16 Jul
2002
16 Jul
'02
4:37 a.m.
On Tue, 16 Jul 2002, Gooijen H wrote:
Going back through my notes I see that the swap RX11
-> RX211
was the last action. So, back to the previous state: RX211 out,
RX11 in. Guess what. The hung condition is back.
The RX11 is a programmed I/O device. I knew that. But thanks
to Tony's e-mail that older cards *do not* wire CA1 to CB1 if
they do not use DMA, I checked the RX11. CA1 is not connected
to CB1. That explains the hung.
And what was the first thing most people said? Check your NPR line!
Perhaps we didn't stress it enough. :-)
But I saw that you had the funny assumption that if a card was inserted
you did't need the check the NPR line. Oh well, you learn something every
day. :-)
Anyway, I'm glad you got it working.
By experiment I will check the NPR wire of the other
UNIBUS slots.
You know why once you have seen the backplane of an 11/70.
Huh? It's no more difficult checking the NPR wire in the 11/70, than on
any other Unibus backplane. And it *is* rather simple to see if the jumper
is there or not. And it's much faster than doing empirical tests. Just
take a peek at the backplane, and you'll see it.
G7273 double grant in slot 41 --> THE HUNG IS BACK!
Huh??
A bad G7273? Inserted wrong? Something like that anyway.
Also learned what that "SWR" column means in
the M9312 manual.
SWitch Register, yes. :-)
Johnny
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt(a)update.uu.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol