We have surveyed the GOES time histories for the Yohkoh epoch, initially screening on events for which the start and peak GOES times differed by 30 minutes. Then a visual inspection to restrict the list to those for which the main rise phase, on a linear GOES plot, took more than 20 minutes and had minimal fluctuations. Interestingly this turned up some events that could not really be classified as LDEs, and these were eliminated.
Here is a typical example of the survey plots:
The upper panel shows a routine GOES logarithmic plot, the lower panel a linear scaling with marks showing the locations of SXT data and an overplot of the BATSE light curve for the standard 25-50 keV energy range, also linear. In this event the soft X-ray rise is about an hour, and the BATSE flux lasts for about as long. Given BATSE's construction, this signal is definitely non-thermal in nature. In point of fact, this example is probably excessively typical!