Not a "problem" but a puzzle: variable screen saver timing delaying sleep

Desktop computer: Dell XPS 8700, built August 2014, originally Windows 7 Pro, now fully updated Windows 10 Pro 64 bit, full repair install November 2022 so AFAIK no problematic "residuals" from Windows 7 except the "legacy" MBR/BIOS boot system. Seems to have been intended as gamer's machine, having 16 GB RAM and NVIDIA graphics card with 4 GB on-card graphics RAM.  Original HDD C: 1,000 GB, I added a second internal HDD D: (for "DATA") 4,000 GB, and moved the standard data folders to that drive.

Years ago I got tired of manual playing with Windows Update and figured out a way to let the computer update itself automatically, at midnight while I was asleep.  That includes manufacturers' drivers, but not generic Microsoft ones.  NVIDIA is a special case because of the profusion of available drivers; I most recently updated the NVIDIA driver two or three days ago.

The "how" is fairly simple.  I have the computer set to sleep after 30 minutes without user input (no mouse or key movements).  I have the "active hours" (Windows Update setting) set to 08:00 to 23:00.  I have a couple of necessary things enabled in the "Advanced Settings" menu (so long ago I can't remember whether or not some of those were default settings, though a couple I know definitely were not).  My system has been running like that for several years, through several Windows 10 version upgrades, with essentially no problems at all.  I have a little startup routine of things to check, taking about three seconds, to make sure all is in order; in the very rare event it is not, a simple restart will invariably fix things (I think twice in five years I actually had to restart twice).

Not content with that, I decided I wanted to know ahead of time when the system was going to sleep, so I could abort the sleep (and subsequent logging in) if I wanted to continue working.  Well, OK, the computer is going to sleep after 30 minutes, so set a screen saver to start 5 minutes before -- in other words, after 25 minutes of no input.  It's been running like that for maybe a couple of years.  I did start noticing that with that setting, the screen saver was actually running a lot longer than 5 minutes, longer in fact than I needed or wanted.  OK, set the screen saver to start after 28 minutes, 2 minutes before sleep.  I changed that setting a few days ago.  Screen saver still running more than 2 minutes before sleep -- and on closer examination, for a rather variable length of time.  I recorded three consecutive "no touch" sleep cycles with the screen saver setting at 28 minutes delay.  The screen saver ran for 6 minutes 45 seconds, for 5 minutes exactly, and for 11 minutes 40 seconds, in those 3 consecutive trials at the 28 minutes delay setting (sleep supposed to start at 30 minutes).

I do not need or want 5 minutes of screen saver -- 2 minutes is plenty -- never mind 12 minutes, or more than 30 minutes I observed on another occasion.  On the surface it looks as if the screen saver is delaying the sleep by a variable amount. So what I am asking here is, does anyone have an idea what is going her, or (even better) an idea how to "fix" it?

Once you start putting screen savers and junk into the mix you will sooner or later end up creating problems,  G' luck with it.

Modern screens do not suffer from burn in so screensavers are no longer necessary and hibernation is a less than useless setting that should be disabled.

Years ago I got tired of manual playing with Windows Update and figured out a way to let the computer update itself automatically, at midnight while I was asleep.  That includes manufacturers' drivers

 

Great idea for important Windows and Windows Defender updates but a massive no no for needlessly updating drivers.

Edited by PhillPower2, 12 August 2023 - 08:34 AM.

You can open an elevated command prompt, and run the powercfg command with the sleepstudy switch. It will usually tell you what is causing the PC not to sleep.

https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/powercfg-command-line-options#option_sleepstudy

In a classic command prompt, the command would be

powercfg.exe /sleepstudy /output "%userprofile%\sleepstudy.html"

In PowerShell

& powercfg.exe /sleepstudy /output "$env:userprofile\sleepstudy.html"

In either case, you will end up with a sleepstudy.html file in your C:\Users\YourName folder.

(Edited with additional details.)

Edited by pseymour, 12 August 2023 - 10:22 AM.

I am of course aware that screen savers are no longer necessary (I used to enjoy the AfterDark ones, in the now far distant past).  I am using one particular one, only, for the sole purpose of telling me when the computer is about to go to sleep.  The screen saver I use (3D Pipes, a "remastered" version of an AfterDark favorite, I think) does not prevent timed sleep, it just delays it by a variable amount.  I am trying to understand (and if possible control) that delay.

I do have to say I have found keeping Windows fully updated at all times to be the preferable approach.  I have not found it desirable to pick and choose among Windows updates.  The "preview" updates are NOT automatically installed.  FWIW I use MS Defender as my primary AV, with Malwarebytes Premium (also automatically updated) in secondary role.  

The way I have this set up, driver updates are very infrequent.  The automatic ones have never caused any difficulty.  I did once, following I think a BC suggestion, manually replace the driver for the Realtek chip in my USB-connected WiFi adapter.  That was a bust, it did not work at all.  I disconnected the adapter, deleted the new driver set, and plugged the adapter back in.  It promptly installed its own software and has worked perfectly ever since.  In general I'm a firm believer in "If it ain't broke  . . . ."

Thanks to pseymour for showing me something I didn't know about.  I'll try that and report back.

I do have to say I have found keeping Windows fully updated at all times to be the preferable approach. 

 

Important Windows and Windows Defender updates are the only two things that should be allowed to automatically update but no way should anything be allowed to automatically update drivers and this includes Windows being allowed to update drivers, I dont go foisting baseless opinions on people and will not chime in unless I can back up what I say and in this instance I will refer you to the following article The Only Safe Way to Update Your Hardware Drivers on Windows

 In general I'm a firm believer in "If it ain't broke  . . . ."

 

That approach goes without saying but somewhat contradictory to allowing your drivers to be needlessly updated, you will one day awaken to an unstable PC caused by one such an update.

Thanks very much for the Chris Hoffman article reference.  Definitely food for thought.  I do, however, get a somewhat mixed message from that article.  He is, to be sure, advising very strongly against the use of third party "driver updaters" -- but that message has long been loud and clear in the BC community.  I do not use any such software.

Out of curiosity I went back and looked at the driver updates in my system for the past 5 years.  A grand total of 15 driver updates, nearly all of them for Samsung software (I have a large Samsung monitor with its own software).  Lots of duplicates -- re-installation or updating of the same item number.  It boils down to half a dozen separate updates, mostly to the USB interface.  I use USB a lot, doing all my backups (image and folders) in duplicate on separate external HDDs, using Macrium Reflect to create the image files.  I've experienced absolutely no difficulties with that at any time.

None of those driver updates bore a Microsoft or Windows label.

It is true that for a good many years I earned my living as a software developer -- at first in the context of scientific research, and later also on behalf of governmental or corporate clients.  I do not, however, claim any "expertise" in the world of computers -- I am a computer user, not a computer scientist or any kind of "expert". Most especially do I wish to avoid arguing (or the appearance of arguing) about something of which my knowledge is obviously insufficient.  I think that perhaps what I am actually doing is not quite so dangerous, and that I have not accurately stated what it is and what it is not.  It is, in a way, something I tinkered together years ago, without realizing until afterward what I had done.  So at the first basic level, my question is how to understand and get control of the unexpected and variable timing delay introduced by the screen saver I am using.  Beyond that are further questions about possible ways to make the system better, and particularly safer.  And re: "If it ain't broke . . . " well, up to now it has given no sign of being broken.  Only not exactly as expected . . . which is not quite the same thing.

It is true that for a good many years I earned my living as a software developer -

 

Nothing that I can teach you then as I`m 95% nuts and bolts and only 5% software and the latter is primarily installing Windows and AVs and advising against driver updates and registry tweaking dangers.

I think we now have two things going on here -- first, my original question about the anomalous clock speed created, it seemed, by using a screen saver to announce the impending sleep of the system, and second the possible perils of over-eager driver updating, called to my attention by PhillPower2.  

pseymour has provided a trail I have not yet followed (too much other stuff on the plate, nothing to do with computers) but will come back to shortly.  In the meantime let me address the driver update question, to see whether or not what I am doing really represents an issue in this situation.  Below I have listed all the driver updates recorded by Windows Update, in chronological order of their occurrence.  For each event I have listed the source or manufacturer, the type of device, the "model number" of the update, and the date on which it was installed.

01    Surface      keyboard    1,0.104.0               09/23/2020

02    Surface      System       1.0.104.0               09/23/2020

03    Samsung   Modem       2.16.14.0                09/28/2020

04    Samsung   USB            2.16.14.0              10.08.2020

05    Intel            System      11.7.4.1045           10/22/2020

06    Samsung   Modem        2.17.6.0                02/08/2021

07    Samsung   USB             2.17.6.0                02/08/2021

08    Samsung   USB             2.17.15.0              09/09/2021

09    Intel            Display        20.19.15.4549      04/08/2022

10    Samsung    USB            2.17.16.0              05/30/2022

11    Samsung     USB            2,17.16.0              05/30/2922

12    Samsung     Modem       2.17.16.0              05/30/2022

13    Samsung     USB            2.19.1.0                11/10/2022

14.   Samsung     Modem        2.19.1.0                11/10/2022

15    Samsung     USB             2,19,1,0                11/14/2022

It looks to me as if these events may record the first initializations of external media -- 4 HDDs (all Western Digital) and a larger number of mostly SanDisk ":thumb drives" I use for storing and manipulating backup images, folders and files.  The dates are about right for when I first put those particular objects into service.  All backup images are made using Macrium Reflect and in duplicate on at least two physically separate media.  So far I have had no difficulty with any of the media.

I can't imagine to what the "Surface" in this tabulation refers.  I gather there is a tablet-like device called "Surface" but I have never owned nor used one.

Hi, would seem Microsoft thinks you are running a Surface.....  No searching will let me find a answer as am combining MS and Surface and get nothing but Surface.  Maybe someone else has a answer for that.

When doing a backup with anything non of your USB devices should be connected to the system.  You would backup these USB devices separately as individual devices.

All the listed Driver Updates are actually not needed.  It would be better if learned how to update your Modem yourself before MS bricks it and destroys access to your data on the USB Devices. (EDITED : to say your)

Edited by Pkshadow, 03 September 2023 - 09:58 PM.

When actually doing a backup, the relevant media device surely has to be connected to the computer.  The same when viewing or reviewing the backup result.  At all other times, the subject devices are physically disconnected from the computer.  In my corporate MIS days, one copy of each daily backup (Exabyte tapes, at that time) was immediately placed in a fireproof bank vault.  Can't  quite manage that degree of rigor as a handicapped private citizen without transport.

The "Surface" event was 3 years ago.  At that time I had a different physical location and a different Internet provider (Comcast) with a completely different hardware setup.  Since then I have also done a complete Repair Install of Windows 10 Pro.

I have lived at my present address only since 8th June 2022.  I have only one modem, property of Spectrum.  They have changed my modem at least four times, by physically removing one and installing a different modem.  My Internet signal enters my apartment via coax (standard TV cable, though I do not, by choice, have TV).  The coax is connected to the Spectrum modem (not a combo device).  From the modem an Ethernet cable goes to the router (my property, WiFi 5, AC1900 class).  The rest is all WiFi.  I have 3 Internet connection options.  The WiFi adapter in my old (August 2014) desktop computer is WiFi 4 class, and 2.4 GHz only. That appears on my screen as "WiFi" and gives me Internet WiFi at around 45 Mbps download speed.  Some months ago I installed a USB-connected WiFi 5 dual band adapter, which appears as "WiFi 2" on my screen.  That gives me a 2.4 GHz connection point at around 110 Mbps download, and a 5 GHz connection point at 300 Mbps.  I can choose any one of those 3 points -- they all work, without conflict or interference, but of course only one at a time.  If because of a faulty contact the "WiFi 2" adapter doesn't work, I can still choose "WiFI" (the original WiFi 4 adapter) and get a normal 2.4 GHz connection at the slower rate of 45 Mbps.  Still quite sufficient for most purposes.

The system does not allow conflicts.  if you choose one of the connection points, that one becomes active, and any other connection is instantly dropped.  Very convenient, actually.  The screen shows clearly which adapter, and if relevant which frequency band, is in use.  No ambiguity whatsoever.

Only the last 3 items of the list, dated 11/20/2022, apply to my existing hardware setup -- and there were half a dozen modem exchanges during the early days at this site.

I understand you to be saying none of the updates visible in that compilation are necessary.  Question arising: how does one decide which updates are "necessary" and which are not?  I thought I had learned that updates from the manufacturer were desirable, while generic ones from Microsoft were to be avoided.  Is that no longer a useful guideline?

Somewhere in this or a related thread I mentioned that occasionally the USB-connected WiFi 5 adapter does not respond, and that in that case I can restore the desired 5 GHz connection by restarting the computer. That almost always works without further action on my part.  Today I accidentally discovered that there's a much faster method to recover the 5 GHz connection -- unplug and replug the connection of the USB adapter, which is plugged into a USB 3.0 socket on the back of the computer tower.  Just pull and replug it, that restores the connection almost instantaneously..  So what does that suggest?  Oxide accumulation on the contacts at the USB 3.0 socket?

There are 4 USB 3.0 sockets on the back ot the tower (and 2 USB 2 sockets).  One of the 4 is empty, not currently in use.  One is connected to the WiFi 5 adapter, one to the UPS (for the software that controls it), and one has the dongle for the Microsoft cordless keyboard (that is a 2.4 GHz connection and would probably work just as well on a USB 2l.0 socket).  I have already had the WiFi 5 adapter plugged into at least 2 other sockets with no apparent difference.  There is another potentially pluggable connection between the adapter chip and the cable, and yet another between the chip and the WiFi receiving antenna.  This connection failure (if that is what it is) is a rare event, maybe a couple of times a month, so up to now I haven't given it much attention (after all, it can always be nearly instantly fixed!)

Troubleshooting a malfunction event that happens once or twice a month is just a wee bit tedious.  It requires more fastidious bookkeeping than I have so far wanted to invest.

Ack, end of 4th line up from bottom of the foregoing message should read USB 2.0

@pkshadow -- I think we have some kind of misunderstanding here.  There is only one modem in my system.  It is the property of my Spectrum ISP and I do nothing with it except (very rarely) physically connect or disconnect.  My Internet signal (from Spectrum) enters my apartment via coax cable (which also carries a TV signal I do not use).  In the first few weeks of my Spectrum contract they changed my modem several times, but the present one (just a modem for the Internet signal, not also a TV box) has not been changed in months and is entirely satisfactory as is.  I could connect it directly via Ethernet cable to the Ethernet adapter in my desktop computer, but in practice I connect it via Ethernet cable to my WiFi 5, AC 1900 class dual channel router.  .My desktop computer has a built-in WiFi 4, 2.4 GHz  only adapter.  I have added a USB-connected WiFi 5, dual channel adapter.  I thus have 3 ways to connect to the Internet by WiFi -- a 2.4 GHz connection point using the built-in WiFi 4, 2.4 GHz only adapter (identified by Windows 10 as "WiFi" and providing WiFi 4 level connection at 2.4 GHz and up to about 45 Mbps download speed), and 2 connection points using the newer WiFi 5 class dual channel adapter, identified by Windows 10 as "WiFi 2", connected to the computer via USB plug and receiving a 2.4 GHz channel at about 110 Mbps download speed, or a 5 GHz channel at about 300 Mbps download speed.  I can use any one of those 3 connection points, of course only one at any given time.  They are mutually exclusive and my computer screen always displays which one is in use.  Needless to say I have the 5 GHz connection set as default, but the others are there if I should want to use them.

When doing a backup I use external media of some kind -- most usually one of several 5 TB external HDDs, but for certain duplicates also high capacity USB "thumb drives".  Those objects must of course be connected to the computer while creating backups, and disconnected at any other time.  Each backup at least in duplicate on at least two physically separate devices, of course.  FWIW I use Macrium Reflect for most of my backup work.

-- and @PhillipPower2 -- I am still perplexed over how to decide which update is desirable and which is not.  Microsoft drivers, no.  Manufacturers' drivers, I thought OK, if not actually mandatory.  And then, are those "drivers" listed in my tabulation actually hardware drivers?  Or do some of those events reflect the initialization of a new, not previously used external device?

@pseymour -- I finally got around to running a couple of System Power Reports.  From those first attempts it is clear that the midnight wake-up occurs at the same precise time every day (within less than one second), but that timing of the end, after screen saver, is variable.  Well, we knew that.  To "filter out" the screen saver effect I will need to run a few reports with the screen saver disabled (easy enough to do, but need to let a few days' worth of data accumulate w/o screen saver).  

At present I have the screen saver set to start one minute before the sleep timer setting.  With those settings, in practice it's usually 4 or 5 minutes of screen saver before the computer goes to sleep.  I can live with that, of course, it just offends my scientist's sense of order.