-1

I was recently diagnosing a time drift issue for a client located in a different city. Their workstations were having issues accessing the Windows 2019 server due to this drift.

Based on the age of the device, my initial instinct was a dying CMOS battery. I thought maybe AIDA64 might provide those details, but it did not. Another user on here recommended using a program called hwinfo, which after trying on my own device, I found works great, but being unable to install unapproved software on this device, I was unable to use it for my purposes..

SOO.. I first changed the time in the GUI, then continued my other tasks. Over that 1 hour, the clock again drifted by almost 1 min. ..CMOS I ran some cmds, mostly w32tm and saw CMOS as the clock source, but being very far away from the server, and not trusting anyone onsite to change the battery, I reassigned the source to pool.ntp.org and changed related settings, with help from this walk-thru: https://icookservers.blog/2014/09/12/windows-ntp-server-cookbook/

I cannot verify the voltage, but based on

  1. the observed drift over 1hr

  2. reg entry: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\W32Time\Config\AnnounceFlags set to 10 (this tells the server to sync time to local CMOS clock)

3.the results from w32tm /query /source (Local CMOS clock)

4.the results from w32tm /query /status /verbose (the computer did not resync because no time data was available)

5.Buncha other things it seemed undeniable that all configs pointed to be dependent upon a failing source...

Questions: how reliably will this work as a workaround to prevent this time drift and does anyone know any methods using CMD or PowShell or built in utilities that will provide a V+ reading on the battery?

If it is not a dying CMOS, and if it wasnt caused by the default windows time server settings, what else can I do to figure out the cause?

I personally would love an on-the-clock, out of town trip to change a small cell battery as much as the client would hate it, but in the meantime...

Thanks

Improve this question
6
  • Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer.
    – Community Bot
    Commented Mar 21 at 20:59
  • harrymc I referenced that post before posting this one, and no it does not. My experience was about the same as the OPs in that link.
    – NerdAlert
    Commented Mar 21 at 22:06
  • That post was helpful though, and is where i discovered the freeware hwinfo, which when i used it on a different computer, did provide a measure on the CMOS voltage.
    – NerdAlert
    Commented Mar 21 at 22:13
  • In that post, which basically asks the same question I am, their issue was hours of forward drift and I assume they could install hwinfo on their device. I am not permitted to install 3rd party software on this device.. if i was then hwinfo would have solved my problem.
    – NerdAlert
    Commented Mar 21 at 22:21
  • Caveat: HWiNFO and HWiNFO 64 can report only what the PC makes available. On the machine I'm using now, CMOS battery voltage is not shown (HWiNFO v. 7.72).
    – DrMoishe Pippik
    Commented Mar 22 at 0:37

1 Answer 1

Reset to default
0

Crystals used on PCs are not that accurate or stable. The clock always drifts, regardless of the CMOS battery condition. You can use a time server to sync the clock once a day. If the CMOS battery is dying and the PC is being turned off, then it will lose time. In which case, replace the battery. IT should last 5 to 10 years.

In your case, since it is drifting faster, I would suggest

  1. Taking them to a repair shop to replace the crystals/TCXO
  2. Writing a small program that syncs it faster. I have written such a program some years ago for PCs being used for industrial control. The operator would input the error/drift (once a day) and then the program would distribute the error over the day and correct the clock by half a second at a time. The frequency of change would depend on how much correction was needed. It only takes a few days to bring it into line.
Improve this answer
1
  • I've always thought the point of the battery and RTC was to keep a 'mostly' correct time, and settings while off, and let NTP clocks sort out post boot
    – Journeyman Geek ♦
    Commented Apr 12 at 14:36

Not the answer you're looking for? Browse other questions tagged .