20 Aug
2002
20 Aug
'02
3:43 p.m.
On Tue, 20 Aug 2002, Tony Duell wrote:
The drive will work fine if it has a '3' lamp cover and is the only drive
connected to the controller. But you must have a 'ready' lamp cover
fitted -- don't assume that if none is installed then the drive will
default to 3. In fact, it won't work at all -- there's an extra contact
in the ready lamp holder/select switch that detects that _a_ cover is
installed.
Yes, it has the '3' ready lamp cover. I also have a 1 or 2 I think, but
definately not a '0'.
Agreed.
I've found someone who will loan me a terminator (thanks Tom!), so once I
get that I can further evaluate the situation.
One cause of 'Fault' is if the clock signal from the controller is
missing. It's around 4.2MHz. But IIRC, if there's no terminator on the
drive bus then the drive won't properly detect the clock (or any other
signals from the controller) so you will get this error.
Incidentally, the clock frequency is critical for running the drive (it
determines the spindle speed amongst other things), but for testing the
fault circuit it's not particularly critical.
That's what I thought... now if I run it at a differenct frequency, can I
get a different data rate out of it? I've thought about creating a
controller that is ISA bus or something, but needed the whole setup to do
any reverse-engineering, of course. Also, I've noticed it's hard to find
8.4 or 4.2MHz (or is it 8.2 and 4.1MHz?) oscillators.
I will confirm that the RL's don't mind if you get the cable in
upside-down. I've done it (using official DEC cables, etc) and never
damaged anything.
Good. I just wanted to make sure I didn't make a big 'boo-boo'.
-- Pat
The drive is set to ID 3, because I don't have a
'0' Ready light cover.
Also never done that. Don't remember if it works or not. We were
always short of 0s and 1s (but seemed to always have a box of 2s and
3s ;-) The RK06/RK07 drives use identical covers, but they
are numbered from 0
to 7 (you can have up to 8 such drives on a single controller). These
covers will work in RLs. The drive number is taken mod 4 in this case, so
that cover 4 -> drive 0, cover 5 -> drive 1, etc
I'll remember that if I come across any.
Also, I don't have a terminator to use.
Major problem.
Previously, when I connected a clock signal to the
drive's clock in,
which was anywhere near 4.2MHz, the fault light would go off.
Hmm... interesting. Never played with RL drives at the signal level.
Not sure what that would do. IIRC, the DEC cable doesn't connect all the pins
(some are just not used)
-- the RK06/RK07 uses a similar-looking cable with more pins wired.
One of the corner pins is deliberately not bussed between the drives.
It's a 5V output from the drive's PSU and is used to power the
terminator. Needless to say, connecting 2 power supplies together is
normally a bad idea. IIRC, though, it's a no-connect on the controller,
so it shouldn't matter if you connect this pin on single-drive system.
On my PDF version of the RL Pocket Guide, it says that all the non-signal
pins are wired to ground... Hmm. There's a copy of it here:
http://www.dadaboom.com/pdp11/rlPocketGuide.pdf
I'll go home and check this with a multimeter tonight.
I also have tried reversing the cable to ensure I
had
pin 1 the same on both ends, to no avail (could that have nuked the
interface on either the controller or drive?).
No. DEC built their stuff to withstand upside-down and backwards
cables (but not modules - a former boss of mine watched an FE