Hybrid-sleep, Win7 trying to solve a strange problem.

I recently changed to a larger backup battery and decided to give sleep mode another try.

1. Sleep/resume works perfectly.
2. Hibernate/resume works perfectly.

When I say they work independently and perfectly I mean each can resume day in and day out perfectly! So there are NO hardware/SW issues with USB, drivers, software, video or the SSD.

If I enable Hybrid sleep the PC will resume from sleep in 2 seconds perfectly – every time.

My reasons for wanting to use this mode is to allow for a quick resume from hibernate if +5v standby power fails.

Now when using hibernate alone you can do ANYTHING to +5v SB. You can remove USB devices when powered down, you can remove all power and reconnect and the PC will resume from Hibernate perfectly, as it should, every time.

Again:
It NEVER fails to resume from hibernate.
It NEVER fails to resume from sleep.

If “Hybrid Sleep” is enabled and I remove +5v SB, the PC will correctly try and resume from the hibernate file, however it fails every time right at the end of resume with a black screen and blinking cursor in the top LH corner.

Windows MUST be doing something different when it writes the hibernate file in Hybrid sleep as opposed to Hybrid alone.

Anyone had/having a similar problem?

Have you seen this thread?

http://forums.tweaktown.com/gigabyte...ower-loss.html

I know it's not your exact motherboard, but maybe updating your BIOS and/or disabling USB3 will help.

Thanks for the link, all interesting info to keep in mind. There are a number of devices that I indentified as causing problems with sleep “only”, these are automatically shut down and restarted by my FE before suspending and when resuming.

However I don’t have a problem with sleep or hibernate.

The problem is most of the posters never test Hibernate separately – the problem will be that hibernate does not work with USB3 enabled or driver loaded. It’s confusing when they lump “Hybrid sleep” as the problem when in reality it appears that Hibernate does not work on that model of MB with USB3 and other setting enabled without a bios update. – Unless I missed someone who had tried it alone.

However, when I eventually reach the cut-out point of my standby battery (just seeing how long it lasts in SB with 380mA SB current) I will disable USB3 just for a try. I don’t think I have a HW/SW issue though. Latest Bios and I can hibernate & resume a 100 times a day perfectly.

Worth a try though!

You are at 380ma in sleep? I just measured mine last night, my PC only (and GPS) runs 190ma, with another 70ma if I leave all my devices plugged in to the USB hub (keyboard, touch-screen, Bluetooth). Not included yet is the ODB-II device as I'm running on the bench.

I'm trying to decide if I want to drop out power to the USB hub or not when it's asleep. It's wasted power, but if it stays on I'll probably have less issues with things not restarting properly.

I take it you've tried your tests with ALL usb devices unplugged?

Nevermind- I'm most of the way though your build thread. Wow.....
Excellent work, and you've given me a lot of ideas. Thanks for documenting everything, it is very much appreciated.

There are quite a few people that have issues when resuming from hybrid sleep when power has been disconnected. Some people see wait times of over 5 minutes when a normal hibernation resume takes seconds. I never use hybrid sleep for that very reason. My E350M1 resumes from hibernation in about 5 seconds (4GB RAM, hibernation file set at 100%, SSD, Windows 7 Ultimate x86, CF as my front-end, with GPS and OBD-II on a RAM tough hub), which is plenty quick. My previous board, a D510MO, took a bit longer to resume, but I didn't mind the wait because it was rock solid and didn't have any problems.

My personal opinion is that some motherboards don't handle hybrid sleep very well after a power loss.

Hi Guys, thanks for the posts. I had this written and was about to post the following when I saw the last two posts. I think my ramblings still apply so here they are anyway.

Zephrant, thanks for the words and you ideas – everything is appreciated with me.

kegobeer, mine resumes from sleep in under 2 seconds, had to use a stop watch as it’s just ridiculously fast. Dam MB takes 12 seconds to do its post on a cold boot (everything optimised and makes no difference with none/all USB devices plugged in. Resumes from Hibernate in around 25 seconds (dam post time).


I have three 4-port USB hubs and 10 USB ports on the PC, along with 2 USB web cameras, GPS mouse, and 4 Interface Micro controllers, optical DSP USB device, touch-screen, mini Mouse, mini Keyboard, OBD, USB3 capture drive, flash music drive, M4-ATX interface and USB-Rs232 lead for OEM factory diagnostic system module, out of 22 ports I have 2 free on one hub and 1 at the PC.

Two of the hubs have 8 LEDS so I’m going to pull them apart and remove these, not sure how much difference this will make. So 380ma is not bad but I think I can get it better

I had the PC in sleep for 24 hours and the AUX battery was reading 12.2 volts, turned the IGN on and the thing resumed from sleep and was playing music before I could finish cranking the car. Prior to that it had been started every 20 minutes all day long and driven 3 times for about 10 minutes, then cranked over another 10 times at different states of suspend and resume throughout the day and has not failed once, and that’s with the M4-PSU.

I’m really surprised just how well this “physical small” 26AH AUX battery is handling this – In conjunction with the Main battery (diode Isolator) which means the Main Battery and the Aux Battery see-saw down until the cut-out point of the main battery isolator drops and then the AUX is on its own. So It’s going to last a lot longer than 36 hours. As I write this it’s up to 30 hours and still above 12v.

As far as dropping USB off in sleep, I think it may fail – at least mine seems very touchy to any disturbance or change to USB while in sleep.

This PC will resume from sleep 100 times a day. It will resume from Hibernate 100 times a day.
It will not resume from Hibernate ONLY if HYB-sleep is invoked and only if sleep fails (because STB power is removal.) I cannot see why the Hibernate file should be any different in this mode. I have heard rumblings that HYB-Sleep is broken in some situations?

I’m tempted to Hibernate the PC, remove the SDD and copy the Hibernate file. Put it into HYB-Sleep and when in sleep, do the same and copy the saved Hibernate file back and see if it resumes from Hibernate correctly. (Naturally it will fail to resume from sleep.) Or try and compare the files.

It’s now so solid running in sleep and lasting so long that I don’t really care if I leave HYB-Sleep off and just let it do a cold boot when the batt can no longer support +5v STB in sleep mode.

I would like to know how you go with powering down hubs in sleep if you go down that track. Might be something I will look at again if there is a way to do it without killing resume from sleep, could be PSU or MB dependant as well?

I ran a few quick tests last night- I hybrid-sleeped the CPU, which leaves the USB +5V up.
I have these items on the hub right now: Andrea USB, keyboard, Bluetooth. In the car it will also have the touchscreen and ODB-II, and maybe the GPS.
Test 1: I unplugged the hub USB cable from the the PC. The PC woke up. I plugged it back in, and things settled down and worked fine.
Test 2-6: I unplugged the USB cable from the hub, then unplugged the power to the hub: The system came back up OK each time I tried it on the bench, but I don't take that as any indication that it'll work every time in the car. : )

For breaking power to the hubs- Since the hub pulls power from the USB line and/or the 5V input, I'd have to cut the 5V line in the USB cable, then relay switch the 5V input to it. Since the M4-ATX turns off the 5V and 12V lines when it goes to sleep, I was thinking of putting a relay on the 12V line that broke power to the hub. Then the hub turns off only after the CPU is done writing the hibernate file, and is turned on before the CPU comes out of reset.
Alternatively, I'm thinking of running the hub off of 5V stolen off of the pins of the M4-ATX (just not the 5VSB). Then it will automatically go up/down at the correct times with no relay needed. I'm pretty sure I have enough power to do so on my system.

I'll try to give that a try this weekend.

Good results. My main concern is that this PC resumes from sleep in about 1.8 seconds. All running software that is accessing ports and USB devices would likely fail as some of these USB devices take a finite time to start, Vehicle and Climate control interface Microprocessors immediately start streaming status data to the Ports the moment of IGN on. Again I guess HW/SW related in my case.

I found that breaking and reconnecting the +5 line in the USB lead itself can cause some Hubs and devices to lock due to corruption of the USB interface in the Hubs and devices, once again maybe related to the Hubs/HW.

Still, it’s worth further investigation and interested to see if you get a working solution.

The M4 has a lot of power for this kind of thing, I power my hubs from the +5 M4 rail. I also backup up power the Micro interfaces with +12v from the M4 during the sleep/hibernate shutdown cycle when the IGN is normally off and during cranking once the M4 has started, this is to stop any chance of data corruption. These changes along with other mods have now made sleep rock solid.

Have you considered holding the PC in reset after IGN-on for a few more seconds if needed to allow all the USB devices to power up?