12 Jun
2008
12 Jun
'08
2:15 p.m.
---------Original Message:
From: "Steve Maddison" <steve at cosam.org>
Subject: Fixing a DEC Letterwriter (LA100)
Hello all,
I have a half-dead Letterwriter 100 KSR here I'd like to get working
again. I get a short buzz from the paper feed motor at power-up but
instead of flashing all it's indicator LEDs, both CTS and LINE stay
illuminated. There's no carriage movement and the keyboard does
nothing either. The basic vital signs are OK but I'm stuck without a
service manual or schematics. Does anyone have access to these docs?
I'd of course also be interested to hear of any common problems and/or
things to check before I dive into all-out mending mode...
Cheers,
--
Steve Maddison
http://www.cosam.org/
----------Reply:
Can't suggest where to start checking but I do have several RO LA100s
awaiting the trip to the scrapper, in case I can help with any non-kbd parts
before they go.
I do also have the tech manual somewhere; unfortunately IIRC it's on 17"
paper so I can't scan it.
I believe Tony also has a copy and more knowledge/experience than I;
maybe he has some suggestions.
m
11 Jun
11 Jun
11:56 p.m.
Can't suggest where to start checking but I do
have several RO LA100s
I guess I'd start by checking all the PSU outputs. I doubt that's the
problem here, but PSU problems can cause the most amazing faults...
awaiting the trip to the scrapper, in case I can help
with any non-kbd parts
before they go.
I do also have the tech manual somewhere; unfortunately IIRC it's on 17"
paper so I can't scan it.
I believe Tony also has a copy and more knowledge/experience than I;
maybe he has some suggestions.
I do have the printset, but it would take me a little time to find it. As
ever I have no scanenr, so I can't provide a copy.
As regards experience, I've worked on these printers, I still have a
couple operational. But I've not really had any trouble with them, so I
can't claim any experience of fixing them.
-tony
13 Jun
13 Jun
4:17 a.m.
2008/6/12 Tony Duell <ard at p850ug1.demon.co.uk>:
That's indeed where I started and, according to the multimeter, all is
OK. Could of course be something gruesome going on that I will only be
able to detect when I take a scope to it...
Can't
suggest where to start checking but I do have several RO LA100s
I guess I'd start by checking all the PSU outputs. I doubt that's the
problem here, but PSU problems can cause the most amazing faults...
I do have the printset, but it would take me a little
time to find it. As
ever I have no scanenr, so I can't provide a copy.
Don't worry - I have found the prints on Manx (they were "hiding" on a
second page of search results that I managed to miss). For the
archive: http://www.vt100.net/manx/details/1,4232
And thanks for clearing up the 7406 story - never came across them before.
--
Steve Maddison
http://www.cosam.org/
12 Jun
12 Jun
10:44 p.m.
I guess
I'd start by checking all the PSU outputs. I doubt that's the
problem here, but PSU problems can cause the most amazing faults...
That's indeed where I started and, according to the multimeter, all is
OK. Could of course be something gruesome going on that I will only be
able to detect when I take a scope to it...
I do have the printset, but it would take me a
little time to find it. As
ever I have no scanenr, so I can't provide a copy.
Don't worry - I have found the prints on Manx (they were "hiding" on a
second page of search results that I managed to miss). For the
archive: http://www.vt100.net/manx/details/1,4232
13 Jun
13 Jun
6:09 p.m.
2008/6/13 Tony Duell <ard at p850ug1.demon.co.uk>:
IIRC this unit is microprocessor-controlled 8085?) and
the LEDs are under
the control of that microprocessor.
Absolutely correct. Everything between the processor and the LEDs in
question looks to be fine so I suspect the program just isn't running
right.
What does it use for RAM? If 2114s, then the lucky-dip
side of me says to
check/replace them. I've had _dozens_ of that chip fail in all sorts of
devices, with various odd results.
Right again, and thanks for the tip.
Other than that, I guess you start by checking the
processor clock and
reset signals, then see if there's any activity on the data and address
buses.
Indeed - reset looks good and pulling it low gives an audible
kick/buzz from one of the motors, just like at power-up. AFAICS the
processor's running and not halted or waiting, which would favour your
RAM hunch. The stickers on the EPROMS' windows are a bit crusty too,
and starting to shine through, so I'll add checking for bit rot to the
list as a last resort.
Cheers,
--
Steve Maddison
http://www.cosam.org/
9:42 p.m.
2008/6/13 Tony Duell <ard at p850ug1.demon.co.uk>:
You know, I can't rememebr the face of somebody I met yesterday, but I
rarely forget a scheamtic. Even though (in this case) I've not worked on
an LA100 for over 10 years...)
IIRC this unit is microprocessor-controlled
8085?) and the LEDs are under
the control of that microprocessor.
Absolutely correct. Everything between the processor and the LEDs in question looks to be fine so I suspect the program
just isn't running
right.
I think the 2114 is the most unreliable chip I've ever seen. And it
doesn't seem to be any particular manufactuer that's to blame either. I
would certainly check these very carefully.
What does it use for RAM? If 2114s, then the
lucky-dip side of me says to
check/replace them. I've had _dozens_ of that chip fail in all sorts of
devices, with various odd results.
Right again, and thanks for the tip. RAM hunch. The stickers on the EPROMS' windows are
a bit crusty too,
and starting to shine through, so I'll add checking for bit rot to the
list as a last resort.
I seem to have dumps of 147E4 and 148E4 here...
-tony
14 Jun
14 Jun
4:24 p.m.
2008/6/14 Tony Duell <ard at p850ug1.demon.co.uk>:
I think the 2114 is the most unreliable chip I've
ever seen. And it
doesn't seem to be any particular manufactuer that's to blame either. I
would certainly check these very carefully.
I whipped all the 2114s out today and popped in a set of IC sockets.
Thinking they couldn't possible /all/ be bad, I got two replacements,
and have experimented with different combinations of old and new. Very
scientific, I know ;-) I thought it was worth a try before hacking up
a test rig.
Anyway - the original problem is gone. With the (half) new RAM I'm now
getting what at least appears to be a proper fault indication on the
LEDs: both DSR and POWER/FAULT are flashing. This combination isn't in
the user manual so my guess is it's failing the POST. If any of the
2114s were still working, I probably fried them getting them out. I
see if I can borrow a couple from a ZX81 or something else lying
around here.
I seem to have dumps of 147E4 and 148E4 here...
Was just looking for those! I'd sure appreciate a copy if you could
mail them or somehow get them online. I might try my luck at tracing
through a bit of code to see which conditions are causing the light
show.
--
Steve Maddison
http://www.cosam.org/
13 Jun
13 Jun
10:54 p.m.
2008/6/14 Tony Duell <ard at p850ug1.demon.co.uk>:
I wouln't put much money on that :-)
I think the 2114 is the most unreliable chip
I've ever seen. And it
doesn't seem to be any particular manufactuer that's to blame either. I
would certainly check these very carefully.
I whipped all the 2114s out today and popped in a set of IC sockets.
Thinking they couldn't possible /all/ be bad, I got two replacements, and have experimented with different combinations of
old and new. Very
scientific, I know ;-) I thought it was worth a try before hacking up
a test rig.
Anyway - the original problem is gone. With the (half)
new RAM I'm now
getting what at least appears to be a proper fault indication on the
LEDs: both DSR and POWER/FAULT are flashing. This combination isn't in
the user manual so my guess is it's failing the POST. If any of the
Maybe soemthing like 'RAM Error' :-)
Incidentally, 2114s fail in different ways. I've had data bits stuck high
or low, I've also had address lines ignored (so that the same data is
repeated twice). If you're going to test them, you want to check them
pretty thoroughly, not jsut make sure you can store something.
I seem to have dumps of 147E4 and 148E4 here...
Was just looking for those! I'd sure appreciate a copy if you could
mail them or somehow get them online. I might try my luck at tracing through a bit of code to see which conditions are
causing the light
show.
IIRC, there is a custom chip in this machine, which sits on the processor
bus. I ahe no idea what the registers in that chip do, so any read/write
access to that is essentially unknown.
-tony
16 Jun
16 Jun
5:06 a.m.
2008/6/14 Tony Duell <ard at p850ug1.demon.co.uk>:
That was my first guess too. Running the ROM through a simulator and
forcing the RAM test to fail confirms it. The problem? I'd managed to
gouge out a bit of track while removing the ICs - just so happens it
was one of the very tracks connecting /all/ the RAM to the address
bus, so checking for continuity between the sockets I'd soldered in
didn't reveal the problem.
So I fixed that and robbed a couple more 2114s from another machine.
So all RAM is replaced but, lo and behold, ONLINE and CSR are back
with a vengeance. So I'm basically back to square one, although not
entirely as it would at least appear the RAM is OK. It could of course
still be failing in one of the peculiar ways you mentioned and still
pass the POST... Otherwise it would appear the original LED
combination was directly asserted by the processor, so it should be
traceable in the ROM code.
Anyway - the
original problem is gone. With the (half) new RAM I'm now
getting what at least appears to be a proper fault indication on the
LEDs: both DSR and POWER/FAULT are flashing. This combination isn't in
the user manual so my guess is it's failing the POST. If any of the
Maybe soemthing like 'RAM Error' :-) > > I seem to have dumps of 147E4 and 148E4
here...
...
I'll get them on my way to you...
Many thanks! You may also want to send them on to Pete Turnbull for
inclusion in his DEC ROM archive.
IIRC, there is a custom chip in this machine, which
sits on the processor
bus. I ahe no idea what the registers in that chip do, so any read/write
access to that is essentially unknown.
I've not come across it yet; let's hope it stays that way ;-)
--
Steve Maddison
http://www.cosam.org/
"The great thing about standards is that there are so many to choose from."
15 Jun
15 Jun
11:08 p.m.
Anyway - the original problem is gone. With the (half)
new RAM I'm now
getting what at least appears to be a proper fault indication on the
LEDs: both DSR and POWER/FAULT are flashing. This combination isn't in
the user manual so my guess is it's failing the POST. If any of the
Maybe soemthing like 'RAM Error' :-) gouge out a bit of track while removing the ICs - just
so happens it
was one of the very tracks connecting /all/ the RAM to the address
bus, so checking for continuity between the sockets I'd soldered in
didn't reveal the problem.
So I fixed that and robbed a couple more 2114s from another machine.
So all RAM is replaced but, lo and behold, ONLINE and CSR are back
with a vengeance. So I'm basically back to square one, although not
Ah, but you've confirmed that the processor is doing something, and that
it's executing the firmware (at least partially). If not, then the RAM
failure, (albeit a man-made one) would not have been detected. Surely
that is a staep forward.
> > I
seem to have dumps of 147E4 and 148E4 here...
...
I'll get them on my way to you...
Many thanks! You may also want to send them on to Pete Turnbull for
inclusion in his DEC ROM archive. 
14 Jun
14 Jun
5:02 p.m.
Tony Duell wrote:
I think the 2114 is the most unreliable chip I've
ever seen. And it
doesn't seem to be any particular manufactuer that's to blame either. I
would certainly check these very carefully.
I've also had disproportionate trouble with 2114 RAM chips.
I have a couple of Compukit UK101s, which use them for 8K
of RAM plus the video memory, and I have quite a few dud
chips now, as well as a tube or two of spares.
--
John Honniball
coredump at gifford.co.uk
16 Jun
16 Jun
12:18 a.m.
You know, I can't rememebr the face of somebody I
met yesterday, but
I rarely forget a scheamtic.
It's a matter of what's important to you. :) I have trouble
remembering names, but I can still recall root passwords from 20+ years
ago. (The first one I ever knew, for example, was "royal" - which I
don't mind saying because that was at least four or five forklift
upgrades, and multiple dozens of root password changes, ago at that
place...which I know because I was there for many of them.)
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
12 Jun
12 Jun
2:54 p.m.
2008/6/12 M H Stein <dm561 at torfree.net>:
Can't suggest where to start checking but I do
have several RO LA100s
awaiting the trip to the scrapper, in case I can help with any non-kbd parts
before they go.
I may take you up on that if repairing the existing bits is going to
be messy. I could definitely use some of the plastic studs that hold
the case together - unfortunately I only figured out that you're only
meant to remove the centre part after persuading the entire thing out
with a screwdriver...
I do also have the tech manual somewhere;
unfortunately IIRC it's on 17"
paper so I can't scan it.
I have since found the full print set on vt100.net, so I'm all set on
that front. I've started tracing back from the logic controlling the
LINE and CTS indicators but haven't found anything obviously wrong as
yet. I have however read some funny-looking voltages off the 7406 in
this part of the circuit: 11.something volts while the datasheet says
30. Could of course be down to my test equipment as I'm limited to a
basic digital multimeter at the moment.
Cheers,
--
Steve Maddison
http://www.cosam.org/
"Hardware: (n.) the parts of a computer you can kick."
midnight
I have since found the full print set on vt100.net, so
I'm all set on
that front. I've started tracing back from the logic controlling the
LINE and CTS indicators but haven't found anything obviously wrong as
yet. I have however read some funny-looking voltages off the 7406 in
this part of the circuit: 11.something volts while the datasheet says
30. Could of course be down to my test equipment as I'm limited to a
basic digital multimeter at the moment.
The 7406 has what's called an 'open collector output'. The output pin is
connected to the colelctor of an NPN transsitor (the emitter being
grounded), but to nothing else inside the chip. What this means is that
the output pin is either low (transistor tuend on) or floating
(transistor turned off), it'll never be pulled high bu that chip.
The 7406 can handle and external voltage oof 30V on the output (that is,
you could connect a resistor between the output pin and any voltage less
than 30V, in shich case that supply voltage essentially becomes the
'high' output voltage). But there's no reason why you _have_ to use 30V.
Any lower voltage is fine. In particualr, using this chip (and its
non-inverting brother, the 7407) to convert between TTL and 12V levels is
quite common, ans what I suspect is happening here.
-tony
5875
days inactive
5879
days old
13 comments
5 participants
participants (5)
-
ard＠p850ug1.demon.co.uk -
coredump＠gifford.co.uk -
dm561＠torfree.net -
mouse＠Rodents-Montreal.ORG -
steve＠cosam.org