No video after resuming from sleep - Windows 10 Anniversary Edition

[Deozaan]

Hi folks,

Shortly after I updated to the Anniversary Edition of Windows 10 Pro x64 I noticed that my primary display no longer gets any video signal after waking my PC from sleep.

I never had this issue before the update. Sometimes I can "fix" it (work around it) by remoting into my PC and logging in. Then the monitor gets a signal and things start working again. But today, when it happened again, even that workaround didn't work for me.

I haven't been able to find anything relevant in my searches for solutions. The closest thing I've found is people complaining about their lock/login screen being black after resuming from sleep. But that's just the background image they're talking about. Other UI elements and the mouse and everything still seems to work IIUC.

Does anyone have any solutions or troubleshooting suggestions?



More info that might be pertinent:

My GPU is an AMD Radeon HD 6870. I have a dual-monitor setup. The primary monitor just flashes the LED at me like it does when it is sleeping due to no signal. The secondary monitor appears to wake up properly, but just displays a black screen since it's normally black/blank on the lock screen anyway. If I move my mouse around, sometimes I see it on the secondary monitor, but it kind of comes and goes over the course of a few seconds.

[Ath]

Update your video-drivers, or installing a WHQL-qualified release should help, IMHO.

[Deozaan]

Update your video-drivers, or installing a WHQL-qualified release should help, IMHO.

It seems I'm already using the latest drivers. But the GPU is considered "Legacy" and drivers for it aren't being updated anymore.



But I've been running Windows 10 since it released a year ago and it's never had this problem until after the Anniversary Update, so I know it "should" work in Windows 10. Honestly it's probably about time to upgrade to a newer GPU anyway (the fan is starting to make a lot of noise) but it's not in the budget at the moment.

[4wd]

There are beta drivers for the HD6000 series (I'm running a 6850):



Beta is available here - if you have trouble getting it, I have the installer available (x64 only).

[Deozaan]

There are beta drivers for the HD6000 series (I'm running a 6850):

I didn't think I wanted the beta because:

1. It's a beta (implies unstable) and won't be updated further.
2. The download link says it is non-WHQL.

But I guess I'll give it a try and see what happens.

[Deozaan]

An aside:

Honestly it's probably about time to upgrade to a newer GPU anyway (the fan is starting to make a lot of noise) but it's not in the budget at the moment.

I'm also surprised that what I thought was a lower-to-mid-range GPU I got 5 years still ranks fairly high up there according to this chart, and spending the same amount of money today that I spent on it 5 years ago will get me a GPU that is only three tiers higher on that chart. It would move me from the 12th best tier to the 9th best tier.

I certainly expected the price to be lower or the performance to be much better by now.

[IainB]

From experience, the relative "newness" of the GPU drivers may be irrelevant. Sometimes the "latest" GPU drivers may not actually be "the best" for a given case. It is worth experimenting.

Again, from experience, the noisy fan thing may also be irrelevant. The word to use here is "maintenance". Try dismantling and cleaning the fan and heat exchanger, and especially cleaning/lubricating the fan spindle with CRC, or something that doesn't offend electrics - that can make a world of difference. Make it just like a new fan.

[Deozaan]

Again, from experience, the noisy fan thing may also be irrelevant. The word to use here is "maintenance". Try dismantling and cleaning the fan and heat exchanger, and especially cleaning/lubricating the fan spindle with CRC, or something that doesn't offend electrics - that can make a world of difference. Make it just like a new fan.

This is the 2nd fan on this GPU after the first one completely failed.

It was noisy, like it is now, so I took it apart and lubricated it with WD-40. It was nice and quiet for a while, but it wasn't long (as in a month or two) before the noise started up again. Eventually (months later) it ground to a halt and when I took it apart again I saw what appeared to be ruined ball-bearings.

You saying "something that doesn't offend electrics" caused me to do some research. I see now that WD-40 is bad for electronics and may have been what caused it to go from just noisy to completely broken. Oops!

Anyway, I contacted the manufacturer after the original fan failed and got a replacement from them (because it seems to be non-standard and I couldn't find a generic/replacement fan online anywhere). Once again it was nice and quiet for a while. But after only about 5 months it started making some noise again, which came and went for awhile. That was nearly 2 years ago and now the noise is pretty consistently there. I haven't attempted lubricating it since I figured it was just a faulty design (and saw other similar complaints in reviews for the product). But now that I know it was probably me who ruined the last one, I may spend a few bucks to get some proper lubrication for it and attempt it once more.

It sure would be nice to have a quiet GPU again. Well, quiet is a relative term, since by default it comes with two speed settings: Loud, and louder! (But I use SpeedFan to reduce the fan speed to make it quieter when it doesn't need to be running so fast.)

[Shades]

Failing fans on GPUs is quite common where I live. Subtropical ambient temperatures and often humid...that doesn't help with longevity of fans. That is to say, GPU fans. Case fans are much "sturdier" in my experience and if you don't care much about the looks of the GPU inside your case, then try to fit a case fan onto the heatsink of the GPU.

Fugly, but your GPU is getting lots of cooling and depending on the model case fan, it can be a very quiet solution. Even if you cannot mount a case fan directly, putting an extra case fan as close as possible to the GPU also works quite well. Especially if you can make some sort of funnel to pull the heat from the GPU. Case fans are much easier to replace or maintain and usually cost much less than a new GPU.

[IainB]

@Deozaan: Yes, using the wrong lubricant can apparently tend to wreck your nearby electrics and conductivity, though being somewhat cautious and having read of others' accidents in this regard, I personally have avoided making the same mistake.

I used to use CRC as that is a superb penetrating lubricant and it says it is good/OK for demoisturising car ignition contacts, but over the last few years I've been using something called inox-mx3 for things like laptop cooling fans as that says it's OK with electric motor windings, etc..
Laptop fans are beautifully simple bits of engineering, albeit a bit delicate, and in their normal operational position (upside down) they hang in the air by magnetic force attraction to their spindle and its base, and they are thus virtually frictionless. With occasional maintenance, they could last for ages. It's probably best not to let them run too long in a noisy or laptop-overheated state though - same probably goes for any PC (CPU/GPU) cooling fan.
I just wish laptop manufacturers would make them easier to get at for maintenance work.

Good idea to use SpeedFan too. Is very useful. I discovered a while back that SpeedFan has an option Info-->DIMM Info-->Read Info that can tell you all about your installed RAM - very informative. It apparently gets the info via the SMBus, but doesn't seem to work with all SMBuses - or maybe the RAM is connected to a different Bus in those cases, I don't know.

[Deozaan]

There are beta drivers for the HD6000 series (I'm running a 6850):

I didn't think I wanted the beta because:

1. It's a beta (implies unstable) and won't be updated further.
2. The download link says it is non-WHQL.

But I guess I'll give it a try and see what happens.

I installed the beta drivers and let my machine sleep overnight. It just woke up fine without a hitch. I'm cautiously optimistic that this issue is fixed, but we'll see how it handles sleeping a few more times before I say case closed.

[IainB]

Failing fans on GPUs is quite common where I live. Subtropical ambient temperatures and often humid...that doesn't help with longevity of fans. That is to say, GPU fans. Case fans are much "sturdier" in my experience and if you don't care much about the looks of the GPU inside your case, then try to fit a case fan onto the heatsink of the GPU.

Fugly, but your GPU is getting lots of cooling and depending on the model case fan, it can be a very quiet solution. Even if you cannot mount a case fan directly, putting an extra case fan as close as possible to the GPU also works quite well. Especially if you can make some sort of funnel to pull the heat from the GPU. Case fans are much easier to replace or maintain and usually cost much less than a new GPU.
___________________________

I just had a thought about that - overworked PC fans.
I'm not trying to teach my grandmother to suck eggs here, but I wondered whether some experiential tips/thoughts that have served me well might also be useful to others.

Humidity helps for cooling:

• In theory, pushing a flow of humid air over the object to be cooled will actually improve the cooling effect. An example of this in practice is as demonstrated by the use of canvas water bags you could buy in the US to place in front of a car's heat-exchanger/radiator (thus obstructing airflow!) when driving through a hot desert region. They worked surprisingly well.
  
• Similarly, when I was using an HP ENVY-14, which used to get a bit hot in summer (it used an Intel i7 chip in a small laptop enclosure), I would have a plastic tray or something on my lap, with a dampened cloth laid out flat and the laptop on top of that with an eraser lifting up either the back or the front (to enable good airflow). When I did this with an already-running and overheated laptop, the temps would drop noticeably within 30 seconds of setting this up, and stay there. Very effective. I wrote about this in a post on the DC forum.

The fans may be defeated from having any effect:

• This would be true, for example, if (say) the airflow of the fan exhaust through the heat exchanger grill was blocked with fluff/dust. This is typically what happens with laptop fans, because the interstices in the heat exchanger grill are relatively small and thus potentially more easily blocked by even small particles accumulating during usage (over time).
  
• I'm not sure whether this might be relevant to the same extent in a PC's heat exchanger (heatsink fins).

The waste heat may not be getting to the heat exchanger:

• If the CPU or GPU in question was apparently working fine and running sufficiently cool when it was new, but now is overheating and failing to work properly, then, if the airflow over the heat exchanger (typically a finned heatsink) is unobstructed, it may be just a symptom of overheating because the path for conducting heat away from the chip is broken.
• The heatsink - usually a piece of copper in a laptop, not sure what it is in a PC - is touching the chip, with a special heat-conducting "thermal" grease sandwiched between the face of the chip and the face of the heatsink. The grease makes up for any uneveness of the two surface areas in contact, maximising the potential for heat transfer.
  
• A few years ago I was using a Toshiba Satellite A100 laptop that started to progressively overheat, the fan running a lot, etc. I did the obvious - cleaned out the heat exchanger - which only improved things marginally. So I bought some of the thermal grease, removed the heatsink, cleaned the faces of the chips and heatsink where they met, smeared on a dab of the thermal grease to the chip faces, and bolted it all back together, taking care not to over-tighten things.
  Problem solved. Laptop ran just like new and with no more "over-heating" problems. So I think that what had happened was that the grease had deteriorated in its quality of thermal conductivity, and thus had ceased to perform its job. What tipped me off to this fixit was my having read about overheating fixes on a forum that specialised in graphic chip technology and drivers. I forget it's name, but I might be able to dig it up from the archives if anybody wants it. At the time, I recall that I was using a GPU controller app that they had produced, which performed better than the GPU manufacturer's (AMD's) own software!

SpeedFan is very useful, but...:

• The thing about SpeedFan is that it will look for and try to access the various heat sensors in the CPU/GPU/HDD chips. It cannot therefore "see" any problems downstream in the heat-flow from the chip itself, so one would probably tend to intuitively assume that fan symptoms of overheating would usually mean that the fan was the problem. The fan is turned on/off by the sensors. Similarly, the operation of the CPU/GPU chips will be automatically suspended when these sensor temps reach certain critical thresholds. This may be what has happened in this case. Repeated/excessive overheating of chips can apparently permanently damage them. I don't know whether this will have happened in this case.

Conclusion:
So what I would suggest is, if you haven't done so already, then replace the thermal grease on those chips (or whatever they use in the PC hardware heatsink interfaces), as described above. This will remove one possible cause. After that, if the airflows are clean, the fans are in good order, yet the chips don't perform like they should, then - whatever the cause - they are probably permanently damaged and they will need to be replaced. To test this hypothesis, try replacing them with any old (working) spare you might have access to, and see if that eliminates the problem.
The fan systems will also need to be inspected/maintained, as they will have probably been operating outside of their normal performance range for extended intervals. This may have caused some mechanical wear and tear.

[Deozaan]

The beta GPU drivers were causing weird issues in some videos:






So I went back to the non-beta drivers today.

I'm about go put my machine to sleep for the night. We'll see if it goes back to having problems waking up the display after sleep. Wish me luck!

[Deozaan]

I'm about go put my machine to sleep for the night. We'll see if it goes back to having problems waking up the display after sleep. Wish me luck!

First day after sleep with the old drivers installed: It woke up from sleep properly! Hopefully the problem is resolved for good.

[Deozaan]

Today my PC didn't wake up from sleep properly.

[4wd]

I've let mine install the Anniversary Update, this is the driver I'm using:



And it resumes from Sleep fine, I'm using HD6850 with HDMI to an AOC I2470Sw and DisplayPort->DVI to a Dell 1701P.

Don't think it'll make a difference but FWIW I'm running an AMD chipset.

EDIT: Spoke too soon, I just had it happen for the first time even though the Anniversary Update was done a week ago.

You'll probably see that the Event Log is full of this:


It appears it can happen even with a reasonably recent card: Event 4101 Display driver amdkmdap stopped responding and has successfully recovered.

More

From here:

Warning Event ID 4101


    Hi TenForumers.
    Few days ago I found some strange warnings concerning the display driver (event ID 4101):
    "Display driver amdkmdap stopped responding and has successfully recovered".
    After trying this and that with no avail, I noticed these warning were raised when I was using my Internet browser (Firefox):
    sometimes my browsing experience suffered of occasional "blocks" of 1-2 secs (with a Warning 4101 coming out in my event viewer).
    I turned off hardware acceleration on my browser's settings page and I eventually got rid of those warnings.

    I'm writing this post hoping it could be useful to someone with the same problem.

I did have Firefox running at the time the machine went into Sleep so I'll disable HW Acceleration and see what happens.

[Deozaan]

I've also noticed that maybe it has to do with the length of time the computer is sleeping. If I put it to sleep and wake it up immediately, it seems to work just fine in my tests. Sometimes when I put it to sleep and wake it up within about 8-9 hours it works just fine. But if I let it go closer to 12 hours before waking it up again, it seems to fail much more often.

That said, I have never properly logged any of the times or paid close enough attention to say for certain that this pattern is true. It's just what it "feels" or "seems" like to me when I experience it, because sometimes it works fine and other times it fails.

I've just started either leaving it on all night long or shutting it down completely at night, so I don't lose any work in the morning when I attempt to wake it from sleep.

[4wd]

Well done Microsoft, second time in less than a week since I installed the Anniversary Update after 12 months of not a single occurance since installing Windows 10.

So, hope for the best, plan for the worst and all that - I've restored the system image I created before I let the Anniversary Update happen.

They can take their Anniversary Update and stick it in a dark place. 

Oddly enough, I had a look at the restored System Event Log, not a single Event 4101 to be seen.


And a FYI re. the times my computer sleeps, almost always less than one hour, (over meal times longer than 20 minutes).  System is shut down overnight unless there is a video encode happening.

[Deozaan]

I was surprised to receive a notification a day or two ago saying AMD had a driver update for me. I thought, "Wow! They updated their old, discontinued driver to support the Anniversary Update! Nice!"

I let it run its thing and then it told me it was running the latest version after all.

[4wd]

That's because you installed the beta driver, the newer Install Manager for the Radeon Crimson software checks for updates but unfortunately it doesn't seem to take into account what graphics hardware you're running before it does.

I completely removed all AMD software and then reinstalled the latest applicable.

Edit: Had the wrong program referred.

BTW, one thing I also noticed with the Anniversary Update, Windows Defender would no longer do scans in the background, ie. you couldn't close the window as it would say it would cancel the scan ... there's a backward step.

[Stoic Joker]

BTW, one thing I also noticed with the Anniversary Update, Windows Defender would no longer do scans in the background, ie. you couldn't close the window as it would say it would cancel the scan ... there's a backward step.

I ran into that the other day too ... Really quite annoying.

[wraith808]

Mine won't update definitions anymore automatically.  I have to download it and install the definitions.  Definitely annoying.

[4wd]

Mine won't update definitions anymore automatically.  I have to download it and install the definitions.  Definitely annoying.

I created a scheduled task that updates them every hour, got tired of it telling me the definitions were out of date.

I'll export and attach later when I'm officially awake.

Here you go, you might need to edit (Author, User ID, etc) but all the info is there:

Code: ( xml)

<?xml version="1.0" encoding="UTF-16"?>
<Task version="1.4" xmlns="http://schemas.microsoft.com/windows/2004/02/mit/task">
  <RegistrationInfo>
    <Date>2015-08-13T19:55:24.4416143</Date>
    <Author>8x8x64\4wd</Author>
    <URI>\Microsoft\Windows\Windows Defender\Windows Defender Definition Update</URI>
  </RegistrationInfo>
  <Triggers>
    <CalendarTrigger>
      <Repetition>
        <Interval>PT1H</Interval>
        <StopAtDurationEnd>false</StopAtDurationEnd>
      </Repetition>
      <StartBoundary>2015-08-13T09:50:00</StartBoundary>
      <Enabled>true</Enabled>
      <ScheduleByDay>
        <DaysInterval>1</DaysInterval>
      </ScheduleByDay>
    </CalendarTrigger>
  </Triggers>
  <Principals>
    <Principal id="Author">
      <UserId>S-1-5-18</UserId>
      <RunLevel>HighestAvailable</RunLevel>
    </Principal>
  </Principals>
  <Settings>
    <MultipleInstancesPolicy>IgnoreNew</MultipleInstancesPolicy>
    <DisallowStartIfOnBatteries>true</DisallowStartIfOnBatteries>
    <StopIfGoingOnBatteries>true</StopIfGoingOnBatteries>
    <AllowHardTerminate>true</AllowHardTerminate>
    <StartWhenAvailable>true</StartWhenAvailable>
    <RunOnlyIfNetworkAvailable>true</RunOnlyIfNetworkAvailable>
    <IdleSettings>
      <StopOnIdleEnd>true</StopOnIdleEnd>
      <RestartOnIdle>false</RestartOnIdle>
    </IdleSettings>
    <AllowStartOnDemand>true</AllowStartOnDemand>
    <Enabled>true</Enabled>
    <Hidden>false</Hidden>
    <RunOnlyIfIdle>false</RunOnlyIfIdle>
    <DisallowStartOnRemoteAppSession>false</DisallowStartOnRemoteAppSession>
    <UseUnifiedSchedulingEngine>false</UseUnifiedSchedulingEngine>
    <WakeToRun>false</WakeToRun>
    <ExecutionTimeLimit>PT72H</ExecutionTimeLimit>
    <Priority>7</Priority>
  </Settings>
  <Actions Context="Author">
    <Exec>
      <Command>"C:\Program Files\Windows Defender\MpCmdRun.exe"</Command>
      <Arguments>-SignatureUpdate -MMPC</Arguments>
    </Exec>
  </Actions>
</Task>

[4wd]

Shortly after I updated to the Anniversary Edition of Windows 10 Pro x64 I noticed that my primary display no longer gets any video signal after waking my PC from sleep.

Did you ever manage to find out what was the reason or did it come good in a later update, Deo?

[Deozaan]

Did you ever manage to find out what was the reason or did it come good in a later update, Deo?

I had stopped putting my PC to sleep and instead either left it running or shut it down completely to avoid the issue.

Eventually I forgot that I ever had this problem and probably around the new year I began putting my PC to sleep at night again and the problem hasn't happened yet.


But as of about mid-December, my other monitor has started having issues when I turn on my PC, though I'm fairly confident it's unrelated. My screen is a pale grey, then looks like "clouds" form on it, then there are either some bright or dark lines of pixels that show up at the top, soon replaced by several lines of glitchy pixels at the top of the screen which gradually expand down toward the bottom until finally the picture comes back, but the colors are messed up. Once it reaches that point, I can shut the monitor off and turn it on again and the colors are all good. But if I try to restart the monitor earlier in the process, it continues about from where it left off. It's like it needs to "warm up" a bit first. It has happened both when resuming from sleep and also when powering on from an off state.

I took a video of it using my tablet. It seems to be worse and last longer since I took this video:

https://www.youtube.com/watch?v=6aLuJqStU1g