CompuServe Thread

#LOCK UP

2 messages in this thread
#33605From: Charles MichalFeb 18, 1992 6:19 PM
David, I noted your response to Joey Williams re Release 10c10a lockups. Please note the following (which I posted to all). Has anyone experienced anything like the following? I recently began using Version 11c2 on early Acer1100 80386's. One machine has 6 meg, the other 2 meg. One has an EGA card, the other a superVGA (using a ADI 4.0 driver). Both use Multisynch II displays. I use QEMM rather than HIMEM, and am running under DOS 5.0. I've read all the readme's and followed lots of discussion of the problems getting ACAD 386 and DOS 5 to work together. I fixed the B-step problem with a hardware patch from Ironwood Technologies. For a while I had these systems working. I kept tinkering with the 6 meg machine trying to optimize things and bring in some other software. Then out of nowhere I got a locked system while in Autocad. I can reliably lock the system by entering an arc or circle command, either from the command line or by any variety of menu or tablet. Once asked for the radius or diameter all entry devices (keyboard, mouse, tablet…) lock. The system also locks up when I try to load for editing an existing drawing containing such entities (or insert a block). The hard to believe part? This now happens on BOTH systems, yet only one had been tinkered with. Want more mystery? I've a third, almost identical, with the same hardware fix for the B-step problem, system that as of this writing does not experience these problems!!! My question – do I have Joey's problem, even with Release 11c2 and my hardware patch? Can you and Autodesk shed any light on this MAJOR problem? I'll give up and go back to Rel 10 soon now. Can anyone suggest something to keep me trying?
#34012From: David GariFeb 21, 1992 4:30 PM
The Ironwood socket will only fix Erratum 21. It will NOT fix Erratum 17. You have the very symptoms of Erratum 17. Have you installed the ERRATA17.EXP patch in ACAD.ADS in Release 11? It should work around the problem, but I recommend replacing all of the CPU chips found by CHKB1STEP (supplied w/ R11 c2) with D-Step 80386 parts to avoid future intermittant lockups in various applications. The reason it is intermittant is because the lockups only occur on certain math instructions during virtual memory paging.