More than six hours after the two-ribbon flare of 21 May 1980, the hard X-ray spectrometer aboard the SMM imaged an extensive arch above the flare region which proved to be the lowest part of a stationary post-flare noise storm recorded at the same time at Culgoora. The X-ray arch extended over 3 or more arc minutes to a projected distance of 95 000 km, and its real altitude was most probably between 110 000 and 180 000 km. The mean electron density in the cloud was close to 10⁹ cm⁻³ and its temperature stayed for many hours at a fairly constant value of about 6.5 × 10⁶ K. The bent crystal spectrometer aboard the SMM confirms that the arch emission was basically thermal. Variations in brightness and energy spectrum at one of the supposed footpoints of the arch seem to correlate in time with radio brightness suggesting that suprathermal particles from the radio noise regions dumped in variable quantities into the low corona and transition layer; these particles may have contributed to the population of the arch, after being trapped and thermalized. The arch extended along the H _∥ = 0 line thus apparently hindering any upward movement of the upper loops reconnected in the flare process. There is evidence from Culgoora that this obstacle may have been present above the flare since 15–30 min after its onset.