
- Dielectric thickness. Thin dielectrics concentrate the field and allow finer lines, but tighten impedance tolerance.
- Via transitions. A via stub or a poor launch can dominate the loss of a short line.
Material options
| Material | Dk (10 GHz) | Df (10 GHz) | Note |
|---|---|---|---|
| FR-4 | 4.2–4.6 | 0.015–0.020 | Not suitable at 77 GHz |
| Hydrocarbon | 3.0–3.5 | 0.002–0.004 | Good balance for radar |
| PTFE | 2.1–2.5 | 0.001–0.002 | Lowest loss, harder to process |
| Modified FR-4 | 3.5–4.0 | 0.008–0.012 | Cost compromise, limited at 77 GHz |
The numbers are typical and vary by supplier. The key is to compare Df and roughness, not only Dk.
Why roughness matters
At 77 GHz the skin depth in copper is a fraction of a micrometre. Standard copper foil has a profile comparable to that depth, so the current path follows the rough surface and the effective loss rises.
Two consequences:
- Smooth foil (low-profile or HVLP) reduces insertion loss.
- The same laminate can measure differently with different foil, so the material datasheet alone is not enough.
Specify the foil type, not just the laminate.
Stackup rules
- Keep the dielectric thin enough for the line width the fabricator can etch; a 0.1 mm line on a thick dielectric needs very low Dk.
- Use the same material for the layers that carry the radar signal; mixing materials changes the effective Dk.
- Control the via stub; at 77 GHz a stub of a few tenths of a millimetre can be significant.
- Avoid connectors with long launches; the launch can cost more loss than the line.
- Specify impedance tolerance and the coupon requirement together.
DFM and assembly
- Etch tolerance on narrow lines is a larger fraction of width; expect tighter impedance spread.
- Solder mask over the RF line changes the effective Dk; specify mask or no-mask.
- Assembly materials near the antenna can detune it; keep the bill of materials clean.
- Conformal coating may change antenna performance; test with the coating applied.
FAQ
Can FR-4 be used at 77 GHz?
Not for the radar signal path. Loss and Dk variation are too high. FR-4 may still be used for control layers in a hybrid stackup.
Is lower Dk always better?
Not always. Lower Dk allows wider lines and lower loss, but the material must also be processable and available in the required thickness.
Does copper roughness affect impedance?
It affects loss more than impedance, but it does change the effective phase velocity slightly. The main effect is insertion loss.
Do I need a coupon for 77 GHz?
Yes. Measure the coupon with the same foil and stackup, and state the measurement frequency.
Material datasheets rarely answer the only question that matters: what does this stackup lose at 77 GHz with the foil you will actually use? Send the stackup and the loss budget, and we will run the numbers before quoting.
