Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!uunet!stanford.edu!leland.Stanford.EDU!elaine27.Stanford.EDU!alee From: alee@elaine27.Stanford.EDU (Andrew Lee) Newsgroups: comp.protocols.appletalk Subject: Re: LocalTalk Cards on 33 MHz 386 Message-ID: Date: 9 Mar 91 09:18:24 GMT References: <17320@crdgw1.crd.ge.com> <1991Mar7.130114.1427@urz.unibas.ch> Sender: news@leland.Stanford.EDU (Mr News) Organization: AIR, Stanford University Lines: 32 I've tried both Apple LocalTalk cards and TOPS Flashcards in my 25 MHz 386, and neither was entirely satisfactory. My machine is a AT-bus clone with 64K SRAM cache and Phoenix BIOS, and the software I'm using is Wollongong's WIN/TCP TCP/IP software, with board drivers from the manufacturers, and a packet driver written here at Stanford. The Apple card doesn't work when my machine is at 25 MHz, or if I have QEMM-386 loaded, so I have to run my machine at 8 MHz without QEMM. I have version 2.00 of Apple's drivers, and LTALKP.EXE fails if my machine is at 25 MHz or QEMM is loaded. LTALK.EXE, the interrupt-driven version of LTALKP, doesn't work under any circumstances. I suspect it's a DMA problem, but LTALKP won't work without DMA. I called Farallon, and they can't get me Apple's version 2.10 drivers, and won't have their own version 2.20 drivers out for a while. Now when I tried the TOPS Flashcard, to get TOPS' drivers ALAP and PSTACK to load under any of the following conditions, I had to have DMA disabled: running at 25 MHz, or having either HIMEM.SYS (from Windows 3.0) or QEMM-386 loaded. Then, no matter what I did, when I tried, when I attempted to load the packet driver, it reported that it couldn't find the gateway. The people at Networking and Communication Systems here at Stanford, who wrote the packet driver (same for both TOPS and Apple cards) told me that it means that it can't talk the the Kinetics box which is the only other thing on this line (about 1300' of twisted pair, plus another 50' of my flat wire), and that they have found that TOPS boards seem to have trouble with cable lengths in their own building. They said that the LocalTalk networks here at Stanford have been checked and found to be within LocalTalk specs. I've tried sending email to TOPS, but they never sent an intellible reply. Andrew Lee alee@leland.stanford.edu, alee@slacvm.bitnet