Re: AB + CAF

From: Doug Little <doug_at_innerworkings.co.uk>
Date: Tue, 10 Mar 1998 09:41:54 -0000

>Jesus!
>
> I mean, why do they say in the manual that you should keep the wires
>as short as possible, ok, the impedance will vary from capacitive to
>inductive depending on the lenght of the wire, but at 25Mhz, it barely
>changes anything!!! The real problem is the interfernce cause by all
>the crossed wires and the live components in the computer.... I mean I've
>got the proof, as soon as I shielded my DMA wire (the white wire), that
>annoying audio popping stopped!!!
>
>(this is simple engineering, when you design a piece of hardware, you make
>sure it is isolated from any interference, especially when you just need
>a shielded wire!!!)
>
>BTW: I suggest shielding all of the wires, better be safe than sorry!


Most of the problems we encountered were due to the shape of the
waveform coming from the buffer - the DMA chip clock pin is calibrated
in such a way that getting it to trigger properly is difficult. You really
have to go enter the realms of analogue electronics to fix the problem.

However, if there are further problems due to interference which lead to
clicks and pops, then shielding the cable is not going to do any harm - and
could even do a lot of good on some machines.

In the two instances where we attempted to shield the cables, DSP and audio
problems were not solved - something which was backed up by the fact that
the audio problems were made worse during SCSI access periods. This was
down to the clock waveform/DMA chip inputs being out of synchronization
with the DMA data bus and could this not be cured by shielding.

But, if shielded wires do make a significant difference to your machine,
then
by all means go for it!

Doug.
Received on ti. mars 10 1998 - 12:24:00 CET