Windows losing time, on a Dell Optiplex 2

[Yiddo]

I having problems with a Dell Optiplex GX1 running windows 95. The clock on the windows desktop is losing time. It varies from beteww 30 seconds to a minute a day.

I have tried replacing the battery, and have even installed a new motherboard to try and solve the problem. This hasn't worked. Does anyone have any other things I can try.

 

[rephlex]

Not gonna be too helpfull I'm afraid but I had the same problem on a Dell Dimension - never found a solution

Sorry but I think you may be stuck with it

Rephlex

 

[iSeriesCodePoet]

If you are worried about the correct time, you can use a cool product from

http://www.analogx.com.

It syncs your time with the national time servers. Mike Wills
RPG Programmer (but learning Java)

"I am bad at math because God forgot to include math.h into my program!"

 

[pfleischman]

If you have an older version of Norton Antivirus running on the system, this could be causing the time loss. Try program updates from Symantec or install a newer version. I assume because NAV causes this problem, a similar problem could exist with other apps that are running in the background. Try disabling some of these to see if the time loss still occurs.

 

[cdogg]

According to just one article I found at Dell's website:

___________________________________


Why does my clock in Microsoft® Windows® 98 lose or gain time?


Troubleshooting time loss or gain on a computer can be complicated. However, there are a few troubleshooting steps that may help determine the root cause if the problem. If the following steps do not help resolve your problem, or if you would like additional information on this subject, please review Dell Knowledge Base Document HO1016518.

To troubleshoot a time problem in Windows 98, perform the following:

Left-click on START | RUN.
Type in "msconfig" in the "Open" dialog box and left-click on "OK."
When the "System Configuration Utility" window appears, the "General" tab will be visible. Left-click on "Selective Startup."
Remove the check marks from each choice except "System.ini," by left clicking on the check marks.
Left-click on "Apply" and then left-click on "OK."
A message will appear stating that "For the changes to take place, the system must be restarted." Left-click on the "OK" button, and the system will begin the restart.
When the "Dell" logo appears, press the [Delete] key on the keyboard 2 to 3 times. The message "Now Entering Setup" will appear. When the initial, or "Main" page appears, the time and date, among other things, will be displayed.
Use the [Arrow] keys to move to the "hours" and type in the proper hour. The

arrow key will then move the cursor to minutes, which should also be set appropriately. Since it is virtually impossible to determine the seconds with any accuracy, do not change them.
Press the [F10] key on the keyboard. A message will appear asking if the changes should be saved and if setup should be exited. Insure that "Yes" is highlighted and press the [Enter] key. The system will reboot.
When the Windows 98 desktop appears, check the digital clock located on the task bar in the lower right corner of the screen. It should now be accurate. Left-click on START | SHUTDOWN | RESTART, then click on "OK" and the system will reboot. When the Windows 98 Desktop appears, check the clock again.

If the system now keeps correct time, then the time loss or gain was caused by a software program loading when the system boots. Please refer to Dell Knowledge Base Document HO1014939 for information on how to troubleshoot using the msconfig utility.

After you have isolated the application or utility that is causing the problem, you will be able to proceed with further troubleshooting. Remember to undo your changes when you have solved the problem. The entry causing the problem can be left disabled if it does not affect system operation. One of the easiest methods to attempt a resolution is to reinstall the problematic application.

For additional help with Windows 98 functions, click on START | HELP. When the "HELP" menu opens, click on the "INDEX" tab. Type in a key word to find the desired task you need help with.

For help with specifics on your Dell system or additional troubleshooting support, refer to

http://www.support.dell.com.

Dell provides you with the troubleshooting tools to resolve most any problem.

_________________________________


Of course, there are other articles on hardware issues as well, so you may want to venture through their website a bit more than I did...​

 

[ErikZ]

The computer clock is exact, but if your computer is busy, updating the clock is the last thing on it's list to do.

The more work you do on your computer, the faster the clock will fall behind. The only way to ensure a reasonable time is to have a time program periodically check an atomic clock on the internet.

 

[cdogg]

Erikz,

I respectfully disagree. Though it would seem to make sense that clock time is lost due to unavailable CPU cycles, the BIOS/CMOS is in charge of keeping this time constant (hence the CMOS battery when the system is offline). So, at the very least, the clock in the OS would be updated once the cycles settled down or after a reboot.

Now, it is natural to lose a few seconds off your clock in a couple weeks, but it's not natural, for any reason, to lose 30 seconds in a day. If this were normal, you would have seen this all the time on older systems running 386/486 processors, or even on the original Pentium. I never encountered such a loss in one day's time.

The problem may lie with the BIOS itself or even be a hardware problem of some sort. The first thing to check always is the battery, which has been done. The other would be to upgrade the BIOS through flashing...

 

[ErikZ]

Cdogg, no, *I* repectfully disagree.
BIOS/CMOS does update the time, it knows what time it is from the frequency crystal by counting the waveforms. However, if it's busy doing other things, it will miss a wave.

Older systems were much slower and accesses the BIOS far less than todays systems. Why do you say that they would be worse that the modern systems?

My friend owns an ISP, he has noted that if he is not sitting in front of the computer doing things, his time doesn't change. if he's working all day and stressing the computer, he looses time.

 

[cdogg]

Well, perhaps it is true that if the system is stressed that a wave might be missed. However, I always thought that clock cycles weren't involved with the quartz clock onboard. This would hold true when the system is offline (no CPU and is using the CMOS battery).

If what you're saying is true, then it seems to me that it is by poor design that the clock wasn't designed to get a constant feed of power from either the battery or DC current. When I first read this post, it sounded like the "OS" was losing time, not the "BIOS". I'm willing to bet that a reboot would remedy the loss of time in the OS...

 

[ErikZ]

I don't think it's possible for just the OS to loose time. It's a system call, and it just asks the Bios.

I also don't think it's possible for a solid state device to tell time without a "Vibrating" crystal. The motherboard can only power it's components with DC power, so if you can power any component with a battery when the board is powered down.

 

[cdogg]

I'm not sure I understand your last post: ", so if you can power any component with a battery when the board is powered down "

I'm not in total disagreement with you, but I still have a feeling of doubt that clock cycles can have a "direct" effect on the CMOS time.

I'll do a little research and post back. BTW, I do appreciate your replies.


~cdogg

 

[ErikZ]

Heh, sorry, bad English. Get rid of the "if" in the confusing statement.

Hey, I found something interesting

The RTC clock is not as accurate as your wrist watch. A typical wristwatch is rated at about plus or minus 15 seconds per month or 3 minutes a year. A typical PC clock is rated at about plus or minus 1 minute per month; many motherboards are off five times that much (10 seconds per day). This is because the chip maker cannot predict the temperature of the motherboard that the RTC chip is plugged into. Some PC's run warmer than others, and dust accumulating will cause even higher temperatures. The crystal's frequency will change slightly as it ages. There are several solutions to this inaccurate clock problem. What most of us do is use the Network to synchronize our clocks, perhaps each day or each time a network connection is established. For more information, see the OTC Web page:

http://www.otc.psu.edu/services/staff/data/ntp.html

 

[cdogg]

Erikz,

I found something as well...

First of all, extreme temperatures inside the case will affect the RTC's abiblity to keep accurate time. Like you mentioned, over time this can cause permenant defects in the crystal resulting in severe losses of time in short periods.

However, back to our original argument...

The "software" clock that runs when the OS loads only synchronizes with the RTC at startup. This isn't constant. As you said before, stressing the system can cause this clock to lose time very easily and depends on the amount of system interrupt requests that are made at once. The more often this occurs without rebooting, the farther off the "software" clock gets.

In an article I found, the RTC cannot display fractions of a second. This means that the software clock can be off as far as .99 seconds at any given point in time (even after rebooting).

To sum it up, we were both right. There is a hardware and software side to the clock. The hardware side is completely sustained by the battery and unaffected by CPU cycles. The only real stressing factor is temperature. However, Yiddo's problem is that his "hardware" clock (or RTC) is damaged. That's why his clock in his OS is off as well. A reboot in his case doesn't make a difference...

check this article to follow up with what I've said here:

http://www.beaglesoft.com/mainfaqclock.htm

~cdogg

 

[cdogg]

One more thing...

When you said:
"BIOS/CMOS does update the time, it knows what time it is from the frequency crystal by counting the waveforms. However, if it's busy doing other things, it will miss a wave."


We should both be able to agree now that the BIOS/CMOS holds the data for the "hardware" clock, but is unaffected by software that's running. The battery in the system keeps the RTC running, not the BIOS/CMOS, even when the system is powered on. The RTC is Not dependent on interrupt requests or CPU cycles. Therefore, a system that's being heavily taxed by software and interrupt requests to the CPU/Memory bus will not affect the RTC directly. Only the software clock in the OS might suffer a loss...

 

[ErikZ]

You know you're a geek when you find yourself facinated by a web-page on how the computer keeps time. Excellent find cdogg! Star for you!

It's obvious from reading that webpage that time-keeping is a multi-layered task. According to the page, OSs like Win95 cause problems by loosing time, and then updating the RTC with the incorrect time.

 

[cdogg]

The funny thing is it didn't take long to find the page, though I ended up spending twice the time analyzing it!

Yes, sometimes my curiosity turns into a crusade!!
B-)

 

[Guest_imported]

On the other hand, Windows can change you BIOS clock as well. Therefore, if there is a windows problem, this could possibly cause the computer to loose time.
A little food for thought from Computer Call Center Agent.

 

[cdogg]

Groda,

Did you not read the last post by Erikz?

"It's obvious from reading that webpage that time-keeping is a multi-layered task. According to the page, OSs like Win95 cause problems by loosing time, and then updating the RTC with the incorrect time."