Angular power spectra for 30 GHz null tests. Pseudo-spectra are calculated on 80% of the sky, with the Galaxy and point sources masked. All spectra are corrected for sky fraction but not for beam smearing effects. For comparison, we also show the simulated odd-even survey difference spectrum and the best-fit cosmological model spectrum filtered by the beam window function. For ℓ ≥ 30 the spectrum of the survey 1 minus survey 2 map difference fully coincides with the half-ring difference spectrum calculated for the same time period, while for the 30 GHz channel there is a small sidelobe contribution at the 10 μK² level on larger scales.