Playing with HP2640B

CuriousMarc curiousmarc3 at
Fri Nov 17 01:12:54 CST 2017

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?

On Nov 15, 2017, at 11:48 AM, Mattis Lind via cctalk <cctalk at> 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


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

Has anyone dumped the HP 2640B firmware already? I didn't find it on


More information about the cctalk mailing list