4 Nov
1997
4 Nov
'97
4:36 a.m.
On Tue, 4 Nov 1997, Tony Duell wrote:
Okay... I'll give it a shot! But if I foul up, I'm stuck - I don't have
a wirewrap tool.
If you can boot RT11, then the grants must be continuous. Is the RX211
nearer the CPU than the UDA50? Normally, if you've left an NPG jumper in
where you shouldn't have, then the devices past that point fail to do
NPR's correctly.
It's closer.
You don't mean the CGJ card, do you? There's something else I'm supposed
No, I mean on the backplane. The little single-height continuity cards
only connect the bus grants. The dual-height ones do the NPG signal as
well, _but_ the jumper I mentioned below is still fitted, since a lot of
older cards do not short CA1 to CB1 themselves, so the wire-wrapped link
is need if you use one of those cards.
to have screwed with. You mean I have to cut one
of the wire-wraps on the
backplane? That doesn't sound like a good idea... is that correct? When
Absolutely. The wire in question is between pins CA1 and CB1 on the _same_
slot - the slot the control cart of the UDA50 is in. It's a little loop of
wire, and is quite difficult to find the first time.
When I re-wrap them (say after removing a card), I normally use red wire
to make them easier to spot the next time.
The jumper shorts NPGin to NPGout on the slot. It must be fitted on slots
that do not contain options that use DMA (either empty slots, or slots
that contain options that use programmed I/O only - like DL11's, DR11-C's,
etc). It must be removed on slots that contain DMA options (like the
UDA50). If you just pulled a single-height grant card and fitted the UDA50
card, then I am _sure_ you need to cut this jumper as well. Otherwise
NPGin is shorted to NPGout on the UDA50 and the internal grant logic goes
crazy on this card (it won't do any damage, though).
Oh! That's what's happening... OK, I'll go look at it.
Thanks!