Antenna pattern ripple in anechoic measurements: absorber placement tricks that actually work

If your measured antenna pattern has ripple that does not show up in simulation, check the chamber before blaming the antenna. I have chased this more than once, and the absorber layout is usually the culprit.
First, the quiet-zone reflection budget. Most small chambers quote a reflectivity number, but that number assumes the absorber is doing its job at your frequency. Below about 2 GHz, pyramidal absorber performance falls off fast, and a "quiet" zone can easily carry -25 dB reflections. That is enough to put visible ripple on a low-sidelobe pattern.
Second, the floor is the usual suspect. The direct path plus the floor bounce creates a classic two-ray interference, and the ripple period tells you the geometry: measure the angular spacing of the ripple, work back to the reflection point, and you will usually land on an uncarpeted floor strip or a metal positioner base.
Third, absorber placement beats absorber quantity. I put extra wedges on the floor reflection point and behind the antenna under test first, then on the side walls at the specular angles, and only then worry about the back wall. Randomly lining every surface wastes absorber and can actually make things worse by creating new edge diffraction.
A quick diagnostic: rotate the antenna under test by 90 degrees and re-measure. If the ripple moves with the antenna, it is the antenna; if it stays fixed in angle, it is the chamber. That one test has saved me more than one board spin.
What chamber artifacts have bitten you hardest - floor bounce, positioner scattering, or something stranger?

Thank you, RF_Bench_Notes — the “check the chamber before blaming the antenna” rule, plus the 90° rotate diagnostic, is exactly what we will keep linking under every “pattern ripple that sim does not show” report.

What we will keep quoting:

1. Quiet-zone reflectivity is frequency-dependent. Below ~2 GHz pyramidal absorber falls off fast; −25 dB reflections are enough to put visible ripple on a low-sidelobe cut. Quote the chamber number only with the frequency it was measured at.

2. Floor bounce is the usual two-ray culprit. Angular ripple period → geometry → uncarpeted floor strip or metal positioner base is the right reverse-engineering order.

3. Placement beats quantity: extra wedges first at the floor reflection point and behind the AUT, then side walls at specular angles, then the back wall. Random full-surface lining can create new edge diffraction.

4. Rotate AUT 90° and re-measure: ripple that moves with the antenna is the antenna; ripple fixed in angle is the chamber. That single test belongs in every first-article pattern checklist.

Open question for the room: when you cannot fully carpet the floor (positioner base / turntable), what absorber thickness / wedge density at the specular bounce point do you treat as “good enough” at 1–6 GHz before you stop chasing chamber ripple and start trusting the cut — and do you also require a co-/cross-pol pair to confirm the residual is not an orthogonal-mode artefact?

— rfforum 编辑部 / Editorial Team