15 Nov
2017
15 Nov
'17
11:48 a.m.
I have been working on a HP 2640B terminal. It was mostly about fixing the
"screen mold" problem and cleaning up the liquids that had been seeping out
from the screen down into the bottom.
The small coaxial wire that connects the 4.9152 MHz clock signal form the
power supply (never seen a crystal controlled SMPSU before!) to the
backplane was broken off, but after fixing that the terminal worked fine.
Just needed some adjustment to the brightness.
With the correct terminfo installed it worked quite well as a serial
terminal to a Linux box.
Then I tried the short 8008 programs that Christian Corti pointed to
http://computermuseum.informatik.uni-stuttgart.de/dev_en/hp2644/diag.html
and
ftp://computermuseum.informatik.uni-stuttgart.de/hp/hp2644
I tried both a couple of times. The terminal enter the LOADER mode but just
hangs completely at the end. I tried different baudrates but no difference.
The selftest STATUS line tell me 40<802 which should indicate that there
are 4k memory in the terminal. However there should be 5k since there is
one 4k board and one combined control store and 1 k RAM board. Maybe there
is a fault in the 1k SRAM? The terminal doesn't complain though.
Regardless, the programs listed either starts at adress 30000 or 36000
which should then be within the available space.
The question is, should these program work for the HP2640B as well? It has
a 8008 but my guess is that the firmware is different from the 2644. What
is the joint experience regarding this? Has anyone ran these small programs
above on a HP2640B?
The HP 2640B firmware consists of four EA 4900 ROM chips which annoyingly
are not anything like normal EPROMs. So dumping will need special
considerations.
Has anyone dumped the HP 2640B firmware already? I didn't find it on
bitsavers.
/Mattis
16 Nov
16 Nov
11:12 p.m.
What did you do for the screen mold? Hot wire method to separate CRT from implosion
window? Put the CRT in a hot water bath? Chip at the glue?
Marc
On Nov 15, 2017, at 11:48 AM, Mattis Lind via cctalk <cctalk at classiccmp.org>
wrote:
I have been working on a HP 2640B terminal. It was mostly about fixing the
"screen mold" problem and cleaning up the liquids that had been seeping out
from the screen down into the bottom.
The small coaxial wire that connects the 4.9152 MHz clock signal form the
power supply (never seen a crystal controlled SMPSU before!) to the
backplane was broken off, but after fixing that the terminal worked fine.
Just needed some adjustment to the brightness.
With the correct terminfo installed it worked quite well as a serial
terminal to a Linux box.
Then I tried the short 8008 programs that Christian Corti pointed to
http://computermuseum.informatik.uni-stuttgart.de/dev_en/hp2644/diag.html
and
ftp://computermuseum.informatik.uni-stuttgart.de/hp/hp2644
I tried both a couple of times. The terminal enter the LOADER mode but just
hangs completely at the end. I tried different baudrates but no difference.
The selftest STATUS line tell me 40<802 which should indicate that there
are 4k memory in the terminal. However there should be 5k since there is
one 4k board and one combined control store and 1 k RAM board. Maybe there
is a fault in the 1k SRAM? The terminal doesn't complain though.
Regardless, the programs listed either starts at adress 30000 or 36000
which should then be within the available space.
The question is, should these program work for the HP2640B as well? It has
a 8008 but my guess is that the firmware is different from the 2644. What
is the joint experience regarding this? Has anyone ran these small programs
above on a HP2640B?
The HP 2640B firmware consists of four EA 4900 ROM chips which annoyingly
are not anything like normal EPROMs. So dumping will need special
considerations.
Has anyone dumped the HP 2640B firmware already? I didn't find it on
bitsavers.
/Mattis
17 Nov
17 Nov
1:09 a.m.
On Thu, 16 Nov 2017, CuriousMarc wrote:
What did you do for the screen mold? Hot wire method
to separate CRT
from implosion window? Put the CRT in a hot water bath? Chip at the
glue? Marc
What we did on one of our 2645 terminals was the hot wire method. We then
attached the "implosion" window to the inner of the case.
BTW is it really an implosion protection? I don't think so because since
the 60s, practically all CRTs have a so-called "integral implosion
protection" (thick glass on the front and metal band around the edge). I
think it is just an anti-glare filter glass. OTOH American CRTs may be
completely different in this aspect compared to European ones.
Christian
1:13 a.m.
Christian do you know the gauge of the wire you used ? And the current?
Maybe I should try that approach again!
David Collins
On 17 Nov 2017, at 8:09 pm, Christian Corti via cctalk
<cctalk at classiccmp.org> wrote:
On Thu, 16 Nov 2017, CuriousMarc wrote:
What did you do for the screen mold? Hot wire method to separate CRT from implosion
window? Put the CRT in a hot water bath? Chip at the glue? Marc
What we did on one of our 2645 terminals was the hot wire method. We then attached the
"implosion" window to the inner of the case.
BTW is it really an implosion protection? I don't think so because since the 60s,
practically all CRTs have a so-called "integral implosion protection" (thick
glass on the front and metal band around the edge). I think it is just an anti-glare
filter glass. OTOH American CRTs may be completely different in this aspect compared to
European ones.
Christian
2:52 a.m.
The screen on my HP2640 had degenerated quite far. It was only a spot in
the middle, 2 by 4 inch, that still attached the glass to the CRT. I used a
thin fish fillet knife to dig through the remaining glue.
Before
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-9/23622163_10155696765784985_65…
The result after:
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-9/23621971_10155696757184985_19…
Anyhow, I researched the glue a bit. The glue is, as far as I understand,
PVAc (PolyVinylAcetate, sometimes also known as PVA). PVAc is not soluble
in water. It takes quite high temperature to melt it. However PVAc is
soluble in many esters. I bought some Butylacetate. It dissolves sample
bits of glue from HP2640 quite well and rapidly. Butylacetate has quite
high boiling temperature (about 120 degrees centigrade) and thus does not
evaporate that quickly. So my idea is now to test on a 2645 screen or VR201
screen by adding some butylacetate and seal with some thin plastic wrap
foil and let it dissolve a bit. Then use the fish fillet knife again and
repeat the process.
/Mattis
4:12 a.m.
I did it by heating the crt to about 50-60 degrees celsius and used a putty-knife.
https://www.flickr.com/photos/hp-fix/albums/72157689357633754
The photos are from a Philips P2000M system but I did it the same way with my 264X
terminals and 9845's systems.
It takes about half an hour to heat and separate the screen from the crt .
-Rik
The screen on my HP2640 had degenerated quite far. It
was only a spot in the
middle, 2 by 4 inch, that still attached the glass to the CRT. I used a thin fish fillet
knife to dig through the remaining glue.
Before
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-
9/23622163_10155696765784985_6518064439030378363_n.jpg?oh=44cbf7f7f
00d6e25155c208124e20a38&oe=5AA7349D
The result after:
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-
9/23621971_10155696757184985_1959733265676657917_n.jpg?oh=36a20689
c0fb5a16de7fc4df138a40e0&oe=5A9993B1
Anyhow, I researched the glue a bit. The glue is, as far as I understand, PVAc
(PolyVinylAcetate, sometimes also known as PVA). PVAc is not soluble in water.
It takes quite high temperature to melt it. However PVAc is soluble in many
esters. I bought some Butylacetate. It dissolves sample bits of glue from HP2640
quite well and rapidly. Butylacetate has quite high boiling temperature (about
120 degrees centigrade) and thus does not evaporate that quickly. So my idea is
now to test on a 2645 screen or VR201 screen by adding some butylacetate and
seal with some thin plastic wrap foil and let it dissolve a bit. Then use the fish
fillet knife again and repeat the process.
/Mattis
10:32 a.m.
Van: Mattis Lind via cctalk<mailto:cctalk at classiccmp.org>
Verzonden: vrijdag 17 november 2017 11:52
Aan: David Collins<mailto:davidkcollins2 at gmail.com>; General Discussion: On-Topic
and Off-Topic Posts<mailto:cctalk at classiccmp.org>
Onderwerp: Re: Playing with HP2640B
The screen on my HP2640 had degenerated quite far. It was only a spot in
the middle, 2 by 4 inch, that still attached the glass to the CRT. I used a
thin fish fillet knife to dig through the remaining glue.
Before
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-9/23622163_10155696765784985_65…
The result after:
https://scontent-arn2-1.xx.fbcdn.net/v/t1.0-9/23621971_10155696757184985_19…
Anyhow, I researched the glue a bit. The glue is, as far as I understand,
PVAc (PolyVinylAcetate, sometimes also known as PVA). PVAc is not soluble
in water. It takes quite high temperature to melt it. However PVAc is
soluble in many esters. I bought some Butylacetate. It dissolves sample
bits of glue from HP2640 quite well and rapidly. Butylacetate has quite
high boiling temperature (about 120 degrees centigrade) and thus does not
evaporate that quickly. So my idea is now to test on a 2645 screen or VR201
screen by adding some butylacetate and seal with some thin plastic wrap
foil and let it dissolve a bit. Then use the fish fillet knife again and
repeat the process.
/Mattis
Wow! Excellent job Mattis <thumb up>
I have seen that HP2640 when it still was in bad shape. The HP2640,
that I have, has many tiny spots on the screen. Your ?result after?
really looks fantastic.
What?s not clear to me, after removing the old ?protection? sheet, did
you apply some new ?protection? sheet, or leave the tube ?as-is? ?
I doubt whether that ?protection? sheet would really help much if
the tube would implode (for whatever miraculous reason) ?
Henk
12:15 p.m.
Wow! Excellent job Mattis <thumb up>
I have seen that HP2640 when it still was in bad shape. The HP2640,
that I have, has many tiny spots on the screen. Your ?result after?
really looks fantastic.
Thanks Henk! I have a few more screens that need fixing. The most difficult
will be the Tek 4016. So I need to do some practice before that one.
What?s not clear to me, after removing the old
?protection? sheet, did
you apply some new ?protection? sheet, or leave the tube ?as-is? ?
I doubt whether that ?protection? sheet would really help much if
the tube would implode (for whatever miraculous reason) ?
No, I didn't add anything between the crt and the glass. If it actually is
a protective shield then of course there is not at all the same protection
afterwards. But as Christian Corti mentioned it might just be an etched
glass to remove reflections. I know at the RICM when doing the same thing
on a DEC VR14 they added a sheet of lexan between the crt and the glass. I
think Michael Thompson can tell more about this.
Henk
/Mattis
5 Dec
5 Dec
7:02 a.m.
I have now dumped the firmware from the HP2640B terminal. Since the EA4900
ROMs are not much like normal EPROMs in regards to pinout and supply
voltages I resorted to use a logic analyzer to dump the contents of the ROM
simply by parsing out the ROM accesses while the terminal did the selftest.
It took a quite many dumps to get the full firmware out of it since my
analyzer HP16550 boards have limited storage. Someone that has a HP16555A
board to sell?
http://storage.datormuseum.se/u/96935524/Datormusuem/HP2640B/2640.hex
One file for all four ROMs.
The good thing attaching the analyzer was that it helped me to find the two
bad 2102 chips on one of the memory boards. Now I get the full 5kKbytes.
While going through the HP2640B I found two boards which I cannot find
relevant schematic and description for, the 02640-60009 Display Memory
Access Module and the 02640-60088 Display Timing Module. Anyone has more
info on these?
With the firmware downloaded I tried to investigate why the small
Diagnostic Download example described here:
http://computermuseum.informatik.uni-stuttgart.de/dev_en/hp2644/diag.html
The example tries to call address 315 octal, but the firmware I dumped has
a HLT instruction at this location which explains why the terminal hangs
completely.
The HP2644 firmware is most likely different from the HP2640B firmware.
http://www.datormuseum.se/peripherals/terminals/hp-264x
7:26 a.m.
On Tue, 5 Dec 2017, Mattis Lind via cctalk wrote:
While going through the HP2640B I found two boards
which I cannot find
relevant schematic and description for, the 02640-60009 Display Memory
Access Module and the 02640-60088 Display Timing Module. Anyone has more
info on these?
The manual for the 02640-60088 is here:
http://www.hpmuseum.net/document.php?hwfile=2542
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
7:35 a.m.
2017-12-05 16:26 GMT+01:00 Mike Loewen via cctalk <cctalk at classiccmp.org>:
On Tue, 5 Dec 2017, Mattis Lind via cctalk wrote:
While going through the HP2640B I found two boards which I cannot find
Thanks! Those numbers fooled me. Some time the number seems to correspond a
bit between the number on the board and some times not.
relevant schematic and description for, the
02640-60009 Display Memory
Access Module and the 02640-60088 Display Timing Module. Anyone has more
info on these?
The manual for the 02640-60088 is here:
http://www.hpmuseum.net/document.php?hwfile=2542
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
7:51 a.m.
On Tue, 5 Dec 2017, Mattis Lind wrote:
2017-12-05 16:26 GMT+01:00 Mike Loewen via cctalk
<cctalk at classiccmp.org>:
The Data Terminal Technical Information Package contains a cross-reference
for PCAs and Module numbers. Unfortunately, it doesn't list the 02640-60009.
http://www.hpmuseum.net/document.php?hwfile=2541
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
On Tue, 5 Dec 2017, Mattis Lind via cctalk
wrote:
While going through the HP2640B I found two boards which I cannot find
Thanks! Those numbers fooled me. Some time the number seems to correspond a
bit between the number on the board and some times not. relevant schematic and description for, the
02640-60009 Display Memory
Access Module and the 02640-60088 Display Timing Module. Anyone has more
info on these?
The manual for the 02640-60088 is here:
http://www.hpmuseum.net/document.php?hwfile=2542
9:14 a.m.
2017-12-05 16:51 GMT+01:00 Mike Loewen via cctalk <cctalk at classiccmp.org>:
Thanks! Those numbers fooled me. Some time the number seems to correspond
a
bit between the number on the board and some times not.
The Data Terminal Technical Information Package contains a
cross-reference for PCAs and Module numbers. Unfortunately, it doesn't
list the 02640-60009.
http://www.hpmuseum.net/document.php?hwfile=2541
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
On Tue, 5 Dec 2017, Mattis Lind wrote:
2017-12-05 16:26 GMT+01:00 Mike Loewen via cctalk <cctalk at classiccmp.org>:
>
> On Tue, 5 Dec 2017, Mattis Lind via cctalk wrote:
>>
>> While going through the HP2640B I found two boards which I cannot find
>>
>>> relevant schematic and description for, the 02640-60009 Display Memory
>>> Access Module and the 02640-60088 Display Timing Module. Anyone has more
>>> info on these?
>>>
>>>
>> The manual for the 02640-60088 is here:
>>
>> http://www.hpmuseum.net/document.php?hwfile=2542
>
>
Looking more into this document revelas that it covers three boards
02640-60009, 02640-60088 and 02640-60112.
Now I had the idea of dumping the character ROM as well. But what is a AMD
27S82? I cannot really find a datasheet online. From the schematics above
it looks not too much off from standard 2716 EPROM with pin 19 (A10) as E3
, pin 21 (VPP) as /E1. But my Data I/O 29B does not seem to be compatible
with 27S82 and as pin 21 is active low it gives that a 2716 setting in the
programmer (reader) won't work.
Does the 27S82 have any equivalents that I could look up instead in the
device list?
10:18 a.m.
On Tue, 5 Dec 2017, Mattis Lind wrote:
2017-12-05 16:51 GMT+01:00 Mike Loewen via cctalk
<cctalk at classiccmp.org>:
I couldn't find that part number either, but I dumped the PROMs on my 2647A
Control Memory PCAs, which use an AMD AM9216BDC PROM. It appears to be set up
the same way as you describe. I hand-wired an adapter to plug into my 29B,
and told it that I was reading a 2716. The datasheet for the AM9216BDC is in
the AMD MOS/LSI Data Book:
http://bitsavers.informatik.uni-stuttgart.de/components/amd/_dataBooks/1980…
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
On Tue, 5 Dec 2017, Mattis Lind wrote:
2017-12-05 16:26 GMT+01:00 Mike Loewen via cctalk <cctalk at classiccmp.org>:
>
> On Tue, 5 Dec 2017, Mattis Lind via cctalk wrote:
>>
>> While going through the HP2640B I found two boards which I cannot find
>>
>>> relevant schematic and description for, the 02640-60009 Display Memory
>>> Access Module and the 02640-60088 Display Timing Module. Anyone has more
>>> info on these?
>>>
>>>
>> The manual for the 02640-60088 is here:
>>
>> http://www.hpmuseum.net/document.php?hwfile=2542
>
>
Looking more into this document revelas that it covers three boards
02640-60009, 02640-60088 and 02640-60112.
Now I had the idea of dumping the character ROM as well. But what is a AMD
27S82? I cannot really find a datasheet online. From the schematics above
it looks not too much off from standard 2716 EPROM with pin 19 (A10) as E3
, pin 21 (VPP) as /E1. But my Data I/O 29B does not seem to be compatible
with 27S82 and as pin 21 is active low it gives that a 2716 setting in the
programmer (reader) won't work.
Does the 27S82 have any equivalents that I could look up instead in the
device list? 
11:44 a.m.
On 2017-Dec-05, at 9:14 AM, Mattis Lind via cctalk wrote:
Looking more into this document revelas that it covers three boards
02640-60009, 02640-60088 and 02640-60112.
Now I had the idea of dumping the character ROM as well. But what is a AMD
27S82? I cannot really find a datasheet online. From the schematics above
it looks not too much off from standard 2716 EPROM with pin 19 (A10) as E3
, pin 21 (VPP) as /E1. But my Data I/O 29B does not seem to be compatible
with 27S82 and as pin 21 is active low it gives that a 2716 setting in the
programmer (reader) won't work.
Does the 27S82 have any equivalents that I could look up instead in the
device list?
FWIW, 27S82 is listed in the 1982 IC Master as
AM27S82M: TTL 1024*8 ROM, 275nS access, OC outputs, 24 pin, 5V supply
So half the size of a 2716.
The book lists PROMs and ROMs separately, so it was probably mask-programmed.
(M would be for mil temp going by other AMD numbers.)
1:08 p.m.
2017-12-05 20:44 GMT+01:00 Brent Hilpert via cctalk <cctalk at classiccmp.org>:
On 2017-Dec-05, at 9:14 AM, Mattis Lind via cctalk
wrote:
I would think bipolar PROM since rest of the AMD 27S series were bipolar
PROMs. I will check this weekend if I can find an old AMD databook. I do
have the "Schottky and Low-Power Schottky Bipolar memory, Logic and
Interface" somewhere.
https://drive.google.com/file/d/0B-rp4vyPPYu1Z3d6MkxRQ1NlVXc/view?usp=shari…
I traced the card to understand if the board actually matched the
schematics and it did. Pin 18 was NC and the it was possible to read the
chip with my trusty Data IO 29B as a 2732 and just use the first 1024 bytes
out of it.
The resulting files are here for those who like to preserve them:
http://storage.datormuseum.se/u/96935524/Datormusuem/HP2640B/hp2640b-charge…
http://storage.datormuseum.se/u/96935524/Datormusuem/HP2640B/hp2640b-charge…
Looking more into this document revelas that it covers three boards
02640-60009, 02640-60088 and 02640-60112.
Now I had the idea of dumping the character ROM as well. But what is a
AMD
27S82? I cannot really find a datasheet online.
From the schematics above
it looks not too much off from standard 2716 EPROM with pin 19 (A10) as
E3
, pin 21 (VPP) as /E1. But my Data I/O 29B does
not seem to be
compatible
with 27S82 and as pin 21 is active low it gives
that a 2716 setting in
the
programmer (reader) won't work.
Does the 27S82 have any equivalents that I could look up instead in the
device list?
FWIW, 27S82 is listed in the 1982 IC Master as
AM27S82M: TTL 1024*8 ROM, 275nS access, OC outputs, 24 pin, 5V
supply
So half the size of a 2716.
The book lists PROMs and ROMs separately, so it was probably
mask-programmed.
(M would be for mil temp going by other AMD numbers.)
1:20 p.m.
On 2017-Dec-05, at 1:08 PM, Mattis Lind wrote:
2017-12-05 20:44 GMT+01:00 Brent Hilpert via cctalk
<cctalk at classiccmp.org>:
The ref did specify it as TTL, so yes, bipolar.
On 2017-Dec-05, at 9:14 AM, Mattis Lind via cctalk
wrote:
> Does the 27S82 have any equivalents that I
could look up instead in the
> device list?
FWIW, 27S82 is listed in the 1982 IC Master as
AM27S82M: TTL 1024*8 ROM, 275nS access,
OC outputs, 24 pin, 5V supply
So half the size of a 2716.
The book lists PROMs and ROMs separately, so it was probably mask-programmed.
(M would be for mil temp going by other AMD numbers.)
I would think bipolar PROM since rest of the AMD 27S series were bipolar PROMs. I will
check this weekend if I can find an old AMD databook.
1:36 p.m.
On 2017-Dec-05, at 1:20 PM, Brent Hilpert via cctalk wrote:
On 2017-Dec-05, at 1:08 PM, Mattis Lind wrote:
Oh, as for being PROM vs ROM, I'm just going by the way things are listed in the ref.
It lists different 27Sxx devices under PROM vs ROM,
for example 27S180, 27S181 are 1024*8 listed under PROM, but 27S80, 27S81, 27S82, 27S83
are 1024*8 listed under ROM.
It's possible they're mis-listed of course, but manufacturers often did make
equivalent devices as both fusible and mask,
I don't know what AMD did for series names in that regard.
2017-12-05 20:44 GMT+01:00 Brent Hilpert via
cctalk <cctalk at classiccmp.org>:
The ref did specify it as TTL, so yes, bipolar. On 2017-Dec-05, at 9:14 AM, Mattis Lind via
cctalk wrote:
> Does the 27S82 have any equivalents that I
could look up instead in the
> device list?
FWIW, 27S82 is listed in the 1982 IC Master as
AM27S82M: TTL 1024*8 ROM, 275nS access, OC
outputs, 24 pin, 5V supply
So half the size of a 2716.
The book lists PROMs and ROMs separately, so it was probably mask-programmed.
(M would be for mil temp going by other AMD numbers.)
I would think bipolar PROM since rest of the AMD 27S series were bipolar PROMs. I will
check this weekend if I can find an old AMD databook.
6 Dec
6 Dec
12:28 a.m.
We just dumped the enhanced character ROMs from a HP2645. 8k bipolar mask ROMs, 90 ns
(they need to be fast since they are used as video RAM really). I read them on the Data IO
using their PROM equivalent, AMD 27S181. The microvectors sets (line set and large
character sets) are 9 bits. The 9th bit comes out on pin 18 in place of an enable pin in
the more conventional 8-bit version. I read them in two passes, wiring the pin for bit 9
in place of bit 8 pin the second time around. 9-bit mask ROMs! Why not 5-legged pigs while
you are at it. Then we found that even HP had misgivings about their weird 9-bit ROM
choices, must have cost a fortune. Carl had a later rev board that had 8-bit version of
the microvector sets, where the board has electronics to create an emulated 9th bit by
shifting all the bits out and duplicating bit 0. Better yet, we burned some new 8-bit
bipolar PROMs (Harris 7681), emulated the bit shifting on my older rev of the board by
cross-wiring, and added new large character sets to my terminal! That was fun.
Marc
On Dec 5, 2017, at 1:36 PM, Brent Hilpert via cctalk <cctalk at classiccmp.org>
wrote:
On 2017-Dec-05, at 1:20 PM, Brent Hilpert via cctalk wrote:
Oh, as for being PROM vs ROM, I'm just going by the way things are listed in the ref.
It lists different 27Sxx devices under PROM vs ROM,
for example 27S180, 27S181 are 1024*8 listed under PROM, but 27S80, 27S81, 27S82, 27S83
are 1024*8 listed under ROM.
It's possible they're mis-listed of course, but manufacturers often did make
equivalent devices as both fusible and mask,
I don't know what AMD did for series names in that regard.
On
2017-Dec-05, at 1:08 PM, Mattis Lind wrote:
2017-12-05 20:44 GMT+01:00 Brent Hilpert via cctalk <cctalk at classiccmp.org>:
The ref did specify it as TTL, so yes, bipolar. On 2017-Dec-05, at 9:14 AM, Mattis Lind via
cctalk wrote:
> Does the 27S82 have any equivalents that I
could look up instead in the
> device list?
FWIW, 27S82 is listed in the 1982 IC Master as
AM27S82M: TTL 1024*8 ROM, 275nS access, OC
outputs, 24 pin, 5V supply
So half the size of a 2716.
The book lists PROMs and ROMs separately, so it was probably mask-programmed.
(M would be for mil temp going by other AMD numbers.)
I would think bipolar PROM since rest of the AMD 27S series were bipolar PROMs. I will
check this weekend if I can find an old AMD databook.
1:10 a.m.
Marc, in addition to Mattis? forthcoming reply, my recent experience with a moldy 2624A
was that the hot wire method was very poor. Too hard to get the wire in, didn?t melt the
?glue? very well, smelly. Gave up when the wire broke.
What worked best for me was a flat blade screwdriver that was small enough to sit sideways
in the gap between the front glass and the tube. I sliced sections of the glue and picked
them out with a hook. I also squirted in a combination of RP7 and household cleaner but
not sure either did anything other than lubricate the surfaces - they may have helped lift
the glue a bit.
My ?glue? was like a layer of silicon rubber which hung on for as long as possible but I
got it all off without any damage.
I replaced the front glass and held it on with a bead of black silicon rubber used for
shower glass. I spaced it from the tube with pieces of wire around the edges and pulled
them out when the silicon dried.
Worked well for me but keen to hear how Mattis went.
I didn?t try the hot water soak but it would probably help.
David Collins
On 17 Nov 2017, at 6:12 pm, CuriousMarc via cctalk
<cctalk at classiccmp.org> wrote:
What did you do for the screen mold? Hot wire method to separate CRT from implosion
window? Put the CRT in a hot water bath? Chip at the glue?
Marc
On Nov 15, 2017, at 11:48 AM, Mattis Lind via cctalk <cctalk at classiccmp.org>
wrote:
I have been working on a HP 2640B terminal. It was mostly about fixing the
"screen mold" problem and cleaning up the liquids that had been seeping out
from the screen down into the bottom.
The small coaxial wire that connects the 4.9152 MHz clock signal form the
power supply (never seen a crystal controlled SMPSU before!) to the
backplane was broken off, but after fixing that the terminal worked fine.
Just needed some adjustment to the brightness.
With the correct terminfo installed it worked quite well as a serial
terminal to a Linux box.
Then I tried the short 8008 programs that Christian Corti pointed to
http://computermuseum.informatik.uni-stuttgart.de/dev_en/hp2644/diag.html
and
ftp://computermuseum.informatik.uni-stuttgart.de/hp/hp2644
I tried both a couple of times. The terminal enter the LOADER mode but just
hangs completely at the end. I tried different baudrates but no difference.
The selftest STATUS line tell me 40<802 which should indicate that there
are 4k memory in the terminal. However there should be 5k since there is
one 4k board and one combined control store and 1 k RAM board. Maybe there
is a fault in the 1k SRAM? The terminal doesn't complain though.
Regardless, the programs listed either starts at adress 30000 or 36000
which should then be within the available space.
The question is, should these program work for the HP2640B as well? It has
a 8008 but my guess is that the firmware is different from the 2644. What
is the joint experience regarding this? Has anyone ran these small programs
above on a HP2640B?
The HP 2640B firmware consists of four EA 4900 ROM chips which annoyingly
are not anything like normal EPROMs. So dumping will need special
considerations.
Has anyone dumped the HP 2640B firmware already? I didn't find it on
bitsavers.
/Mattis
2572
days inactive
2593
days old
20 comments
8 participants
participants (8)
-
bhilpert＠shaw.ca -
cc＠informatik.uni-stuttgart.de -
curiousmarc3＠gmail.com -
davidkcollins2＠gmail.com -
henk.gooijen＠hotmail.com -
hp-fix＠xs4all.nl -
mattislind＠gmail.com -
mloewen＠cpumagic.scol.pa.us