ANTHISTLE SYSTEMS & PROGRAMMING LTD.
563 Patricia Drive,
Oakville, Ontario,
CANADA L6K 1M4

Telephone 905-845-7959

19 February, 1999

For those who are concerned about the Y2K problem as it relates to 
computer CMOS clocks and BIOS here are some web pages copied from the 
internet. This will give you additional information.  If you have web 
access you can view the originals and follow additional links at the 
http:// addresses given.  For those who do not have web access I have 
converted the web pages to text so you can read them.

------------------------------------------------------------
From http://www.electrocution.com/2000.htm
------------------------------------------------------------

Just to recap (in case you've been asleep), the problem started when 
early programmers and equipment makers used 2-digit date fields to 
represent the century, because resources were scarce and expensive - 
even 
10 years ago, the hard disk space to store an extra two digits for the 
date fields for 40,000 records would have cost over $2,000. Even earlier 
than that, when using punched cards, it was important not to spill over 
to a second one, so you didn't double processing time and have to spend 
money on a second shift of operators. Separators, 
such as forward slashes or hyphens to distinguish between years, months 
and days were discarded as well, so 01/01/1980 would look like 010180. 
On that basis, at the end of 1999, 311299 would evolve into 010100, 
which 
brings us to the way computers (or rather, programs) calculate on the 
dates concerned.

Without the century figures to work with, the second date above (that 
is, for 
the first day of the year 2000) looks to the computer like a lesser 
number, 
or an earlier date, so 1965 works out (in computer speak) to be earlier 
than 
in 1898, making the software used by your local coroner think you died 
before 
you were born. That might not bother you particularly, but it could if 
your 
mortgage payments are calculated wrongly, or your endowment policy 
doesn't 
get paid when it matures. If you're running a business, watch out for 
those 
automatic invoicing routines (and the cashflow, or lack of it), or 
systems 
that Personal Information Managers use to delete past events (you didn't 
want that lunch anyway, did you?). On top of all that, the year 2000 
must also
be recognised as a leap year.

However, companies tend not to spend money on upgrades and often simply 
make 
do, usually only thinking as far ahead as the next Board Meeting, so 
software 
that was expected to have a working life of around 5 years is still 
being used 
today. Unfortunately, 1996 was the last point something could be done by 
average programmers in a mid-size company without some sort of 
automation, 
and have a reasonable chance of finishing. Since the programming 
problems 
revolve around COBOL-based systems that weren't used on PCs (COBOL 
stands for 
Common Business Oriented Language, and revolves around accountancy), 
this 
leaves some of you with very much less to worry about, as much of all 
this 
is a management problem, concerning the sheer number of changes to be 
made inside
a certain time rather than the specific actions to be taken. Don't stop 
reading, though, because you may still have concerns about hardware.

In a PC, the problem boils down to the date in the Real Time Clock not 
rolling 
over automatically, and not necessarily whether a date over 2000 can 
actually be 
coped with, because BIOS manufacturers appear to have solved that 
already and 
DOS has been able to cope with it for years (but see below about Time 
Dilation). 
If you set your system time and date to just before midnight 1999 and 
leave the 
machine running until the rollover happens, you will find that DOS 
(version 6.22, 
anyway) copes with the problem quite easily (boot the machine from a 
floppy, 
to make sure that date dependent software is not affected). Although it 
only 
shows two digits when you perform operations such as the dir command, 
four a
re used internally. As its official operating span is between 1980 and 
2099, it 
can figure out that 00 equates to 2000, although it may cough and 
splutter a 
bit if the Real Time Clock specifically hands it a date of 1900, or any 
other 
date it can't cope with.

In practice, the BIOS converts it as well, so this shouldn't happen; 
some correct 
the time automatically at boot and others will supply DOS with 2000 
instead of a 
hardware date of 1900, although many can't produce a date later than 
1999 anyway - 
Award BIOS 4.5G prior to Nov 1995 can only accept dates between 1994-
1999. Still 
others just add 20 years before passing the date to the operating 
system.

The official story from Microsoft about DOS (6.22) is that it will allow 
a date of up to 2099 to be input, but the operational range is actually 
up to 2035. 1980 is regarded as year 0. msbackup has other problems, 
but noone uses that anyway. Windows 3.x depends on DOS, so it will react 
similarly, with the addition that you have to set the leap year from the 
keyboard. The 2 main problems described here, such as the RTC not 
updating 
the century byte, and Time Dilation (see below) are both affected by the 
complex interaction between the POST and the hardware.

For example, changing 99 to 00 takes the PC outside its operating 
envelope, so it 
returns a date of 01/04/1980, which is actually an error message 
indicating an 
out-of-range date, instead of the 01/01/1900 it gets from the Real Time 
Clock (RTC), 
which is inside its own specification. Also, changing the century byte 
from 19 to 20 
causes the POST to use a different logic path and exceed the 244 
microsecond grace 
period so that, occasionally, if the RTC is being read at just the wrong 
time 
(after Y2K), Time Dilation occurs. Incidentally, the date 01/01/1980 is 
set if 
your CMOS contents are lost, and 01/03/1980 means an invalid BCD, as 
mentioned 
below. And the 244 microsecond grace period?

Well, you can't read the RTC while it's updating, and the status is not 
checked after 
every read, so the designers set the status early and guaranteed the 
data for 244 
microseconds, so it could be obtained even if it changed afterwards. 
Hence the problem 
if the POST takes a different logic path (which is because the RTC and 
PC keep the date 
in different formats and therefore need more conversion time if you now 
have to adjust 
for the new century and count the years since rollover on top of the 
normal 
calculations - more details below). Timing is important, too - if you 
read the RTC 
at the beginning of a second, you have almost the whole second minus 
about 600 
microseconds (for updating) to do it in, but if you do it at the start 
of the grace 
period, you only have the 244 microseconds.

The Real Time Clock, or RTC, lives in a chip on the motherboard that 
also has 
some memory in it, which is used by the computer to remember its 
configuration 
from day to day, including the time and date. When the computer is 
switched off, 
there is a battery that keeps that information alive, and which usually 
lasts 
about 2 years. The clock was introduced with the IBM AT, and it's ironic 
that 
anyone out there still using an XT-class machine (and there are many) 
won't have 
this problem, because it didn't have one. Instead, it used the DOS 
clock, which 
is set in motion every time the computer is switched on (it's actually a 
device 
driver called clock$, which is really a counter that increments at a 
known rate 
of about 18 times a second). It was never intended to be a precision 
time-keeping 
device, but for "operator convenience". There were certainly no 
specifications 
written down.

The DOS clock gets its starting information from the RTC, or from you 
typing it in, 
as was done with the XT. The RTC is then promptly ignored until the 
machine is switched 
on again. The date in the RTC is kept in second, minute, hour, day, 
month and year 
format, but is converted to the number of days since Jan 1 1980, plus 
the number 
of seconds since midnight on the current day. The latter is stored in 
the counter 
by the BIOS, and when DOS needs to read the clock, the BIOS is called to 
read the 
counter, and the number of ticks is converted back to seconds. This is 
to save 
going as far as the RTC for the time, which is a slower process, and 
explains 
why there is often a difference between your watch and your PC's clock 
at the end 
of the working day; the system clock has to compete for attention with 
other 
devices, and is often reprogrammed by games and other programs, that use 
it for 
their own timing purposes, mostly running the video faster. If they 
don't set it 
back, your DOS clock will be running very fast indeed! In short, being 
interrupt 
driven, the DOS clock's accuracy depends on system activity.

In all IBM PC/AT compatible systems, the current date and time are 
stored in internal 
CMOS memory in Binary Coded Decimal (BCD) format, with each component 
occupying one byte, 
such as second, minute, hour, day of week, month and year (the year uses 
up two bytes, 
one for the lower two digits, and one for the upper two). In case you 
were wondering 
what BCD means, I'm indebted to Peter Farrow for the following...

In BCD each decimal digit is mapped to a 4-bit binary number. BCD is 
simply a binary sequence that terminates at the equivalent of decimal 
"9", so a 
sequence for 4 bits is:

0000	0
0001	1
0010	2
0011	3
0100	4
0101	5
0110	6
0111	7
1000	8
1001	9

after which it returns to 0000, so any value above "9" is not valid. 
When the BIOS 
complains about an invalid BCD code, one of the 4-bit BCD numbers is 
carrying over 
to a value that would correspond to a decimal digit value of more than 
"9" (1001) 
which is, of course, invalid.

And to Mike Echlin for this:
It's just a decimal number stored in hex. i.e. BCD 12 is hex, but means 
decimal 12, 
and 0a in BCD is an error, but valid, that is, it's a valid hex number, 
but when 
converted to decimal is an error (Confused yet? ;-). In other words, a 
BCD number 
is just a hex number that when printed as hex, looks like decimal, and 
is read as 
if it was decimal. So it is stored as 0x12,printed in c as:

x = 0x12;
printf( "%x", x);

which looks like "12", and is interpreted as 12 decimal.
In summary, all BIOSes have the millenium bit problem, because it is a 
hangover 
from the IBM architecture, but Award Software, Phoenix and AMI appear to 
have sorted 
it out in all versions after 1994. However, whether the motherboard 
designer has is 
another story! Here is the official chart of BIOS dates and the action 
you need to 
take, but after you've taken a quick look we will discuss what might 
happen in 
reality - we've already seen that some Award BIOSes after Nov 95, sorry, 
1995, 
can only use a certain date range.



BIOS date

Action


Before 26 April 1994

Reset your system clock once, by turning the system off before midnight 
on 31 December 1999, and turning it on again after midnight (that is, on 
1 Jan 2000), then reset the date manually in setup.


26 April 1994 - 31 May 1995

You either need an update or must be willing to reset the clock every 
day.


After 31 May 1995

Don't worry about it.




As of October 1997, all Compaqs are compatible - pre Y2K-compliant 
Presarios 
will read 1900 after the rollover. You can check out your machine and 
download a patch (if required) from Compaq's web site. For 
IBM products, you may just have to 
re-enter the date. All Toshiba computers manufactured after April 4, 
1996 
have Year 2000 compliance code written into the BIOS (version 5.0 or 
higher). 
All machines produced after 1992 will accept the date manually. 
Otherwise, 
Toshiba and its software partners have incorporated 4-digit dates into 
their 
products.

Mac users only have to worry about software, because the hardware and 
the way it handles dates is completely different from the PC. Unix keeps 
track of time in numbers of seconds since 1970. This method will track 
time until the year 2038 when the date function will then require more 
than 32 bits.

As far as software goes, Windows 3.x has the millennium bug. The most 
inexpensive solution is to download the file year2000.zip from 
Windows Magazine and 
follow the instructions in the readme file.

To determine if a particular Microsoft application has a date-conversion 
limitation, visit the Microsoft web site. Microsoft Access 95 has a date 
limit of 1999, for which you need an upgrade. You can create a macro in 
Excel to change 20th century dates to 21st century ones.
Time Dilation
This is also known as the Crouch, Echlin effect, referring to time and 
date instabilities that will occur in and beyond 2000 on PCs and other 
embedded 
hardware. In other words, the time and date have been observed to 
intermittently 
and abruptly leap forward (or backward) when PCs are powered down and/or 
rebooted
 after 2000. In extreme cases, COM ports have ceased to work. For 
example, one PC 
produced a date in December after 2 weeks into 2000.

More information can be obtained from 
Mike Echlin's web page 
or Jace Crouch's.

Testing The Hardware
All this really takes is common sense. Although there is lots of 
software out 
there, they all essentially use the routine that follows. It would 
appear that 
tests for the leap year aren't necessary, but you may want to check 
anyway.

Again, making sure you boot the machine from a floppy, to protect any 
date 
dependent software you may have, set the time and date to a few minutes 
before midnight on 31 December 1999, and wait. As we have already seen, 
DOS will probably cope, but the acid test comes when you switch the 
machine 
off, then on again (if DOS can't handle it, change to a later version 
and 
try again to eliminate it from the equation). You might well see a 
system 
date starting somewhere in 1980, which is usually what you get if the 
machine 
has no other information, but if you see the correct figures, your 
hardware 
needs no further treatment, save to check for February 29th (2000 is a 
leap 
year). You won't lose your hard drive details or other BIOS settings, 
which 
is what usually happens when you lose the CMOS. Repeat to make sure, but 
you might 
want to keep the machine running a while longer to check for Time 
Dilation (see above). 
Incidentally, the date 01/01/1980 is usually set if your CMOS contents 
are 
lost, and 01/04/1980 if an out-of-range date is encountered. 01/03/1980 
means 
an invalid BCD, as mentioned above. There is a utility called viewcmos 
which 
will let you inspect the RTC directly and see what's really going on.

If your machine didn't pass the rollover test above, you might be able 
to 
reset it manually, in which case, set the time a few minutes after 
midnight, 
and switch off again (very important, that). If the date is still 
correct when you switch on, your hardware is useable. Otherwise, you 
will 
have to change the motherboard. You could try changing the BIOS, but 
that's 
often not cost-effective. If you feel the need to try, there are 
several links to BIOS and motherboard manufacturers 
who may have one you can use.

Replacement CMOS RAM/RTC chips can be obtained from Dallas Semiconductor 
for many machines.

There appears to be some memory-resident software available that takes 
up only 256K of memory. Presumably it traps any calls for the date and 
feeds back the correct one. However, 
once Windows '95 starts rearranging memory to its own liking, as it is 
wont to do, this might well 
become ineffective. A better solution is to use software that acts as a 
device driver, and which 
can therefore control interrupts.

HTML RETAINED IN REMAINDER OF DOCUMENT FOR LINKS

Also, there are plug-in cards (e.g. from <A 
HREF="http://www.unicore.com">Unicore</A> or <A 
HREF="http://www.eurosoft-uk.com">Eurosoft</A>) that occupy an ISA slot.
<P>
For more information, here's a list of <a 
href="http://www.liant.com/whatsnew/lnk2000.htm">Year 2000 web 
sites</A>, or Peter de Jager's <a href="http://www.year2000.com">The 
Year 2000 Information Center</A>. 
Try also <A 
HREF="http://www.microsoft.com/cio/articles/year2000faq.htm">Microsoft</
A> for a series of FAQs, or these websites:
<MENU>
<LI><A HREF="http://www.rightime.com/">www.rightime.com</A>
<LI><A 
HREF="http://www.sbhs.com/y2k/index.htm">www.sbhs.com/y2k/index.htm</A>
<LI><A HREF="http://www.rigel.co.nz/">www.rigel.co.nz</A>
<LI><A HREF="http://www.everything/">www.everything2000.com</A>


------------------------------------------------------------
END OF http://www.electrocution.com/2000.htm
------------------------------------------------------------


------------------------------------------------------------
From http://www.firmware.com/pb4ts/year2000.htm
------------------------------------------------------------


The Y2K Date Rollover Problem on PC Systems

Filename: YEAR2000.TXT
WWW URL:  http://www.firmware.com/support/y2k/year2000.htm
FTP URL:  ftp://ftp.firmware.com/y2k/year2000.txt
Author:   Terry Slade, Micro Firmware Technical Support
Revision: 02/15/99
Summary:  Some information on the Y2K date rollover problem on PCs and
          how it relates to the PC's system ROM BIOS and RTC.


Contents:

1. Introduction
2. A Technical Explanation of the Y2K Date Rollover Problem on PCs
3. Y2K Compliancy of Micro Firmware BIOS Upgrade Products
4. Solutions to the Y2K Date Rollover Problem on PCs
5. Testing PCs for Y2K Date Rollover Problem
6. Does a Y2K-compliant BIOS Make a PC "Year 2000 Compliant"?
7. Time Dilation / Crouch/Echilin Effect / Time-leap
8. For more information:


1. Introduction

There are some interesting and important issues concerning computers
and the year 2000. Collectively, these issues have come to be known by
various names such as "the year 2000 problem" or "the millenium bug"
or just "Y2K". Basically this has to do with how various computer
software and hardware will deal with the date change from DEC 31, 1999
to JAN 1, 2000 and with how software will handle dates beyond
DEC 31,1999.

This problem seems to be of more concern on mainframe systems than on
PC systems and is also more of a software issue than a hardware issue.
Another issue is embedded systems - the billions of chips that are
built-in to various machines. There is, however, a hardware issue that
affects almost all PCs. On PC systems, the Real Time Clock (RTC) chip
will not roll the date over to the year 2000 without some help. Many
newer BIOSes will account for this.

We began to investigate this issue in 1996 as we were starting to get
phone calls from customers and non-customers wondering if our BIOS
upgrades have this problem or solve this problem. It is the purpose of
this article to address this issue as it pertains to PC hardware, the
PC BIOS, and our own BIOS upgrades in some detail. Please note that
there are additional date-related issues concerning PC operating
system and application program software. These are not issues that we
will investigate or address. We would suggest looking at the web sites
listed at the end of this document as a starting place to begin
learning about any year 2000 issues not addressed here.

Also - see our web site for information on our Flash 2000 card, which
is an 8-bit ISA card with a ROM extension that corrects for the Y2K 
date rollover problem. A separate text file, FLSH2000.TXT, discusses
that card in detail. That file is available on our web site and ftp
site at these URLs:

http://www.firmware.com/support/y2k/flsh2000.htm

ftp://ftp.firmware.com/y2k/flsh2000.txt

The specific problem that occurs on PCs is that, under most older
BIOSes, if the system is powered off during the date change from
DEC 31, 1999 to JAN 1, 2000, when the system is turned back on the
DOS date will be indicated as 01/04/1980. This is easily solved by
just setting the date correctly with the DOS DATE command. The system
should keep the date correctly from then on. If the system is running
at midnight on 12/31/99, it may roll over the date correctly, but
still show a date of 01/04/80 when first rebooted. After setting the
date correctly it should keep the date properly from then on. On
some systems, other solutions may be needed, since it may be critical
for the date change to be handled automatically, without having to
depend on someone remembering to set the date.


2. A Technical Explanation of the Year 2000 Problem on PCs

The year 2000 date change problem on PCs is caused by the way the date
is stored in the clock chips used on PCs. All PCs (besides XTs) have a
Real Time Clock (RTC) which keeps track of the date and time when the
system is powered off. This is usually a dedicated chip, but the RTC
may also be integrated into an I/O chip or the main chipset. IBM used
the Motorola MC146818 in the original AT. The RTCs in all PCs since
then are based on this chip. The RTC chip is usually also used to
store CMOS setup values. The battery on the PC motherboard is present
to hold the values stored in the clock chip and to keep the clock
running. (Some clock chips, such as the Dallas 1287/1287A/12887/12887A,
contain the battery within the chip). The RTC has dedicated registers
for seconds, minutes, hours, days, months, and years, which it
increments itself. The register that holds the year value holds only
two digits, so 1999 is stored as 99. The year 99 will roll over to 00.
The clock chip does not have any way to keep track of the century by
itself. The original PC BIOS specification does provide for keeping
track of the century by way of a particular byte in CMOS RAM (32h),
which also is normally located in the clock chip. This byte is updated
whenever a BIOS call is made to set the date (INT 1Ah Function 05),
which happens whenever the DOS DATE command is used to set the date.
The century byte is also updated whenever a BIOS call is made to
read the date. DOS makes a BIOS call to read the date on every boot,
so it can set up its counter. DOS does not make a BIOS call when the
DOS DATE command is used to READ the date, since DOS is already
keeping track of the date, once the system is booted up.

When a PC is booted up, DOS gets the date and time from the BIOS,
which gets the date and time from the clock chip. As long as the
system is powered on, the clock chip and DOS are both keeping track of
the date and time separately and using different methods. The clock
chip uses a signal from an oscillator (which may be built-in to the
chip) to increment its clock. DOS calculates the date based on the
number of days since 01/01/1980 and calculates the time based on a
counter maintained by the BIOS in the BIOS Data Area. DOS then
increments its clock using this counter, which is based on a different
oscillator than the clock chip is using. As long as a PC is running,
DOS will maintain the date using 4-digit year format and the date will
roll over properly from 1999 to 2000. However, DOS is maintaining its
date and time separately from the clock chip, which is using a 2-digit
year format. DOS does not update the date or time in the clock chip
unless the DOS DATE or TIME commands are used to set the time or date.
(or actually if a BIOS call to read or write the time or date is
made). So if a PC is running at the turn of the century, DOS will
advance the date correctly but when the system is then powered off and
back on, DOS will retrieve the current date and time from the clock
chip which will include 00 for the last 2 digits of the year and 19
for the century. The clock chip does not have a means to account for
the century itself, and also has no way to alter the byte in CMOS RAM
that the BIOS uses to keep track of the  century. When the DOS DATE
command is used to set the date to the year 2000, the century byte in
CMOS RAM will be updated. Newer system BIOSes account for this problem
by assuming that if the 2-digit year accessed from the clock chip on
bootup (or whenever a BIOS call is made to read the date) is 00 (or
less than some value such as 80), then the century byte is set to 20
instead of 19. 

Windows 3.x will have no effect on this problem. That is, if the
system is running under WIN3.1x over the century change, the date
should roll over properly, just it would under DOS, but after
rebooting the system it would show a date of 01-04-1980.

Windows 95 also does not correct for the date rollover problem,
although it may appear to if testing is not done carefully.

Windows 98 does appear to correct for the date rollover problem,
however this correction is thwarted if the system is booted from a
DOS diskette on the first boot after the century change. The date
would then be 01/04/80 and when Windows 98 is restarted it will keep
this date until manually changed.

Keep in mind that software issues are a separate matter. A PC system
that is declared to be "century-compliant" is not immune to all
kinds of problems caused by software that is not century-compliant.

Another aspect of the year 2000 problem concerns the fact that leap
years occur every four years EXCEPT when that year is a century year
(such as 1800, 1900, 2000, etc.), in which case it is not a leap year
- EXCEPT every four centuries. The year 2000 is a leap year. This
could be a problem in some software that does not properly account for
leap years. This doesn't seem to be a common BIOS problem. We have
seen some BIOSes which will not allow a date of 02/29/2000 to be
manually entered in CMOS setup, but this does not prevent that date
from being handled properly otherwise.


3. Y2K Compliancy of Micro Firmware BIOS Upgrade Products

Micro Firmware BIOS upgrades which are based on Phoenix 4.05 code 
are Y2K compliant. These include I4HS10, I5HS10, I5HS15,
I5HS20, I5HS25, I5HS30, I5HS35, M4HS45, M4HS45G, M4HS45GC, M4HS45GP,
M4HS50, M4HS60, M5HS10, M5HS15, M5HS20, M5HS60, P4HS00, P4HS10, and
P4HS20. All revisions of these BIOS upgrades are Y2K compliant BIOSes.

Older revisions of our BIOS upgrades based on Phoenix 4.03 code are
not Y2K compliant. The current revisions of all of these upgrades are
Y2K compliant.

The following chart indicates which revision of each of our 4.03 BIOS
upgrades is needed to fix the year 2000 date rollover problem:

MFI Part Number    Minimum Revision Needed for Y2K Fix

M3GS20             4.03.6.09  Released 02/04/97
M3GS30             4.03.6.12  Released 02/21/97
M4GS20             4.03.6.09  Released 02/24/97
M4GS30             4.03.6.14  Released 02/26/97
M4GS25             4.03.6.17  Released 01/30/97

The M4GS25 BIOS upgrade is a flash BIOS. The current revision can
be downloaded from our updates page at:

http:/www.firmware.com/support/updates/

The other 4.03 BIOS upgrades listed above are supplied on EPROM
chips. Please contact us for information on replacing older revisions
of these BIOSes.

The I4GS10 BIOS is a flash upgrade based on 4.03 code. This BIOS is
not Y2K compliant, but has been replaced with the I4HS10 BIOS, based
on Y2K-compliant 4.05 code. The current I4HS10 BIOS (which will flash
in on top of the I4GS10 BIOS) can be downloaded from our updates page:

http:/www.firmware.com/support/updates/

Note that any responsiblity to support or update a BIOS on any PC
system rests entirely with the manufacturer of the motherboard and/or
system. Companies such as Phoenix, AMI, and Award produce only the base
code used in creating BIOSes. Normally the manufacturer of a motherboard
will adapt code licensed from a provider of source code to create a
BIOS for a specific motherboard. Phoenix, AMI, Award, etc. normally
have no control over what is done with the source code and do not have
any means to provide an updated BIOS for any particular motherboard.
Micro Firmware has no plans to produce any new BIOS upgrades for the
purpose of providing Y2K compliancy.

Unless a PC is left on all the time and is performing a critical task,
we do not recommend purchasing a BIOS upgrade just to fix this
problem, as (in most cases) it can be fixed very simply as described
elsewhere in this document by resetting the date on January 1, 2000.

What do we mean by a Y2K compliant BIOS?

A Y2K compliant BIOS will handle the year 2000 date rollover problem
at the BIOS level. When an operating system or application program
makes a BIOS call to get the current date (INT1Ah Function 04h),
if the BIOS finds that the year value in CMOS RAM is 00, it will
change the century value in CMOS RAM from 19 to 20.

Any PC BIOS should update the century value in CMOS RAM whenever a
BIOS call is made to set the date. A Y2K compliant BIOS will
additionally update the century value in the RTC anytime a BIOS call
is made to read the date if it finds that the year is 1900 (or within
some specified range). A Y2K compliant BIOS does not make a PC
completely Y2K compliant, since the RTC still may return an incorrect
date to an operating system or application if the BIOS is by-passed
and the RTC is accessed directly. (Note that this can only occur if
the century value has not already been changed). Also, there may be
any number of date-related problems present in operating system or
application program software.

There is one date-related issue that we know of in our Phoenix 4.03
and 4.05 BIOS upgrades. The Date field in CMOS setup will allow a year
value only up to 2030. This is only a problem with entering the date
manually into this field in CMOS setup. The BIOS will not prevent the
date from advancing correctly past 2030. If CMOS setup is entered
after the year has advanced beyond 2030, the correct year value will
be seen in the Date field. If this field is accessed and the year
value is changed, it will again be subject to the 2030 limitation. The
date can then be set at the DOS prompt and will continue to be kept
correctly.


4. Solutions to the Y2K Date Rollover Problem on PCs

Solutions include manually setting the date after the century change,
upgrading the system BIOS, installing a TSR program, installing a card
with a BIOS extension (such as the Micro Firmware Flash 2000 card),
replacing the clock chip, and replacing the motherboard. Each of these
solutions has advantages and disadvantages. Note that none of these
solutions (except replacing the clock chip) results in a "completely
Y2K-compliant" PC (and even then - we are not counting the software).

Manually resetting the date:

  Advantages:
    No cost (assuming no one is being paid to do it).

  Disadvantages:
    Requires manual intervention at the century change.
    May be impractical for sites with a large number of PCs.
    Some programs may need to run uninterrupted overnight.
    It may be forgotten until after problems have been caused by
    the incorrect system date.

Upgrading the system BIOS

  Advantages: 
    If an upgrade is available this is a preferred solution.
    Requires no manual intervention at the century change.

  Disadvantages:
    For most systems with this problem there is no BIOS upgrade
    available.

Installing a TSR:

  Advantages: 
    Requires no manual intervention at the century change.

  Disadvantages:
    May be removed when the TSR program itself or AUTOEXEC.BAT is
    accidentally deleted or when hard drive is replaced or
    reformatted.

Replacing the clock chip:

  Advantages: 
    Fixes the problem at the root.
    Requires no manual intervention at the century change.

  Disadvantages:
    Only certain types of clock chips can be replaced. The clock chip
    may be integrated into the chipset or it may be soldered onto the
    motherboard. Cost is about $25.00.

Flash 2000 card:

  Advantages: 
    Requires no manual intervention at the century change.
    Cannot be accidentally erased (as with a TSR program).

  Disadvantages:
    Takes up an ISA slot.

Replacing the motherboard or system:

  Advantages: 
    Requires no manual intervention at the century change.
    Most new motherboards will have a BIOS that corrects for the 
    problem. (Only a very few new motherboards will have a 
    Y2K-compliant clock chip).

  Disadvantages:
    High cost.


5. Testing PCs for Y2K Date Rollover Problem

We have written a test utility, Y2KTEST.EXE, which is supplied with
our Flash 2000 card and which can be downloaded by anyone from our web
site or ftp site.

ftp://ftp.firmware.com/y2k/y2ktst.exe

This is a small DOS program which can be run under Windows 95,
although it may be more accurate if run under a clean DOS boot. We
provide this program so that any PC can be tested to see if it may
need a solution for the date rollover problem and so that any PC in
which the Flash 2000 card has been installed can be tested to insure
that the Flash 2000 card is working in that system.

Some systems may have time-sensitive software. It is safest to run
any date-testing programs from a clean boot (boot from a bootable
diskette).

The text screens outputted by Y2KTEST.EXE are shown below. After
starting the program, a screen appears explaining the program. After
pressing Y to continue, the date is set to just before midnight on
12/31/99 and the RTC, BIOS, and DOS are tested to see what happens at
the century change. The program can return three different results -
the RTC rolls the date over correctly and no BIOS correction is
needed, the RTC does not roll the date over correctly and the BIOS
does not correct for it, or the RTC does not roll the date over
correctly and the BIOS corrects for it. After installing the Flash
2000 card into a system that was previously reported as having an RTC
that does not roll the date over correctly and a BIOS that does not
correct for it, the Y2KTEST program should then report that the RTC
does not roll the date over correctly and that the BIOS does correct
for it.

There may be some systems with Y2K compliant clock chips which our
Y2K test utility may indicate as having non-compliant clock chips.
This is because some newer Y2K compliant clock chips may use a
non-standard location for the century byte. It will be very rare that
a PC motherboard will use a Y2K-compliant RTC. There is only one
Y2K-compliant RTC that we know of which uses the standard location for
the century byte. This is the Dallas DS12C887. This chip can be used
as an upgrade on some motherboards and there are some new motherboards
that are supplied with this chip. If a different type of Y2K compliant
RTC is used which does not use the standard location for the century
byte, this chip may still not provide Y2K-compliancy at the RTC level
for applications or operating systems that access the RTC directly,
since such programs may be assuming that the century byte is in the
standard location.


This is the text that will display when Y2KTEST is run:

Y2KTEST Revision 1.03
Copyright 1998 Micro Firmware, Inc.  All rights reserved.
Website: http://www.firmware.com/
Email:   sales@firmware.com  support@firmware.com
Phone:   800-767-5465 or 405-321-8333

This program will test the Real Time Clock Chip and the BIOS on this PC 
to
determine whether it can successfully roll the date over from Dec 31, 
1999 to
Jan 01, 2000.

Note that it is expected that the Real Time Clock chip on almost all PCs 
will
not roll the date over correctly. What is of more concern is whether the 
BIOS
will correct for this hardware problem.

The date and time will be temporarily changed and will be changed back 
after
the test. You may wish to check your date and time after running this 
test.
Phoenix Technologies, Ltd. and Micro Firmware, Inc. assume no liability 
or
responsibility for any harm this free program may cause. If you do not 
agree,
please exit now by pressing N.

Press Y to continue or N to exit.


This text appears after testing if the RTC is not compliant and the
BIOS does not correct for it:


* The Real Time Clock chip on this PC does not successfully roll the 
date   *
* over from December 31, 1999 to January 01, 2000                           
*
*                                                                           
*
* The BIOS on this PC does not correct for the date rollover problem in     
*
* the clock chip.                                                           
*


The date and time were temporarily changed and were changed back after 
the
test. You may wish to check your date and time to ensure that they were
correctly restored. Phoenix Technologies, Ltd. and Micro Firmware, Inc. 
assume
no liability or responsibility for any harm this free program may cause.


Copyright 1998 Micro Firmware, Inc.  All rights reserved.
Website: http://www.firmware.com/
Email:   sales@firmware.com  support@firmware.com
Phone:   800-767-5465 or 405-321-8333


This text appears after testing if the RTC is not compliant but the
BIOS succesfully corrects for it:


* The Real Time Clock chip on this PC does not successfully roll the 
date   *
* over from December 31, 1999 to January 01, 2000                           
*
*                                                                           
*
* The BIOS on this PC corrects for this problem so the date will            
*
* successfully rollover to January 01, 2000.                                
*


The date and time were temporarily changed and were changed back after 
the
test. You may wish to check your date and time to ensure that they were
correctly restored. Phoenix Technologies, Ltd. and Micro Firmware, Inc. 
assume
no liability or responsibility for any harm this free program may cause.

Copyright 1998 Micro Firmware, Inc.  All rights reserved.
Website: http://www.firmware.com/
Email:   sales@firmware.com  support@firmware.com
Phone:   800-767-5465 or 405-321-8333


This text appears after testing if the RTC is compliant, which will be
true on only a small percentage of PCs:


* The Real Time Clock chip on this PC will successfully roll the date 
over  *
* from December 31, 1999 to January 01, 2000 and needs no correction in 
the *
* BIOS.                                                                     
*


The date and time were temporarily changed and were changed back after 
the
test. You may wish to check your date and time to ensure that they were
correctly restored. Phoenix Technologies, Ltd. and Micro Firmware, Inc. 
assume
no liability or responsibility for any harm this free program may cause.


Copyright 1998 Micro Firmware, Inc.  All rights reserved.
Website: http://www.firmware.com/
Email:   sales@firmware.com  support@firmware.com
Phone:   800-767-5465 or 405-321-8333


Manual Testing

Any PC can also be tested for the date rollover problem manually
(without running a testing program). Boot clean to a DOS prompt.
Using the DOS DATE command, set the date to 12/31/99, using the DOS
TIME command, set the time to just before midnight (11:57pm). Power
the system off, and then turn it back on after waiting long enough    
for the date to have advanced to 01/01/2000. On most PCs the date will
then be found to be 01/04/1980. 01/01/1980 is where DOS starts
calculating the date. DOS will usually set the date to 01/01/1980 if
the date and time info is lost, such as when a battery is replaced.
The reason for the 01/04/1980 has to do with diagnostic date codes
used by DOS. If DOS finds an out-of-range date (year must be
1980-2099, month must be 1-12, day must be 1-31), the date will be set
to 01/04/1980. Another example of the diagnostic date code is that an
invalid BCD (binary coded decimal) code in the calendar registers of
the clock chip will result in the date being set to 01/03/1980.

When performing a manual date rollover test, it may be necessary to
disconnect a network cable from the system, as some networks will
automatically synchronize the system time with the network server
time on bootup. This may occur even when not logging onto the network.
Also it may be necessary to disable the Daylight Savings Time 
adjustment feature in Windows 95. Otherwise, depending on the time of
year in which the test is run, Windows may set the time back an hour
after rebooting.

When performing the above manual testing in a system with a
Y2K-compliant BIOS (or with a Flash 2000 card installed) and setting
the date to test the century change with the system running, it may
be observed that if the system is turned off and back on and CMOS
setup is entered right away, the year may be indicated as 1900. This
is because there has not yet been a BIOS call, which is what causes
the Y2K-compliant BIOS to correct the century byte in the clock chip.
This does not mean that the BIOS is not correcting the date as far as
any programs that may get the date from the BIOS are concerned. To
test for this, the following DEBUG script can be run. This DEBUG
script makes a BIOS call to ask the BIOS what date it is.

C:\>DEBUG           (type DEBUG at DOS prompt)
-R AX               (enter command at dash prompt)
:0400               (enter this value at colon prompt)
-E 100 CD 1A CC     (enter command at dash prompt)
-G=100              (enter command at dash prompt)

AX=0000 BX=0000 CX=2000 DX=0101

                    Ignore the above register dump except for the
                    contents of CX (year) and DX (month/day)

-Q                  (type Q at dash prompt to return to DOS prompt)


Award BIOS dated 04/26/94 - 05/31/95

Some Award BIOSes will not accept any date with a year outside the
range of 94-99. The BIOS apparently checks the date during POST on
each boot and resets the year value if it is not in the range 94-99.
It will normally be reset to 1994 or 2094 (the century value is left
unchanged). If powered off over the century change, the date will
probably be set to 01-01-1994 on the first boot in the year 2000.

This problem is not corrected by the Flash 2000 card, although the
Y2KTEST program will erroneously report that the BIOS is Y2K compliant
after the Flash 2000 card is installed. 

This problem can be corrected by a BIOS upgrade (if one is available)
or by resetting the date every day. Actually, just setting the date on
each boot is sufficient - the system should maintain the correct date
as long as it is left on. One way to deal with this problem is to just
leave the system on, and set the date anytime the system is rebooted.
It could also be helpful to put the DATE command on a line in the
AUTOEXEC.BAT file, so the date will have to be checked and set on
every boot.

A BIOS upgrade will need to be obtained from the motherboard
manufacturer (or possibly from a third party).

Here is a URL to website with some information on how to identify
the motherboard manufacturer from the Award BIOS version. Contact
info for many manufacturers is also given here. The manufacturer
can then be contacted to ask if a BIOS upgrade is available.

http://www.ping.be/bios/numbers.shtml

To test for the type of BIOS discussed above, set the date to
01-01-2000 (or any date in the year 2000) and reboot the system.
A BIOS with the problem discussed above will have set the date to
01-01-2094, with or without a Flash 2000 card installed.

Phoenix Technologies, Inc. purchased Award in 1998. In early 1999,
the contents of Award's website was integrated into Phoenix's
website, which resulted in the loss of some links to Award's
site which were given in previous revisions of this document and in 
the printed documentation we supply with the Flash 2000 card.

In the Y2K FAQ on Phoenix's site, they state that work is being done
to develop a software fix for the Award BIOSes with the problem
described above. Here is a link to the Phoenix Y2K FAQ:

http://ww.phoenix.com/support/y2k.html


Recent Y2K-compliant Award BIOSes may not pass Y2K testing.

Many recent Award BIOSes will fail our Y2KTEST utility and various
other Y2K test programs, but will show the correct date when doing
a manual DOS date rollover test.

Award seems to be explaining that this is only an issue with passing
the test utilities, and that the BIOS is actually Y2K-compliant. Since
there is no "official" specification for Y2K-compliancy, these BIOSes
could be said to be Y2K-compliant, however, they are less "compliant"
than other BIOSes that correct for the Y2K date rollover problem.
These Award BIOSes will check the date on each boot to see if the
century needs to be corrected. This is why they will roll the date
over correctly in a manual DOS test. Other modern BIOSes will correct
the century anytime a BIOS call to get or set the date is made. So if
a system with this type of Award BIOS is left on during the century
change, any program that makes a BIOS call to get the date will get
the wrong date until the system is rebooted.

This problem was corrected by Award in November of 1996, but there may
be Award BIOSes with later dates that show this problem. BIOS updates
to correct this problem may be available from the motherboard
manufacturer. The Flash 2000 card can be used to correct for this
problem.


6. Does a Y2K-compliant BIOS Make a PC "Year 2000 Compliant"?

That depends on what is meant by compliant. On most PCs, a
Y2K-compliant BIOS (which may have been provided on the system,
or accomplished by a BIOS upgrade or by installing a card such as
the Flash 2000) will cause the date to roll over correctly. For most
users, this solves the Year 2000 problem. But for a system to be
declared completely "Year 2000 Compliant", it should have a compliant
clock chip. This is because there may be programs that access the
clock chip directly for the date and time. Such programs are not
common and would normally be found in industrial, military, or big
business environments where specialized, custom-made applications are
used. Such programs could also be considered non-compliant since they
are depending on the RTC to roll the date over correctly when it is
known that the RTCs in almost all PCs will not. Some operating
systems, including some versions of Unix, may access the clock chip
directly. Such OSes may account for this problem in current revisons
or have fixes available. The Flash 2000 card or a Y2K-compliant BIOS
will not solve the Y2K problem for OSes that access the RTC directly.

It is possible to upgrade the clock chip in many PCs to a Y2K
compliant chip. Dallas makes a chip called the 12C887, which can be
used to replace 12887 and 12887A and compatible chips and in most
cases can replace the 1287 and 1287A chips. This was released in
December of 1997 and can be obtained through a company called Resource
800. A URL for their web site appears below. Their phone number is
800-430-7030. Although most computer users will not need to be
concerned with replacing the clock chip, this issue may need to be
taken into account in cases where there is some government or
corporate regulation that needs to be met. Note that this will not
be an option for most PCs, since the clock chip is often soldered
onto the board or it may be a surface-mount chip or it may be
integrated into the motherboard's main chipset.

Also, for a system to be declared completely "Year 2000 Compliant",
all software installed on that system must be compliant. All kinds of
applications may have various problems with dates beyond 12/31/99.
These problems are solved only by updating the software and can not be
solved by the clock chip or the BIOS or any similar solution such as
the Flash 2000 card.


7. Time Dilation / Crouch/Echilin Effect / Time-leap

We have had some inquiries regarding "time dilation". Time dilation is
actually the name of a theory that Einstein came up with that has to
do with time slowing down due to motion. There is a web site or two
devoted to a problem involving the time and date randomly leaping
ahead on PCs after 01/01/2000. This was first referred to as time
dilaton, later renamed to the Crouch/Echilin effect, and sometimes
called "time-leap". They say that this occurs on 286, 386, 486, and
Pentium systems. They seem to have narrowed this down to systems that
have unbuffered clock chips. There are references to test utilities
but none seem to be available. No information is given as to what
clock chips are in the systems tested. No information is given as to
what type or version of BIOS is in any system tested. No information
is given as to what brand or model of systems have been tested.

Based on our own knowledge and consultation with a trusted authority,
we think that what is being seen here is that some older systems may
have broken BIOSes that do not follow correct procedures for getting
data from the RTC, which may cause this type of a problem on systems
which also have a non-buffered clock chip. (The Motorolla 146818 and
compatilble RTCs used in the original IBM AT are unbuffered as are
many RTCs built in to chipsets. All Dallas RTCs are buffered). It
should follow that any such system would have the same problem both
before and after the century change. We would then conclude that this
is just another of the thousands of very specific bugs that can be
found on specific hardware and that this is not a Y2K issue.

Although we would tend to think that this is a "non-issue" (at least
for most people), it is possible that there is more to this. Perhaps
more information on this subject will be forthcoming in the future.

Here are links to a couple of documents from Intel, in which they
say they have thoroughly investigated all proposed theories and have
found no evidence of the existence of the Crouch-Echlin Effect:

RETAINED HTML IN REMAINDER OF DOCUMENT FOR LINKS

<P>
<A HREF="http://www.intel.com/support/year2000/c-e-
wp.htm">http://www.intel.com/support/year2000/c-e-wp.htm</A> - Overview
<P>
<A 
HREF="ftp://download.intel.com/support/year2000/ceeffect.pdf">ftp://down
load.intel.com/support/year2000/ceeffect.pdf</A> - White paper
<P>
Here are links to Crouch and Echlin's web sites:
<P>
<A 
HREF="http://www.intranet.ca/~mike.echlin/bestif/">http://www.intranet.c
a/~mike.echlin/bestif/</A> - Mike Echlin
<P>
<A 
HREF="http://www.nethawk.com/~jcrouch/dilation.htm">http://www.nethawk.c
om/~jcrouch/dilation.htm</A> - Jace Crouch
<P>
And links to a couple of interesting articles:
<P>
<A 
HREF="http://www.y2ktimebomb.com/Computech/Issues/hsmith9845.htm">http:/
/www.y2ktimebomb.com/Computech/Issues/hsmith9845.htm</A> 
<P>
<A 
HREF="http://www.y2ktimebomb.com/Special/Opinion/Readers/hsmith9847.htm"
>http://www.y2ktimebomb.com/Special/Opinion/Readers/hsmith9847.htm</A> 
<P>
<HR>
8. For more information on Y2K:
<P>
<A HREF="http://www.year2000.com">http://www.year2000.com</A> - Lots of 
info including FAQ and many links.
<P>
<A HREF="http://www.rightime.com">http://www.rightime.com</A> - Detailed 
PC-related technical info.
<P>
<A HREF="http://www.dell.com/r&d/vectors/vect_1-3/v1-
3_century.htm">http://www.dell.com/r&d/vectors/vect_1-3/v1-
3_century.htm</A>
<P>
<A 
HREF="http://www.dell.com/r&d/whitepapers/y2k/y2k.htm">http://www.dell.c
om/r&d/whitepapers/y2k/y2k.htm</A>
<P>
<A 
HREF="http://www.mitre.org/research/y2k/">http://www.mitre.org/research/
y2k/</A> - Non-profit engineering group
<P>
<A HREF="http://www.resource800.com">http://www.resource800.com</A> - 
Source for Dallas 12C887 chips.
<P>
<A HREF="http://www.nstl.com">http://www.nstl.com</A> - National 
Software Testing Labs.
<P>
<A 
HREF="http://www.microsoft.com/year2000/">http://www.microsoft.com/year2
000/</A> - Microsoft FAQ and Product Guide.
<P>
<A 
HREF="http://www.ionet.net/~tslade/y2klinks.htm">http://www.ionet.net/~t
slade/y2klinks.htm</A> - huge collection of links
<P>
<HR>
<P><CENTER><H5><A HREF="/sales/index.htm">SALES -</A>
<A HREF="/support/index.htm">SUPPORT -</A>
<A HREF="/contact/index.htm">CONTACT -</A>
<A HREF="/links/index.htm">LINKS -</A>
<A HREF="/index.htm">HOME</A>
</H5></CENTER></P>
<HR>
<ADDRESS>Micro Firmware, Inc.<BR>330 W. Gray Street<BR>Norman, Oklahoma  
73069-7111<BR>
Toll-Free Sales (USA/Canada): 1-800-767-5465 or 1-888-4-PC-BIOS<BR>
Support & Sales:             +1 405-321-8333<BR>
Sales Department FAX:        +1 405-573-5535<BR>
Technical Support FAX:       +1 405-321-8342<BR>
Email: <A HREF="mailto:sales@firmware.com">sales@firmware.com</A>  <A 
HREF="mailto:support@firmware.com">support@firmware.com</A><BR>
<HR>
<ADDRESS>&#169 1998 Micro Firmware, Inc. - All Rights Reserved<BR>
While every attempt is made to ensure the accuracy of our support
documents, Micro Firmware provides these documents on an AS-IS basis
for information purposes only.

------------------------------------------------------------
END OF http://www.firmware.com/pb4ts/year2000.htm
------------------------------------------------------------



------------------------------------------------------------
From http://www.resource800.com/ho17035.html
------------------------------------------------------------
   
   Dallas Semiconductor (Clock/CMOS RAM Chips) Frequently Asked 
Questions (FAQ)


Dallas Semiconductor (Clock/CMOS RAM Chips) Frequently Asked Questions 
(FAQ)
If my computer uses a standard battery, can it also have a Dallas 
Semiconductor Clock/CMOS RAM chip?
How do I recognize Dallas Semiconductor modules on my system board?
Someone put a password in my Dallas Semiconductor DS12887 chip.  Is 
there any way to remove it?
My system won't boot after I replaced by DS1287.  What's wrong?
My Dallas Semiconductor DS12887 chip failed almost immediately after I 
installed 32 MB of RAM?  What's wrong?
Is there a way to trouble-shoot my system if it has a Clock/CMOS RAM 
chip installed?
When will Year 2000 compliant parts be available for DS12887s and 
DS12887As?
Can I substitute a DS12887 for DS12B887 or DS12887A?
Is the DS12C887 compatible with ANY of the Award 4.5X BIOS series?
Are the DS1397 and DS1497 being discontinued?
I'm confused!  What's the difference between Y2K compliant and Y2K 
capable?
My new system board uses a DS17887.  Is it Y2K compliant?
My PS/2 with Micro-Channel architecture uses a DS12887.  Can I make my 
system Y2K compliant with the DS12C887?
Can I use a software program to correct my operating system clock so I 
can be Y2K compliant?
I've seen ISA and PCI boards that seem to take care of the Y2K issue.  
Do they work?
I need a new power supply.  Where can I get one?
I took my old chip out and plugged it in the wrong way when I put it 
back.  Is it fried?
Will the DS12C887 correct an older BIOS that cannot generate dates past 
12/31/99?
I'm REALLY confused!  Is there a check-list or a single document that I 
can use to see if the DS12C887 will work in my system?

If my computer uses a standard battery, can it also have a Dallas 
Semiconductor Clock/CMOS RAM chip?
No.  Clock/CMOS RAM devices (Dallas Semiconductor, Odin, Benchmarq, 
Twinhead, et al) use tiny Lithium batteries that are encased in the 
device.  If your system uses a Nickel Cadmium (NiCd) battery soldered to 
the board, a coin cell
(soldered or in a cup socket) or a wire harness battery it won't use one 
of these devices.

Clock/CMOS RAM devices with onboard power CANNOT be added to systems 
with board mounted clock and CMOS powered by one of the above types of 
batteries.  If you are unable to find anything that looks like a battery 
on your system board, see the next
question in this FAQ.
How do I recognize Dallas Semiconductor modules on my system board?


Dallas Semiconductor "Smart Batteries" and computer clock Modules with 
on-board batteries are easy to identify.*  The most commonly used types
are:


DS1260

DS1287

DS12887

DS12887A


Dallas Semiconductor modules are distinctively marked with an alarm 
clock like the one shown below:



Service life on these devices is typically in excess of 5 years.  The 
manufacturer rates them at 10 years.  Heat can play a role in early 
failure, so 5 years is the recommended replacement period.

These devices are typically mounted in sockets on the computer's system 
board (see CAUTION below).  Disconnect the power cord and observe the 
Electrostatic Discharge (ESD) safety rules before touching any part of 
the computer.

 CAUTION:  We have encountered some systems where the module has been 
soldered directly to the systemboard instead of using a socket.  Don't 
attempt to remove the old unit and resolder a new unit in place unless 
you are an experienced,
board-level repair technician!

DS1260 "Smart Batteries" are available in 250, 500 and 1,000 mAH (milli 
Ampere Hours) rated packages.  These devices sense when the AC power to 
the system is on and switch themselves out of the circuit to conserve 
battery resources.  The
DS1287 and 12887 are completely self-contained CMOS RAM modules with 
their own on-board power and a real-time computer clock.  According to 
the Dallas Semiconductor data book, the DS1287 can be replaced by the 
DS12887.  The service life on these devices
is in excess of 5 years.  The manufacturer rates them at 10 years.


* other manufacturers (like Odin and Benchmark) that make competitive 
modules typically use the same number as Dallas Semiconductor models.  
They are usually pin compatible.  For more information, see the cross-
reference information in the section
entitled Dallas Semiconductor Devices.

Someone put a password in my Dallas Semiconductor DS12887 chip.  Is 
there any way to remove it?


Unfortunately, no.  Dallas Semiconductor products without an "A" at the 
end of the model number can only be reset using a password.  Our 
recommendation is to password protect all systems using this type of 
device with the same password and keep the
password under lock and key in one or more places.

Dallas Semiconductor devices that do have an "A" at the end of their 
model number can be completely erased even if they are password 
protected.  This is usually accomplished with a jumper on the system 
board.  It is also possible to use a device with
an "A" at the end of the model number in systems that don't have this 
feature.

To eliminate tampering on those systems that do not have a RAM clear 
jumper, we recommend that users password protect the device and tape a 
copy of the password INSIDE the computer case.



My system won't boot after I replaced by DS1287.  What's wrong?


We have encountered some systems that simply will not recognize the 
system clock or the CMOS settings after the clock/CMOS RAM device has 
been replaced.  It appears that the only way to get the system to 
recognize the clock is to use the
manufacturer's set up disk.  This allows the system BIOS to recognize 
the clock/CMOS RAM chip so the user can enter the setup information.


My Dallas Semiconductor DS12887 chip failed almost immediately after I 
installed 32 MB of RAM?  What's wrong?


The answer (as of 3/13/97) is: we don't know.  The following e-mail was 
graciously provided by Mr. Ken Price, director of research and 
development,  of Simulator Labbs.  Mr. Price has extensive knowledge of 
computer system boards and took the time to
reply to our query regarding the recent rash of computer clock devices 
(Odin, Benchmark and Dallas Semiconductor) with on-board power.  The 
information is provided by Resource 800 with Mr. Price's permission.

Date: Tue, 11 Mar 1997 23:54:44 -0800
Organization: Simulator Labbs
To: rescueme@onramp.net
Subject: Battery & Clock failure
X-URL: http://www.dallas.net/~r800/r800.htm

There has been a rash of failures recently in Dallas Semiconductor 
components.  From what I can gather given the facts I have available.

1.) Most failures occur after an upgrade in memory(Most 16 upto 32 or 32 
upto 64 Meg.)

2.) Most failures are in large cases acting as servers w/2+ drives 
active at all times.

3.) Random number lots and ages of the failed parts.

I would surmise that failure has to do with the memory subsystem in some 
way.  The most likely cause of this failure would in my opinion be 
caused by a slight overload in the bank, Ras and Cas switching logic.  
This can in my opinion be because
of:

1.) improperly set resistor packs on motherboard indicating for instance 
a single sided 72-Pin bank when it is really a two sided.

2.) Barely spec, or improperly rated memory.

3.) Mixed banks of single and double sided memory on motherboards using 
interleaved placement of SIMMs. 1-2-1-2 instead of 1-1-2-2, etc.

4.) Power supply being run two close to it's rated output.

The end result of this is both overshoot and undershoot of the power 
supply.  The way this happens is that the power coming into the MB is 
split up according to IBM's original specs which call for specific pins 
on CN8 and CN9 to do specific jobs.

This means that instead of having single ground and power planes for 5-
Volts you have multiple planes and end up with what is called switching 
noise on the power lines. This is a byproduct of the over and under 
shoot found on the logic lines. What
happens here though is that during power up there is twice as many or 
four times as many capacitors being charged in the DRAM and it puts a 
greater load on the lines going to both the memory and memory
controller. If the power supply senses this extra load(Seen as lower 
voltage) it will increase the amperage available(Raise voltage) to 
compensate.

So while the CPU with it's own line(Pin Not found on XT's) will see only 
a slight drop or raise in voltage because of local filters. The memory 
subsystem will feel it a great deal more. When the memory is charged the 
load goes away the discharging
inductors in the power supply will see a lesser load and raise voltage 
due to the lesser load. This causes a normally small spike in the 5-Volt 
line to the memory. If this spike is large enough(>0.7V) then it could 
in theory overload the input diodes in
the Dallas Semi. Devices causing failures with no reason for the 
failures detectable at a later time because the remainder of the machine 
is unaffected.

If this should be found to at fault then possible solutions are:

1.) Use a larger U.L. Rated supply(Add 100 Watts Min.).

2.) Use a higher frequency switching supply(Most are 30Khz., Use min 
50Khz.) as this will reduce the amount of compensation necessary during 
bootup and power on.

3.) Use a double filtered power supply so that spikes are absorbed by a 
second inductive/capacitive filter (More expensive).

4.) Sequence the power on of H.D.'s to lower the power on draw.
Most SCSI and some IDE drives have this feature.)

5.) Use a UPS so as to give a much more steady voltage to the power 
supply.  Another great source of spike is power fluctuations.

6.) Use larger single drives with larger partitions so there are fewer 
drives powering on at once.

Is there a way to trouble-shoot my system if it has a Clock/CMOS RAM 
chip installed?


If your system is 3-4 years old and you haven't made any significant 
changes like adding drives, newer modem or more RAM, replace the chip.

If you've made hardware changes to the system and begin experiencing 
problems, then you need to see what's causing them.  It might be a disk 
drive, modem, network interface card (NIC) or even a floppy drive.

The following e-mail may be of interest to those persons experiencing 
difficulty with these devices.  It tends to support our contention that, 
in many cases, the problems users are experiencing are with the system 
board or the power supply and not
with the Dallas Semiconductor device itself.



Hello Barry:

Since our conversation about Dallas Real-Time CMOS battery failure  I 
disconnected all peripherals and started reconnecting them one at a 
time.

When I got to the 5 1/4 floppy, it caused the CMOS to lose all  prior 
settings.  I reset the CMOS date, time and configuration information and 
left the drive disconnected.  It has functioned for almost three weeks 
without incident

The problem seems to be a weak power supply or some kind of high  
resistance short to ground.  If anything else develops in the near 
future, I will let you know

Thanks for your help, Barry.

Bob Leeling

P.S. It was one of your suggestions to do this.





The best way to trouble-shoot any system is to disconnect all of the 
peripherals EXCEPT for the video card and boot hard drive.  Remove all 
but 8 Megabytes (up to a maximum of 16 Megabytes) of RAM.  Perform a 
cold start (using the
power switch, NOT Ctrl-Alt-Delete (warm boot).

If the system starts, then you have either a peripheral malfunction or 
the power supply Wattage is inadequate.  Please note that you should 
have AT LEAST a 250 Watt power supply for a system using 32 Megabytes of 
RAM with a full load
of peripherals (fast video card, floppy disk, modem, CD-ROM, large (or 
multiple hard drives, tape drive, etc.).  Systems with 64 Megabytes of 
RAM should use a 300 Watt power supply.

In either case, a UPS is recommended to stabilize Voltage in the system.  
It is our experience that systems with weak or under-rated power 
supplies will surge and DESTROY the Lithium wafer batteries in 
clock/CMOS RAM devices like the
DS12C887.  This can happen in as little as 24 hours!  Surges also tend 
to damage other devices that oscillate at high frequency including the 
video card, modem, network interface card (NIC), CD-ROM and processor 
(CPU).

Begin connecting after each peripheral is reconnected to the system.  
Note: if you're using Windows 95, press F8 and go straight to the 
command prompt!  Then power down, reconnect another peripheral and 
restart.  If you don't this,
the Windows 95 will try to reconfigure itself for those items that are 
disconnected.  If you connect a peripheral (or add more RAM to the 
system) and it exhibits problems or fails to boot, then you will know 
that either the peripheral is malfunctioning
OR the power supply is failing or too small for the peripheral load.  If 
you need a replacement power supply, you can link directly to 
http://www.pcpowercooling.com/.

In the case of excess RAM memory, reduce the amount of RAM memory to the 
amount installed before the upgrade.  Most users report this as the last 
significant change to the system before the problems started.


When will Year 2000 compliant parts be available for DS12887s and 
DS12887As?


DS12C887 Year 2000 compliant devices are available for world-wide 
shipment.  Resource 800 will provide these parts for $29.95 US per 
device plus shipping and tax (where applicable).  Quantity discounts 
will be available.  Call 1-800-430-7030 for
details or send e-mail.

For complete information about the DS12C887 and the devices it will 
replace, see Dallas Semiconductor Devices.




Can I substitute a DS12887 for DS12B887 or DS12887A?


Yes.  The DS12887 or the Year 2000 compliant DS12C887 will replace 
either device.  However, the use of the RAM clear feature found on these 
devices will not be accessible.  In this case, we recommend that the 
device be password protected to defend
against possible tampering.  Make sure that a copy of the password is 
taped inside the computer case in the event of an emergency.


Is the DS12C887 compatible with ANY of the Award 4.5X BIOS series?


Dallas Semiconductor says that a small segment of devices in the Award 
4.5X (i.e. the 4.50G and 4.51PG) series are incompatible with the 
DS12C887.  It is best to test your system with one of the Y2K compliance 
analysis programs PRIOR
to ordering a replacement.

Year2000.Com is a small resident program for DOS, OS/2 and Windows 
published by The RighTime Company that fixes the year 1999 to 2000 date 
change flaw of the CMOS RTC in AT-class PCs and PS/2s, 286 through 
Pentium and its clones.  It works, and it's
free for personal, non-business use (a licensed business version is 
available; see below).  You can get a copy of the program from: 
www.RighTime.com or via their BBS at (305)644-6185.

The program will deteremine quickly and safely whether or not your 
system is:



Year 2000 compliant
Year 2000 compliant with intervention (date must be manually set at 
2000)
Not compliant and requires BIOS replacement, BIOS "flash" or clock/CMOS 
RAM chip replacement.



Contact information for the company is:



Tom Becker
The RighTime Company
Miami, USA
Phone: (305)644-6500 
Email: GTBecker@RighTime.com





Are the DS1397 and DS1497 being discontinued?


Yes.  According to Dallas Semiconductor, these devices will no longer be 
made past May 1999.  Users needing replacements will have to rely on 
existing stocks as these devices fail.  Compaq and other manufacturers 
used these devices in file servers.
Resource 800 will continue to provide these parts to users as long as 
supplies last.


I'm confused!  What's the difference between Y2K compliant and Y2K 
capable?


This is a confusing issue.  The simplest explanation is that Y2K 
compliant means that the system will AUTOMATICALLY roll over to the year 
2000 after 12/31/99 at 11:59:59.

Systems that are Y2K capable can be RESET to function in the new century 
after 12/31/99 at 11:59:59.  In other words, these systems require 
manual intervention in order to be able to provide the correct date.


My new system board uses a DS17887.  Is it Y2K compliant?


The DS17887 along with the DS1687, DS17287 and DS17487 are Y2K 
compliant, however the location of the century counter does not conform 
to the PC-AT or PS2 specifications.  The century counter is located in 
bank one location 48h.


My PS/2 with Micro-Channel architecture uses a DS12887.  Can I make my 
system Y2K compliant with the DS12C887?


The answer is "No," according to knowledgeable sources at Dallas 
Semiconductor.  The reason is that the Micro-Channel architecture (MCA) 
used by IBM in the PS/2 series maintains the century bit information at 
37HEX.  AT-type systems
maintain century bit information at 32HEX.  The DS12C887 will not work 
in systems that use Micro-Channel architecture (MCA).  Users who need 
more information on this subject should contact IBM directly.



Can I use a software program to correct my operating system clock so I 
can be Y2K compliant?


Yes and no.  Yes, some software programs that run at boot time can 
correct the operating system clock and allow you to provide the correct 
date and time to application programs.  However, if ANY application 
programs makes a call
DIRECTLY to the hardware clock, it will get the wrong date.

NOTE - most applications use the software clock.  Others, especially 
engineering, scientific, machine control, CAD and accounting programs 
use the hardware clock for date information (See the next question for 
additional information
about ISA and PCI cards that attempt to correct hardware clock dates).


I've seen ISA and PCI boards that seem to take care of the Y2K issue.  
Do they work?


It depends on the board.  You should also be aware that, when pressed on 
this issue, the board manufacturers will tell you that there is always a 
latency issue when using this type of retrofit solution.

These devices implant the century information in the system in place of 
the BIOS and the associated hardware clock information.  Application 
programs that make all their calls to the operating system clock (or DOS 
clock) won't have any problems.
However, ANY application that makes a call to the hardware clock MAY may 
encounter the latency between the time the Y2K board intercepts the 
hardware clock information and relays it to the operating system clock.  
In
this case, the application can get the wrong time and spawn bad dates 
throughout application software and corresponding data.


I need a new power supply.  Where can I get one?
You may be able to find one locally.  Be sure to look for a at least a 
250 Watt, double-filtered, UL-rated supply.  If you have more than 32 
Megabytes of RAM memory and/or a full load of peripherals, purchase a 
300 Watt,
double-filtered, UL-rated supply.

We also recommend using a UPS on the system.  These devices help 
stabilize Voltages in the system and help protect from brown-out 
conditions and power surges.  Be sure to purchase one with a phone-line 
surge suppression socket as most power surges
(especially from lightning!) enter your system via this route.

If you can't find a power supply locally, visit 
http://www.pcpowercooling.com/.  They provide a complete line of well-
built power supplies and accessories.  You can easily match the power 
supply in your
system to the detailed drawings at their web site. 
I took my old chip out and plugged it in the wrong way when I put it 
back.  Is it fried?
Probably.  These devices are made to be plugged in ONLY ONE WAY!  Please 
be sure to note the position of the dot (lower left-hand corner) of the 
chip.  We recommend putting a small dot of ink (or Liquid Paper) on the
system board to note the proper placement.
Will the DS12C887 correct an older BIOS that cannot generate dates past 
12/31/99?
Yes.  According to Dallas Semiconductor, the DS12C887 CAN correct older 
BIOS versions by providing four century digits at the 32H address.  In 
this case, the system performs correctly (i.e. falls in the Y2K capable
category) once the correct date is set.
I'm REALLY confused!  Is there a check-list or a single document that I 
can use to see if the DS12C887 will work in my system?


This is what you need to do or know, arranged in priority, BEFORE 
ordering a DS12C887 Y2K compliant device for your system.  Resource 800 
provides this information as a consumer service.  Please note that we 
cannot provide technical
assistance via our toll-free sales number.  We will gladly provide 
copies of the following information, our price list and the relevant 
Frequently Asked Questions (FAQ) document via e-mail.

At a minimum, you MUST know ALL of the following:



Who manufactured the device in the system?  This could be: Dallas 
Semiconductor, BenchMarq, Twinhead, Odin or Houston Tech.
How long the original device lasted, especially if it failed in less 
than five (5) years.
Is the device socketed or soldered?
The total amount of RAM memory is in the system.
Whether or not you or someone else made significant changes to the 
system, including adding peripherals and RAM memory up to or in excess 
of 32 Megabytes.
Whether or not there are peripherals that are plugged into the serial or 
parallel port that remain on AFTER the system is powered off.
The BIOS manufacturer and FULL version number.
The power supply Wattage.
Whether or not the system has been tested for Y2K compliance using one 
of the programs available via this site or a commercial software 
offering.




If you can't answer the above questions, you must read and implement the 
following BEFORE ordering or inserting the DS12C887 Y2K-compliant 
clock/CMOS RAM device.





STEP ONE - Open your system case and determine which (if any) device is 
used in your particular system.  If you are not familiar with the 
various types of computer clock/CMOS batteries and semiconductor hybrid 
devices used read Computer Clock (CMOS) Battery Types BEFORE continuing.  
Please note that you cannot determine whether or not a clock/CMOS RAM 
chip has been used in your system without visually inspecting the 
system.  Please do
not rely on manufacturer schematic diagrams as they are notoriously 
inaccurate.

If your system uses a board-mounted battery like a soldered-on NiCd, a 
coin-cell or Lithium barrel or wire-harness (i.e. external) battery, 
STOP AT THIS STEP!.  You CANNOT use the DS12C887 in your system.  The 
system
contains a chip set clock and you will have to replace the BIOS in order 
to effect Y2K compliance.

For your convenience, here is a list of the major BIOS manufacturers 
along with their contact information and URLs:



AMI (American Megatrends, Inc.)

According to AMI, "All current versions of AMIBIOS 95, core date 7-15-
95, and future versions of AMIBIOS are year 2000 compliant.  All 
previous OEM versions of AMIBIOS can be changed only by the system OEM 
to support the year 2000 feature.  This
change can be given directly to the system OEM from American 
Megatrends."

Specific information on Year 2000 (Y2K) compliance is available via: 
http://www.amibios.com/2000.html.

The contact information for American Megatrends, Inc. is:



American Megatrends, Inc.
6146-F Northbelt Parkway
Norcross, GA 30071-2976 USA
Phone: 770-246-8600
Technical Support: 770-246-8645 



The URL for the AMI website is: http://www.megatrends.com or 
http://www.amibios.com.


Award Software International

The following information characterizes the v4.50G and v4.51PG BIOS 
series:

Systems using the Award v4.50 series BIOS (most commonly v4.50G) may not 
accept any dates after 1999 nor prior to 1994.  In some cases, the Award 
v4.51PG BIOS yields the year 2096.  Award recommends that end-users "get 
a BIOS upgrade from the OEM (the
machine or motherboard's supplier or manufacturer)."  Users that are 
unable to determine the Original Equipment Manufacturer (OEM) or who 
cannot contact them, should determine the BIOS part number from the top 
of the CMOS Setup screen and contact Award
directly.

The contact information for Award Software International is:



Award Software International Inc.
777 East Middlefield Road
Mountain View, CA 94043-4023 USA
Phone: 415-968-4433 
FAX: 415-968-0274 
E-mail: support@award.com



The URL for the Award web site is: http://www.award.com.



Phoenix Technologies

Phoenix Technologies does not provide BIOS upgrades directly to end-
users.  BIOS upgrades must be obtained either from the Original 
Equipment Manufacturer (OEM) or from Micro Firmware, a company 
specializing in PhonenixBIOS upgrades.

We recommend that you read the information found in this document found 
at the Micro Firmware website: 
http://www.firmware.com/pb4ts/year2000.htm.

The contact information for Microfirmware is:



Micro Firmware, Inc. 
330 West Gray Street, Suite 120
Norman, OK 73069-7111 
Toll-free: 800-767-5465
Phone: 405-321-8333
E-mail: support@firmware.com



The URL for the Micro Firmware web site is: 
http://www.microfirmware.com.



STEP TWO - If you determine that your system uses a clock/CMOS RAM 
device, you should next check the BIOS manufacturer and series for 
compatability.  You MUST do this by booting the system, noting the 
information
presented by the system.  Reading the information on the BIOS chip 
itself is inadequate and will almost never provide the provide the 
actual series number.


A. If the system will not boot, read the following:



The best way to trouble-shoot any system is to disconnect all of the 
peripherals EXCEPT for the video card and boot hard drive.  Remove all 
but 8 Megabytes (up to a maximum of 16 Megabytes) of RAM.  Perform a 
cold start (using the
power switch, NOT Ctrl-Alt-Delete (warm boot). VERY IMPORTANT: if you 
are using Windows 95 or Windows 98, press shift-F5 during the boot 
process and go straight to the DOS command prompt!  If you don't this, 
the Windows
95 will try to reconfigure itself for those items that are disconnected.  
If the system boots properly, power down, reconnect another peripheral 
and restart using the power switch.

If the system misbehaves or won't boot at all during this process, then 
either a peripheral malfunction exists or the power supply Wattage is 
less than adequate for the configuration.  Systems that will not hold 
configuration information in CMOS RAM
during a warm-boot or those that stall and/or lock up at boot time are 
indicating malfunctions in some sub-system OTHER THAN the clock/CMOS RAM 
device!

SPECIAL NOTE - systems equipped with a device that ends with or contains 
an "A" (i.e. DS12887A, BQ3287A, etc.) may have the RAM clear jumper set 
in the wrong position.  READ THE SYSTEM BOARD
DOCUMENTATION to determine which position the device must be in TO 
RETAIN the CMOS settings.  In some cases, the RAM clear jumper must be 
engaged to clear the machine BEFORE a new clock/CMOS RAM device can be
inserted.

Please note that you should have AT LEAST a 250 Watt power supply for a 
system using 32 Megabytes of RAM with a full load of peripherals (fast 
video card, floppy disk, modem, CD-ROM, large (or multiple hard drives, 
tape drive, etc.).
Systems with 64 Megabytes of RAM should use a 300 Watt power supply.

In any case, a UPS is recommended to stabilize Voltage in the system.  
It is our experience that systems with weak or under-rated power 
supplies will surge and DESTROY the Lithium wafer batteries in 
clock/CMOS RAM devices like the
DS12C887.  This can happen in as little as 24 hours!  Surges also tend 
to damage other devices that oscillate at high frequency including the 
video card, modem, network interface card (NIC), CD-ROM and processor 
(CPU).

If you connect a peripheral (or add more RAM to the system) and it 
exhibits problems or the system fails to boot, then you will know that 
either the last peripheral installed is malfunctioning OR the power 
supply is failing or is too
small for the configuration.  If you determine that you need a 
replacement power supply, you should consider  linking directly to 
http://www.pcpowercooling.com/.

In the case of excess RAM memory, reduce the amount of RAM memory to the 
amount installed before the upgrade.  Most users report this as the last 
significant change to the system before the problems started.

NOTE: These devices are rated by the manufacturer for a ten (10) year 
life.  We rate them at five (5) years, nominally.  If the device in your 
system failed in less than five (5) years (typically about two (2) 
years)
and/or you have made major changes to the system peripherals including 
adding hard drives (especially SCSI devices), other high-drain 
peripherals like storage devices or RAM memory you have a system-based 
problem.

Typically, this is due to a weak or inadequate power supply.  In this 
case, you may have to spend $100.00 US, or more, plus the cost of a BIOS 
(flash program or new EPROM chip at a cost of $49.95 US) from a BIOS 
manufacturer, plus the cost of the
clock/CMOS RAM chip ($29.95 US).  In some cases, it may be less 
expensive to purchase a new system board.  Resource 800 makes a business 
out of trying to solve more problems than it creates.  Therefore, this 
option may be better for you in the long-run.



B. If the system starts, determine whether or not the clock/CMOS RAM 
device is socketed or soldered to the system board.  If the device is 
soldered, you will have to remove the system board from the case, de-
solder the old device, solder in a 24 pin
DIP socket (available from Resource 800), re-install the system board 
and peripherals and insert the new device into the socket.  Caution:  
You MUST remember to mark pin one on the system board BEFORE de-
soldering the
old device!  The 24 pin socket can be inserted backwards but the 
DS12C887 CANNOT!.


C. Verify the manufacturer and device model for the clock/CMOS RAM 
device.  The DS12C887 CANNOT replace the following: DS1387, DS1488, 
DS1397, DS1497 and DS14287.  These devices ARE NOT Y2K compliant and 
there are no
replacements planned for this series.  For more information, please read 
the appropriate entry in the Dallas Semiconductor (Clock/CMOS RAM Chips) 
Frequently Asked Questions (FAQ).



The current list of pin-compatible items for the DS12C887 includes the 
following:



BenchMarq 

   
   BQ3287MT 
   BQ3287AMT 
   BQ3785P DIP 
   

Dallas Semiconductor 

   

   DS12885 DIP 
   DS1285 DIP 
   DS12B887 
   DS1287 
   DS1287A 
   DS12887  
   DS12887A 

   

Houston Tech * 

   

   HT12888A 

   * ONLY if DS12885 DIP used as base assembly. 

   

Odin 

   
   OEC12C887
   OEC12C887A
   

Thomsen-Mostek

   
    M48T86 
    MK48T87 
   







D. Verify the BIOS manufacturer and series:



If the BIOS is an Award 4.50G or 4.51PG, contact the BIOS manufacturer 
(see contact list above) for an upgrade if you want to purchase the 
DS12C887 Y2K compliant device.  The DS12C887 WILL NOT work with these 
BIOS versions!  In many
cases, the system will not even boot!  You can, however, purchase one of 
the non-compliant device (i.e. DS12887, DS12887A) but your system will 
most likely not calculate dates beyond 12/31/99.


If the BIOS is an AMI or Phoenix BIOS, you should test the system with 
one of the Y2K compliance analysis programs.  We currently recommend the 
program provided by The RighTime Company.  The free download contains a 
test program and a
terminate-and-stay-resident (TSR) program that can correct some time-
related defects in older systems.

DO NOT INSTALL the program without reading the documentation!  Un-ZIP 
the program, read the documentation and run the test FIRST!.

The program will tell you quickly and safely whether or not your system 
is:



Year 2000 compliant
Year 2000 compliant with intervention (date must be manually set at 
2000)
Not compliant and requires BIOS replacement, BIOS "flash" or clock/CMOS 
RAM chip replacement.



Year2000.Com is a small resident program for DOS, OS/2 and Windows 
published by The RighTime Company that fixes the year 1999 to 2000 date 
change flaw of the CMOS RTC in AT-class PCs and PS/2s, 286 through 
Pentium and its clones.  It works, and it's free
for personal, non-business use (a licensed business version is 
available; see below).  You can get a copy of the program from: 
www.RighTime.com or via their BBS at (305)644-6185.

Contact information for the company is:



Tom Becker
The RighTime Company
Miami, USA
Phone: (305)644-6500 
Email: GTBecker@RighTime.com



Note: Resource 800 CANNOT provide technical support for this program.  
You should read the documentation in its entirety and then contact The 
RighTime Company for assistance.


STEP THREE - 

If you have determined that your system is operating properly, has a 
power supply with enough Wattage to cover the system's needs, has a 
compliant (or capable; see definition above) BIOS that is compatible 
with the DS12C887, call 1-800-430-7030 to
order or use the secure form at http://www.resource800.com/ho13008.html.

Remember that there are ALWAYS exceptions and we have found some 
situations  where the DS12C887 will not run with older BIOS versions.  
In these cases, you must upgrade the BIOS via the manufacturer (see BIOS 
manufacturer contact
information above)

Thank you very much for taking the time to read this information.  We 
hope that it was informative and provided you with the data needed to 
make a decision about upgrading your system to Y2K compliance.

The Resource 800 Technical Support Team


Copyright c1998 by Resource 800 - ALL RIGHTS RESERVED


------------------------------------------------------------
END OF http://www.resource800.com/ho17035.html
------------------------------------------------------------

