From: wolff@liberator.et.tudelft.nl (Rogier Wolff) Subject: Re: Wild Interrupt? (IRQ7) - huh? Date: Tue, 16 Mar 1993 17:57:27 GMT
root@umibox.hanse.de (Bernd Meyer) writes:
>I dived into this yesterday - Seems like something is really broken in
>hardware (at least with me). I got the iwld interrupt IRQ7 even with just
>the at-bus controller and the vga installed. It went away
>a) when you unconnected the harddisk (well, no solution to me)....
>b) when you used another, 8 bit vga (occured rarely, though)
>c) when you used the original vga in an 8 bit slot (occured rare, too)
>BUT: Adding some more boards (i/o, soundblaster et al) made the situation
>worse and worse, no matter what boards were used. Also, with the original 2
>board setting, both boards don't connect ANYTHING to IRQ7 (this is pin B21).
>So I've decided I'll blame a bad termination for now. But I still wonder -
>IRQ 7 worked somed time ago.....
>I've posted about this to an ibm-pc hardware group. Maybe they'll know
>Bernie
The problem is that every card that uses an IRQ should drive the signal.
That is it provides both the low and the high level. Anybody who ever
designed a computer knows that you should specify a 1k resistor to
power on the motherboard and an open collector output on the devices
that drive it. IBM didn't do this back in 1982 when they designed the
PC.
I recently came across a story from someone who had the same problems
It seems that the floating IRQ7 picks up some Electro magnetic interference,
and goes active every now and then. Before you didn't notice, since the
thing was not enabled. For debugging purposes Linux now prints the
interrupt....
Roger.
-- **** a 486 in V86 mode is like a VW buggy with a 6 litre V12 motor. **** EMail: wolff@duteca.et.tudelft.nl ** Tel +31-15-783643 or +31-15-142371