So we carried out a modeling. This modeling is much more demanding than the previous one, because under weak p-a diffusion, we have to track down all the pitch angle drift, diffusion and variation along the inhomogeneous magnetic loop. I figure that because of the pa variation within a magnetic loop with field str. Variation it also depends on the magnetic field variation. We put a specific magnetic field variation of the loop being observed. And then electron evolution in phase space is obtained and  used for calculation of the microwave flux at the two observing frequencies. For a systematic search, we , 2. the spectral index variation. This is the first time to tell the time delay due to pitch angle diffusion in addition to the energy space diffusion, and this will later be found to be the key to explaining the observation.