Msg: 2524 *Conference*
05-01-92 10:43:46
From: LEX JENKINS
To : TRACY ALLEN
Subj: MORE GREMLIN STUFF
Hi, Tracy, I've done a few more experiments in trying to duplicate the UR-II related crashes, with these results: 1. The extRAM image appears to be the most resistant to crashing, especially since I relocated the blue wire/pin to #40 on the 100's bus. In fact, I can now Plot preview multi-merged files (both comma-delimited data AND text file merges) without getting an OM error. 2. The ROM installed in the 102's Booster Pak is the least crashworthy. Crashes while Plot-previewing data-text merges usually corrupt all files in RAM, and even seem to corrupt the ROM itself, although cold-starting and reinitializing the ROM cures the problem. 3. I disconnected the Booster Pak and installed the ROM directly into the 102. While Plot-previewing merged files still caused OM errors, pressing reset unlocked the 102 and no files were corrupted. All in all, considering these limited and not particularly scientific experiments, it appears the extRAM maintains a very stable image with the relocated blue pin. I'm speculating that the Booster Pak somehow uses the ROMs differently than does a stock 100 or 102, leaving it somewhat more susceptible to crashes in extreme circumstances. I say extreme because I experienced no problems whatsoever with UR-II after several months of extensive use, until I began really pushing T-base data file and T-word text file merges to the limits of the system's capacity. I'll also acknowledge it's possible that I actually exceeded the ROM's limits in constructing my T-base screen files, although I was very careful to stay just below those limits. In the end, I modified the screen files to leave a much wider safety margin, and have had no further problems - unless I try to Plot preview. I'm going to do a write-up on this experience and offer it for the TechRef section. Despite the hangups, I still believe data-to-text file merges are a very efficient way to utilize the limited RAM. <Lex>