🚀 go-pugleaf

In article <1993Jan5.134929@informatik.uni-kl.de> m_sattle@informatik.uni-kl.de (Matthias Sattler) writes:
>A few days ago, I tried to increase the CPU clock of my A540. But I encountered
>a strange phenomenon. The computer boots until the message RISC OS .... then
>the cursor stops flashing and the A540 obviously crashes. BUT when I press
>reset now the computer enters the desktop happily and works fine??!!!
>I can't even run the ARM at 30 Mhz without having this problem......

When the Archimedes powers up, a memory check happens to see :

1) if your DRAM is ok,
2) and to find out the optimum DMA parameters.

Probably your enhanced CPU clock confuses the test and wrong DMA parameters
are being used (that's why an ARM3 has to have its 'cache off' during that
test).  If you simply press 'reset' then this test is being skipped.  I
presume Risc OS will take standard values for the DMA parameters.

I believe also that the CPU clock in the A540 is also being used for other
chips (isn't there a clock share (with a few basic dividers) between the MEMC
and ARM3 ? My A540 & A5000 Technical Reference Manuals are at home...).

BTW, what happens if you do a CTRL-reset ?

>But ALL A5000 I know run fine at 30 Mhz or even more and they only have the 
>'bad' ARM3 (as the A540 have the selected good ones..).

The A5000 has other circuitery than the A540 (see the A540 and A5000 Technical
Reference Manuals).  So comparing the 'same' hardware upgrade for these two
machines is a bit trickier than what you do.

>(At the moment I use a 74F74 on the processor card to get the CPU clock, maybe
>this is important. By the way, what is the 68ohm resistior between the 7474
>and the ARM3 good for?)

Mmh, you probably never heard about reflexions in transmission lines due to
differences in impedance ? Electronics is a bit more than replacing crystal
clocks... ;-)

John
tytgat@esat.kuleuven.ac.be
BASS