You Are Here: Greyghost > Features > Twinhead Technology TH6887A External Battery Rework
Twinhead Technology TH6887A
External Battery Rework
Recently, I had
the occasion to collect several integrated clock modules as part of a much
larger load of IBM PS/2 equipment. And as I've written before, integrated clock
modules are great...right until the battery runs down. The modules I collected
bore date codes from 1993 and 1994. Given that they had been stored in a
building with no climate control, the modules had undoubtedly been exposed to
extreme heat, cold and everything in between. The batteries inside them would
surely be depleted.
(It is interesting to note that a 1993 era
Benchmarq BQ3287 in this bunch of modules appeared to have a good battery, where
the TH6887A modules had depleted their batteries. I don't know why this is. It
is hard to imagine that the BQ3287 module was never used and had never blown its
DS1287/12887 and variants were common and popular modules for which replacements
can still be had if you can justify spending the money on them. There are also
many clones of the DS1287 and DS12887 parts out there. I've seen such modules
branded Symphony, Houston Tech, VIA, Benchmarq (now Texas Instruments), along with
various other fairly generic markings--with no brand name--that resemble part
numbers. I've also seen a few modules that had the names of motherboard makers
printed on them. Some are compatible with the DS1287/12887 and others are
This page discusses one of the clones I ran across.
Someone modified this module before I did. However, their web page describing
the process is long gone and the Internet Archive's Wayback Machine did not
archive it. That is why I created this page, to document the process of
modifying a TH6887A real time/NVRAM module and compare it to the modification of
a Dallas Semiconductor part.
Many of these are
nothing more than clones of the DS1287/12887 modules. Some are possibly modules that were made by Dallas
Semiconductor and labeled as requested by the company using them in finished
products. Twinhead Technology's TH6887A
module is an example of such a clone, and it uses slightly different
construction as compared to a Dallas Semiconductor part. I don't know if
Twinhead ever had any ability to fabricate their own parts, but I've seen some
other ICs bearing their name, typically on CGA video cards.
Whatever the case with its manufacture, the only
important thing is that the TH6887A module can be re-enlivened just as the
Dallas Semiconductor DS1x87/1x887 parts can be. I used much the same procedure
as explained by Peter Wendt here, with the notable exception of the Dremel tool.
I've never had any luck opening these modules with a Dremel tool--instead, I'd
end up destroying them or flinging them off to points unknown if they escaped
the grip of the vice.
A key difference
with the TH6887A part is the location of the wiring leading to the internal
battery. They are in the same general location as the ones in a Dallas module,
but the leads themselves are much closer to the surface of the potting compound.
It is also possible to see the other pins of the embedded IC once you've broken
through the rigid top layer (or "shell") on the module. You won't have to dig as
far into the module to get at the battery leads. Once you have opened the
module, disconnecting the battery leads is the same process as it is on the
Dallas module. Despite some exterior differences in the TH6887A's design, it
appears that the battery circuit was routed much the same way as it would be on
a DS1287/12887 part.
I don't know if the
TH6887A has a batteryless counterpart. I guess it would surprise me a little if
it did, as I don't know if Twinhead Technology had any ability to fabricate
their own semiconductors or any need for a part that gets its power from an
external battery in their motherboard products. And I don't feel like destroying
a module simply to find out whose IC it is based on.
It can be safely said that the TH6887A is 100%
compatible with the DS1287/12887 components. After re-energizing one, I put it
in a PS/2 Model 55SX and was able to configure the system successfully. And, as
with the Dallas modules, I had to run a DEBUG script to trash the contents of
the CMOS RAM before the configuration would work correctly and allow the system
to boot from its internal hard disk.
Go Back >
2009 William R.
Walsh. All Rights Reserved. Written 11/19/2009. Permission is
granted to mirror this page
in its unedited entirety as long as a link back to this site and credit
for the material is provided. You may not charge a fee or exchange
items of value to provide access to this page or its content, other
than an amount reasonably necessary to cover the cost of connection
time, data transfer or printing supplies. Content from this page may
not be displayed alongside advertising content of any type. You may use
portions of this page in other products only if you provide credit and
a link back to this page and only if the finished product is freely
accessible to anyone interested in having a copy. If you use this
material in work of your own, you may not charge or exchange items of
value to provide access (other than as reasonably necessary to cover
connection time, data transfer fees, or to cover printing supply costs)
nor may you display advertising materials alongside any content you use
from this page. Images may not be edited other than to resize them or
to provide for faster downloading.