Windows 10: Excessive battery drain during S0 sleep (modern standby)

For my Lenovo X1 Yoga 3rd Gen (Win 10 20H2 Build 19042.928) I have had issues using hibernation. About 30% of the time, the WiFi device will not work when resuming from hibernation. The device will not even appear in device manager. A reboot is required to make it appear and function again. I did lots of searching regarding this problem and I found that if I update to the latest BIOS (1.41) then for "Sleep Mode" in the BIOS I can select "Linux" instead of "Win 10" and this will make Sleep Mode S3 available. This worked great and I would use sleep instead of hibernate. But then I noticed that if the sleep was initiated by closing the lid of the computer, the touch screen would not work (disappears from device manager) when resuming from the sleep. This can be fixed by disabling the Thunderbolt port (but I need it for an ID Token). The touch screen works fine if the sleep was initiated in the Start menu or with the power button. But the problem is that if I put the computer to sleep with the power button and it enters sleep, the computer will immediately wake when I later close the lid and I have been unable to figure out why this happens. So after all of that, I decided to try to get Modern Standby (S0) to work for me. I set the Network to be disconnected when in the S0 sleep state so now a powercfg /a shows these as the available sleep states: Standby (S0 Low Power Idle) Network Disconnected Hibernate I did this so that I could get as low power as possible, and I have no need for the computer to do anything while sleeping (which is why I prefer S3). I started with a fully charged computer and closed the lid. 11.75 hours later I opened the lid and the battery remaining was around 28%. So that is about 6.1% per hour discharge rate, which is quite excessive. I did powercfg /sleepstudy and it shows Mobile Intel(R) Processor Family I/O PCI Express Root Port #9 - 9D18 (\_SB.PCI0.RP09) 100% active during that time. I look in device manager and it reports the device is working properly, there are no conflicts, the version is 10.1.1.45 with a date of 7/18/1968 (odd date!), and provider is INTEL. When I look on the Lenovo support page for my computer, the chipset driver version is shown as 10.1.18228.8176 (12/25/2019) and if you look in the ReadMe it says 10.1.1.45 (1/30/2018) is two versions back. So I am not sure why the Lenovo auto updater isn't installing this. I am not even sure what a PCI Express Root Port #9 is, so I am really at a loss to determine why it won't go to a low power state. I have performed all of the updates using the Lenovo support page auto update and all Windows updates have been performed. One thing I noticed is that for this device in device manager, when I go to the drive tab and select driver details, it lists several files: C:\windows\system32\DRIVERS\nipbcfk.sys (National Instruments 3.0.0f0) C:\windows\system32\DRIVERS\nipcibrd.sys (National Instruments 18.5.0f0) C:\windows\system32\DRIVERS\nipxibrc.sys (National Instruments 17.3.0f0) C:\windows\system32\DRIVERS\nischifk.sys (National Instruments 2.0.0.0f0) C:\windows\system32\DRIVERS\pci.sys (Microsoft, 10.0.19041.1) All but first one lists Microsoft Windows Hardware Compatibility as the Digital Signer. This laptop has a fair amount of National Instruments (LabView, etc) installed on it, so maybe that is what is causing the high power usage during sleep. I am considering uninstalling all of that. I really don't understand why so many driver files are listed in "Driver File Details", and why none of them are version 10.1.1.45 (the version reported in the Driver tab. Thanks for any suggestions. - - - Updated - - - I updated the chipset driver to the version 10.1.18228.8176 found on the Lenovo website (Device Manager now shows it as 10.1.5.2 for the Mobile Intel(R) Processor Family I/O PCI Express Root Port #9 - 9D18 device). This time I started with the battery at 100% and closed the lid for 3 hours and the battery lost just over 30%. So that is over 10% per hour. The report has a 1 hour session where the computer was active (not in sleep or hibernate) and it only used 14% in an hour.

:)

 

Excessive battery drain during S0 sleep (modern standby)

S0 Modern standby is not sleep but a glorified idle condition. That's why it uses so much power - the computer is on during S0.

Denis

 

Excessive battery drain during S0 sleep (modern standby)

For my Lenovo X1 Yoga 3rd Gen (Win 10 20H2 Build 19042.928) I have had issues using hibernation. About 30% of the time, the WiFi device will not work when resuming from hibernation. The device will not even appear in device manager. A reboot is required to make it appear and function again. I did lots of searching regarding this problem and I found that if I update to the latest BIOS (1.41) then for "Sleep Mode" in the BIOS I can select "Linux" instead of "Win 10" and this will make Sleep Mode S3 available. This worked great and I would use sleep instead of hibernate. But then I noticed that if the sleep was initiated by closing the lid of the computer, the touch screen would not work (disappears from device manager) when resuming from the sleep. This can be fixed by disabling the Thunderbolt port (but I need it for an ID Token). The touch screen works fine if the sleep was initiated in the Start menu or with the power button. But the problem is that if I put the computer to sleep with the power button and it enters sleep, the computer will immediately wake when I later close the lid and I have been unable to figure out why this happens.

So after all of that, I decided to try to get Modern Standby (S0) to work for me. I set the Network to be disconnected when in the S0 sleep state so now a powercfg /a shows these as the available sleep states:

Standby (S0 Low Power Idle) Network Disconnected
Hibernate

I did this so that I could get as low power as possible, and I have no need for the computer to do anything while sleeping (which is why I prefer S3). I started with a fully charged computer and closed the lid. 11.75 hours later I opened the lid and the battery remaining was around 28%. So that is about 6.1% per hour discharge rate, which is quite excessive. I did powercfg /sleepstudy and it shows Mobile Intel(R) Processor Family I/O PCI Express Root Port #9 - 9D18 (\_SB.PCI0.RP09) 100% active during that time.





I look in device manager and it reports the device is working properly, there are no conflicts, the version is 10.1.1.45 with a date of 7/18/1968 (odd date!), and provider is INTEL. When I look on the Lenovo support page for my computer, the chipset driver version is shown as 10.1.18228.8176 (12/25/2019) and if you look in the ReadMe it says 10.1.1.45 (1/30/2018) is two versions back. So I am not sure why the Lenovo auto updater isn't installing this. I am not even sure what a PCI Express Root Port #9 is, so I am really at a loss to determine why it won't go to a low power state.

I have performed all of the updates using the Lenovo support page auto update and all Windows updates have been performed. One thing I noticed is that for this device in device manager, when I go to the drive tab and select driver details, it lists several files:

C:\windows\system32\DRIVERS\nipbcfk.sys (National Instruments 3.0.0f0)
C:\windows\system32\DRIVERS\nipcibrd.sys (National Instruments 18.5.0f0)
C:\windows\system32\DRIVERS\nipxibrc.sys (National Instruments 17.3.0f0)
C:\windows\system32\DRIVERS\nischifk.sys (National Instruments 2.0.0.0f0)
C:\windows\system32\DRIVERS\pci.sys (Microsoft, 10.0.19041.1)

All but first one lists Microsoft Windows Hardware Compatibility as the Digital Signer. This laptop has a fair amount of National Instruments (LabView, etc) installed on it, so maybe that is what is causing the high power usage during sleep. I am considering uninstalling all of that. I really don't understand why so many driver files are listed in "Driver File Details", and why none of them are version 10.1.1.45 (the version reported in the Driver tab.

Thanks for any suggestions.

- - - Updated - - -

I updated the chipset driver to the version 10.1.18228.8176 found on the Lenovo website (Device Manager now shows it as 10.1.5.2 for the Mobile Intel(R) Processor Family I/O PCI Express Root Port #9 - 9D18 device). This time I started with the battery at 100% and closed the lid for 3 hours and the battery lost just over 30%. So that is over 10% per hour. The report has a 1 hour session where the computer was active (not in sleep or hibernate) and it only used 14% in an hour.

 

Excessive battery drain during S0 sleep (modern standby)

Yes, that is true, but I have heard of people getting decent drain rates in S0. I uninstalled all of the National Instruments software and also found that the Thunderbolt port (in Device Manager) had "allow this device to wake the computer" checked so I unchecked that. One of these seemed to solve the problem I had of it remaining awake during sleep.

During a previous period (posted above) it showed it active 100% of the time during sleep and it consumed an average of 2743mW (6.1% battery life per hour). I updated the drivers and did another sleep test and it consumed an average of 4570mW (10% battery life per hour). Both showed that device active 100% of the time during sleep. After the changes with the Thunderbolt port and removing the NI software, all of the "top 5 offenders" were green and showed 0% active time. Also it showed %Low Power State Time of SW:100% and HW: 98%. However the average power consumption was 1848mW (4% battery life per hour). I thought I had read that you could get better than this, but if everything is in a low power state and the network is disconnected, I am not sure I can get better.





I am not happy with this amount of battery usage during sleep, but like you said, it is still on. Oddly enough when I had S3 sleep state enabled, the power consumption wasn't drastically lower than this. Maybe 2% per hour. I saw someone with the same laptop and driver version attain 327mW which is nearly 6x better than what I am seeing. I would be perfectly happy with that value. My sleepstudy shows that nothing is active, so I am not sure what is using all of the power.

I don't need instant on and would prefer to hibernate, but unfortunately the WiFi doesn't always work after resuming from that state.