21 Nov
1999
21 Nov
'99
7:24 p.m.
OK, I've discovered something, but have no idea what it means. For all
that goes, I am honestly not sure why I discovered it other than I was
fooling around with the front panel while checking voltages on the memory
module.
I had the following results when depositing 7777 to the memory.
Memory locations: xxx0 - xxx3 don't save anything
Memory locations: xxx4 - xxx7 save 7760
This holds true to at least location 0200.
On to other matters, I'm trying to do the resistance checks on pg. 4-16.
I've run into a couple of problems. It says that each winding for one bit
is
~3 ohms, however I show them at ~6 ohms.
Also step 6 tells me to check the resistance of diodes FSA2501. Um,
someone mind explaining how to do this exactly? Beware, I don't have the
equipment to put the core plane on extenders and run cables to the G104 &
G227 boards. Near as I can figure out they're a diode array from Fairchild
Semiconductor, however thier web site doesn't have a data sheet. Shouldn't
the resistance readings from one chip match that of another chip?
Oh, I do have the scope at home now, but haven't tried anything yet, or
checked into how horrible the probes are (I remember they're pretty
pathetic).
Zane
| Zane H. Healy | UNIX Systems Adminstrator |
| healyzh(a)aracnet.com (primary) | Linux Enthusiast |
| healyzh(a)holonet.net (alternate) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
22 Nov
22 Nov
6:13 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 5013 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991122/3e8fe20e/at…
23 Nov
23 Nov
8:47 a.m.
> I had the following results when depositing 7777
to the memory.
>
> Memory locations: xxx0 - xxx3 don't save anything
> Memory locations: xxx4 - xxx7 save 7760
My guess is that eitehr E46 is faulty, or it's
never being enabled. The
enable is applied to input D of the chip (pin 2) , from the gates in E34
(74H10). This is not going to be hard to trace out.
Found the chip, and assume I should test it with a Logic Probe and Pulser,
with the data in the back I think I might be able to figure that out. One
question, in looking through the pathetic little manual with the logic
probe and pulser I have (not even sure why I have them, I've had it since
the late 80's and never used them), it looks like the board needs to be in
the system and powered up, as they get their power from the circuit.
to the entire card. You might look at E24 (8881) and
E28 (384). These
chips are quad line drivers and receivers and it's possible for an entire
I'm guessing the next step is to also take the logic probe to these, but I
don't find any data on these chips...
Do you know if the CPU board is correctly handling data
on the bottom 4
lines? Since everything is open-collector, it's safe to try grounding (to
the system 0V rail) the MD lines during a read cycle to see if the CPU
sees them as asserted (everything is active low). And to monitor them
during a write cycle to see if the CPU can pull them low.
Was hoping to look into this, but it took so long to figure out the
O-scope, that I'm out of time today, gotta be at work in 15.
Unfortunatly the next couple days are going to be busy so not sure how much
time I'll have to work on this until Friday.
Zane
| Zane H. Healy | UNIX Systems Adminstrator |
| healyzh(a)aracnet.com (primary) | Linux Enthusiast |
| healyzh(a)holonet.net (alternate) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
10:24 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 4306 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991124/59515021/at…
12:13 p.m.
Forget the pulser, which isn't that useful,
actually...
Somehow that doesn't surprise me, and it's actually a relief to hear that!
(I assume you have a TTL databook giving the pinouts
of the common TTL
chips -- if not, GET ONE!).
I've got a copies of the TTL and CMOS Cookbooks, but suspect I should keep
my eyes open for actual databooks. If needed I think I know where I can
borrow one.
These sound like the ones that Radio Shack/Tandy were
selling. FWIW, I
You guessed it.
So don't think your logic probe is preventing you
finding the fault here.
It's not. It'll easily do the job.
Don't worry, even though it is from Radio Shack it looks more than up to the
task at hand.
to the entire card. You might look at E24 (8881) and
E28 (384). These
chips are quad line drivers and receivers and it's possible for an entire
I'm guessing the next step is to also take the logic probe to these, but I
don't find any data on these chips... during a write cycle to see if the CPU can pull them
low.
Was hoping to look into this, but it took so long to figure out the
O-scope, that I'm out of time today, gotta be at work in 15. For example. Put the logic probe on MD0 (MSB) on the
backplane and
deposit 7777. It blips low, right? Now examine that location (assumed to
be a 'working' location) -- it blips low again.
OK, one quick question on this that I've been meaning to clear up, but think
I've got right. '0' is the component side, and '1' is the back side,
right?
The manual seem to cover every other part of finding a specific pin except
that.
Zane
24 Nov
24 Nov
4:56 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 3004 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991124/a18284f1/at…
23 Nov
23 Nov
8:31 p.m.
New subject: Some Progress on -8/m Mem problems
> >> I had the following results when
depositing 7777 to the memory.
> >>
> >> Memory locations: xxx0 - xxx3 don't save anything
> >> Memory locations: xxx4 - xxx7 save 7760
>
> >My guess is that eitehr E46 is faulty, or it's never being enabled. The
> >enable is applied to input D of the chip (pin 2) , from the gates in E34
> >(74H10). This is not going to be hard to trace out.
Hmm, I'm almost safe with this thing, looks like it's a good thing that the
system appears to be fairly robust since it doesn't seem to be any more
messed up that it's been. Just realized I was testing the G227 board with
+15V hooked up to the Logic Probe, which explains why the probe was just
beeping, instead of showing the change from high to low and vice versa (of
course I realized this after swapping the G227 and G104).
However, it looks like E46 is indeed bad. It doesn't seem to function like
it should. On E39 some of the legs get a single blip every eight examines
(yes, I know that isn't a very good way of describing it, but I didn't take
notes, plus I wasn't using the probe right).
BTW, both E46 and E39 looked to be getting blips at the right spots on Pin 2.
to the entire card. You might look at E24 (8881) and
E28 (384). These
chips are quad line drivers and receivers and it's possible for an entire
I'm guessing the next step is to also take the logic probe to these, but I
don't find any data on these chips... For example. Put the logic probe on MD0 (MSB) on the
backplane and
deposit 7777. It blips low, right? Now examine that location (assumed to
be a 'working' location) -- it blips low again.
Now look at MD11 (LSB) and do the same things.
All 12 lines of the MD Bus look to be functioning both ways. This was
tested with the proper voltage hooked up to the probe.
Zane
| Zane H. Healy | UNIX Systems Adminstrator |
| healyzh(a)aracnet.com (primary) | Linux Enthusiast |
| healyzh(a)holonet.net (alternate) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
24 Nov
24 Nov
5:57 a.m.
New subject: Some Progress on -8/m Mem problems
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1243 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991124/e6fb9439/at…
8:14 p.m.
New subject: Some Progress on -8/m Mem problems
You can deduce the pinouts from the printset :
The problem here is the only PDP-8 printsets I've got access to are the
ones on Highgate. I suspect a lot of other people have the same problem.
8881 :
1 Output A
2 Input A1
3 Input A2
4 Output B
5 Input B1
6 Input B2
7 Ground
8 Input C2
9 Input C1
10 Output C
11 Input D2
12 Input D1
13 Output D
14 Vcc
Each gate is an open-collector NAND gate, e.g. Output A = Input A1 NAND
Input A2. On the G104 card, E24 :
Section Bit
A 11
B 10
C 9
D 8
Question: Isn't this functionally the same as a 7401, at least it looks to
me based on the above and looking at info on the 7401 as if should be.
DEC put a 4 digit code on these ICs. It's still a
384 logically, but it's
been selected for something (timing?). Of course _getting_ a replaement
now is non-trivial.
Unfortunatly the 384 looks rather non-standard compared to the info I've
got on hand (bit more than yesterday, while out taking care of stuff this
morning stopped at one of the local electronics places and picked up a NTE
Reference manual).
OK, one quick
question on this that I've been meaning to clear up, but think
I've got right. '0' is the component side, and '1' is the back
side,
right?
Off by 1... '1' is the component side, '2' is the track side.
MD0 means memory data bit 0, which according to the prints is on pin
AK1. You _do_ have the Omnibus pinout, right? However, it
looks like E46 is indeed bad. It doesn't seem to function like
it should. On E39 some of the legs get a single blip every eight examines
(yes, I know that isn't a very good way of describing it, but I didn't take
notes, plus I wasn't using the probe right).
That sounds like what it should be doing. I assume E46 is not doing this. OK, change E46. Now, where you get one (8251 decoder
IIRC, no relation to
the 8251 from Intel that's a USART) is another matter. I don't have any
spares :-(
Well, there is at least one person locally that might have some spares if
I'm lucky. I don't suppose they're used on MicroVAXen or if necessary
Q-Bus PDP-11's
For example. Put the logic probe on MD0 (MSB) on the
backplane and
deposit 7777. It blips low, right? Now examine that location (assumed to
be a 'working' location) -- it blips low again.
Now look at MD11 (LSB) and do the same things.
All 12 lines of the MD Bus look to be functioning both ways. This was
tested with the proper voltage hooked up to the probe.
25 Nov
25 Nov
5:54 a.m.
New subject: Some Progress on -8/m Mem problems
For example.
Put the logic probe on MD0 (MSB) on the backplane and
deposit 7777. It blips low, right? Now examine that location (assumed to
be a 'working' location) -- it blips low again.
Now look at MD11 (LSB) and do the same things.
All 12 lines of the MD Bus look to be functioning both ways. This was
tested with the proper voltage hooked up to the probe.
7:02 a.m.
New subject: Some Progress on -8/m Mem problems
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 2368 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991125/1ba9f1ce/at…
11:10 a.m.
New subject: Some Progress on -8/m Mem problems
Yes, it's very similar to the 7401 (and to one
other 74xx chip that I
forget).
There's some difference in drive capability and/or output saturation
voltage, but practically, if your Omnibus/Unibus isn't that long you can
use a 7401 as a bus driver in DEC machines.
OK, seeing as how the 7401 can more or less be substituted for a 8881 I've
got a question that belongs in the "Wacko File". I've been trying to see
if there are any equivalents to a 384 or 8251. I've not found one yet for
the 384, but the 7442 and the 7445 both look fairly simular to the 8251,
unfortunatly the pinouts are wrong. What is to stop me from building a
setup that will convert a 7442 to a 8251. Or doing the same thing for the
384 when I find a chip that will work?
It would probably involve running wires from where the chip is supposed to
go, to a socket with the actual chip. It looks to me like it would be
work, but I don't know what kind of problems that might cause.
Well, there is
at least one person locally that might have some spares if
I'm lucky. I don't suppose they're used on MicroVAXen or if necessary
Q-Bus PDP-11's
Unlikely. You might find the odd one on Unibus boards, but I'd rather not
strip one of those. 
11:22 a.m.
New subject: Some Progress on -8/m Mem problems
I've got some 8251's on some PDP-8 prototype boards, presumably they could
be carefully desoldered and re-used. If you are interested in trying this
Zane let me know. I think I have one board extra that I can send you.
--Chuck
12:04 p.m.
New subject: Some Progress on -8/m Mem problems
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1864 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19991126/f897a613/at…
9164
days inactive
9167
days old
13 comments
3 participants
participants (3)
-
ard＠p850ug1.demon.co.uk -
cmcmanis＠mcmanis.com -
healyzh＠aracnet.com