thank you for reaching out. It appears that we have some RA values in overlap regions that ended up with a (wrong) 180 deg offset. We are correcting it as we speak. I'll comment here again when this is done.
Thank you for alerting us to the problem.
One unrelated recommendation: if you care that the search radius be always the FOV you specified, please do:
imgTable = svc.search((ra,dec), (fov/np.cos(dec*np.pi/180), fov), verbosity=2).to_table() # ra,dec,fov in degrees
to account for the cos(dec) shrinkage.