The visibility amplitudes for the phase ref are almost constant as a function of projected baseline length (uv distance in metres), showing that it is point-like to e-MERLIN with a resolution of 50 mas.
The amplitudes are noisiest out to about 130 km because these are the Defford baselines and that telescope is least sensitive. There is a little data with anomalously low amplitudes but because we know that the target is bright we won't worry about that for now.