cavedon@hslrswi.hasler.ascom.ch (Gilbert Cavedon) (09/25/90)
(I am posting this request for a friend) I made this changes on my Amiga : - Replace the 7Mhz Version of the 68000 with a 14Mhz 68000 using the 28Mhz clock provided by the Fat Agnus via a D-Flip-Flop. I have still some problems; I can't access neither the hard-disk nor the floppy. Only the bootblock of a normal disquette is read, if you use a NODOS disquette then it works. I think this problem is due to wait states that can't be see by the devices when the clock speed is increased. Does anybody know what to do? Has anybody made the same changes to his machine? If there is a program that can solve this problem can anybody post it to the net? Thanx. +--------------------------------------------------------------------+ | Gilbert CAVEDON voice : +41 31 63 12 50 | | Ascom Hasler AG email : cavedon@hslrswi.hasler.ascom.ch | | Berne, Switzerland | +--------------------------------------------------------------------+
thyssen@batserver.cs.uq.oz.au (Anthony Thyssen) (09/26/90)
cavedon@hslrswi.hasler.ascom.ch (Gilbert Cavedon) writes: >I made this changes on my Amiga : >- Replace the 7Mhz Version of the 68000 with a 14Mhz 68000 using the > 28Mhz clock provided by the Fat Agnus via a D-Flip-Flop. > >I have still some problems; I can't access neither the >hard-disk nor the floppy. Only the bootblock of a normal disquette is >read, if you use a NODOS disquette then it works. >I think this problem is due to wait states that can't be see by the >devices when the clock speed is increased. I am currently doing timing on the CAI chips and can tell you straight out that that is one place a problem lies. The `E' line from the 68000 is clocking at a speed of 1/10 of the incoming cpu clock line This is done so as to insure that the `E' line is less than 1Mhz in speed as the chips that use it can't go above this speed (the CIA) (thay were originally designed for the 6800 1Mhz cpu (like the C64) This means that your CIA's are being over clocked and and if thay are running are timing system events at double speed. The vidio parts of the 2 CIA's is still find as thay use the 50/60Hz clock signal. The Disk drive, and general system timer on CIA-A timer B however is running at double speed - And thus interuping too early. Solutions :- Reprogram the timing interupts! OR half the clock output of the 68000 cpu `E' line. Thanks for letting me blab on.... Anthony Thyssen - (Dragon Computing!) thyssen@batserver.cs.uq.oz.au ------------------------------------------------------------------------------- ``Never argue with a computer.'' -- Avon, Blake's 7 -------------------------------------------------------------------------------
billsey@agora.uucp (Bill Seymour) (09/27/90)
In article <1560@hslrswi.UUCP> cavedon@hslrswi.hasler.ascom.ch (Gilbert Cavedon) writes:
:(I am posting this request for a friend)
:
:I made this changes on my Amiga :
:
:- Replace the 7Mhz Version of the 68000 with a 14Mhz 68000 using the
: 28Mhz clock provided by the Fat Agnus via a D-Flip-Flop.
:
:I have still some problems; I can't access neither the
:hard-disk nor the floppy. Only the bootblock of a normal disquette is
:read, if you use a NODOS disquette then it works.
:I think this problem is due to wait states that can't be see by the
:devices when the clock speed is increased.
:
:Does anybody know what to do? Has anybody made the same changes to his
:machine? If there is a program that can solve this problem can anybody
:post it to the net?
There has been a discussion on this topic a couple of times
before. What you have to do is generate the syncronous bus signals
yourself. VPA, VMA and E-Clock are expected by the 8520s to be syncronous
to the 7MHz system. Since you're generating these at twice the normal
speed, the timings that the 8520s use are wrong. The most typical
failure is in the floppy drives. The time between turning the drive motor
on and when you're allowed to write is determined by the 8520s, and
when things happen too quickly there, you run the risk of either trying
to read or trying to write to a drive that isn't spinning quite fast
enough. If your drives spin up quickly, you won't see the problem.
If your drives spin up slowly, they'd be unusable. I can't see where
that would affect your HD, but they may be using the slow clocks
somehow...
:Thanx.
:
:+--------------------------------------------------------------------+
:| Gilbert CAVEDON voice : +41 31 63 12 50 |
:| Ascom Hasler AG email : cavedon@hslrswi.hasler.ascom.ch |
:| Berne, Switzerland |
:+--------------------------------------------------------------------+
--
-Bill Seymour ...tektronix!reed!percival!agora!billsey
=============================================================================
Bejed, Inc. NES, Inc. Northwest Amiga Group At Home Sometimes
(503) 281-8153 (503) 246-9311 (503) 656-7393 BBS (503) 640-0842martijn@dnlunx.pttrnl.nl (Reinalda M.) (09/27/90)
Well, I did the same thing on my 500. The internal drive had the same problem but my external 5"1/4 drive worked fine. I did some timing measures on the track-step and it seemed that the 3"5 had a much lower track-step-speed then the 5"4. So I replaced the drive with a faster 3"5 drive and it loads fine!! We come in peace, shoot Intel.
martijn@dnlunx.pttrnl.nl (Reinalda M.) (10/09/90)
Here's what I found found out sofar : Halving the CLK works. But the drives go way too fast. Halving CLK and E doesn't work. Halving CLK, VPA, VMA and E doesn't work. Gee, I'm stuck !!