home *** CD-ROM | disk | FTP | other *** search
- Path: yama.mcc.ac.uk!usenet
- From: Christos Dimitrakakis <mbge4cd1@fs1.ee.man.ac.uk>
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: 'Overclocking' the damn thing (060)
- Date: Wed, 17 Apr 1996 14:54:18 +0100
- Organization: Sirius Cybernetics Corporation
- Message-ID: <3174F80A.BA6@fs1.ee.man.ac.uk>
- References: <3165A359.13DB@ios.chalmers.se> <316A93DC.1A10@netvision.net.il>
- <927.6674T1136T2860@fuchal.demon.co.uk> <4khfpc$uvs@hearst.cac.psu.edu> <1233.6675T1133T1748@fuchal.demon.co.uk>
- NNTP-Posting-Host: meehph.ee.man.ac.uk
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.01 (X11; I; HP-UX A.09.03 9000/715)
-
- > TP> You can overclock it to 100Mhz if you want but it probably won't work for
- > TP> long.
- It will not work at all, as the clock signal will not arrive in sync
- with other signals in the chip. (note that it can never arrive exactly
- in
- sync, but increasing clock frequency makes matters worse.)
-
-
- >The rating on your 030 means that it passed all the tests for that
- Are u sure about that? I think they usually design chips so that they
- can
- run at higher speeds (clock buffering, handshaking speed/sync etc)
- and they dont just make and say "Hm...Let's see how fast it can go and
- then
- slap a max clock frequency on it".
-
-
- > TP> 40Mhz probably wouldn't work(maybe if it were super-cooled). 33 or even
- > TP> 35 might.
- Cooling is there only because higher temperatures change the resistance
- of
- silicon and sometimes even the shape of the chip. It does not help very
- much with
- speed. You still get the same signal delay ratios and thus your speed is
- still limited. You could normally expect to gain 1Mhz from a 40K
- decrease in temperature.
-
- >
- > Super-cooled? How about a open-cube ( |_| ) made out of 5 heatsinks screwed
- > together with a large fan on top? :)
- >
- > TP> Now what if you could rig a slider control up so you could set your clock
- > TP> speed anywhere between 0 & 100Mhz. Add a small digital readout to
- > TP> indicate the current speed ;-)
- That is only possible if none of the handshaking within the processor
- does not assume
- a maximum clock delay. About which I don't know.
- As I _think_ that it is the case for 680x0 processor, yes, you could in
- theory
- do something like that, though you should watch for the clock edges, as
- they
- might get too slow. Which is a MAJOR problem. The only real solution to
- having
- variable frequencies is that these be provided from a digital chip,
- perhaps with a small amp afterwards to make the edges sharper.
-
- >
- > Well, could I, for example, get a 33mhz crystal, and flick between these with
- > a switch? Actually, I never looked, is the crystal socked I wonder?
- >
- I think that most of the 68k series feature an external clock pin.
- You could hack your motherboard a bit and provide it. :)
- Note though that the length of the track between the 68k and the clock
- chip
- must be really small. (fast edge problems)
-
-
-
- Good luck. I hope you don't burn anything!
-