home *** CD-ROM | disk | FTP | other *** search
- Path: isar.de!news
- From: imd@m.isar.de (Thomas Doerfler)
- Newsgroups: comp.sys.m68k
- Subject: Re: 68302 DRAM Refresh & External Masters
- Date: Tue, 09 Jan 1996 08:19:47 GMT
- Organization: IMD
- Message-ID: <4ct87t$jed@nixe.isar.net>
- References: <NEWTNews.821149036.27321.kevinb@king.tutsys.com>
- NNTP-Posting-Host: imd.m.isar.de
- X-Newsreader: Forte Free Agent 1.0.82
-
- Kevin Braun <kevinb@tutsys.com> wrote:
-
-
- >Hi all,
- >I am looking for signal timing and sequence for using the '302's DRAM refresh
- >with asynchronous bus masters (i.e. 83902 Ethernet MAC). This subject doesn't
- >seem to be covered in the data book. While the 83902 is mastering, it can hold
- >the bus for a number of refresh cycles. I don't see any way of using !BCLR to
- >stop the 83902 from releasing the bus before it wants to. Does anyone have any
- >experience in this area or point me to some reference material? Thanks
-
- >..kevin braun
-
- Well, as far as the User's Manual (Rev 2) mentions, you really have a
- problem. The Refresh Controller doesn't queue refresh requests, so if
- a given request is not granted within one refresh interval, it is
- lost.
-
- Maybe you don't have a problem at all, because when the MAC really
- writes a block to your DRAM, it automatically refreshes it (but just
- the rows it wrote to...)
-
- Did you estimate, how long the MAC will use the bus at most? If it is
- less than two refresh periods, it might still be acceptable...
-
- Or can you change your hardware to the MC68EN302 (internal ethernet
- controller)? I saw a product brief some weeks ago, quite interesting
- chip...
-
- The only solution I can guess is a sort of an external up/down
- counter. It will be incremented every time the refresh counter
- expires, counting the number of refreshs needed. And for each refresh
- cycle performed (look for special function code for this...), you can
- decrement it again. As long as it is not zero, refresh request should
- keep active. This solution has a drawback, the 68302's CP will be busy
- refreshing quite a long time and therefore your SCCs might be stuck.
- But they will be stuck aswell, if they cannot reach their buffers.
-
- Hm, bye,
- Thomas.
- --------------------------------------------
- IMD Ingenieurbuero fuer Microcomputertechnik
- Thomas Doerfler Elilandstrasse 12
- D-81547 Muenchen Germany
- email: imd@m.isar.de
-
-