Model with beamed injection (n10, μ, Δμ)
A beamed injection:
 Δθ  = 30°
near vertical incident angle,
   μο= 0
into a low density trap
n = 4 ·109 cm-3
at a scale of
τinj = 32 s fits the observed microwave spectra.
This conclusion is also new, I think, enabled by microspectrum. it works because microwave radiation has the sensitivity to the pitchh angle distribution as well as energy distribution. What would be for hard X rays? Thick target –better for injection properties. This is slower process and could be timely resolved as spectral evolution.