Sleep and Wake under S1, S3 and S1&S3 problem -- related to PSU?

[BonzaiDuck]

Posting this also under "Motherboards."


As per myriad recent posts and threads I've made, I'm trying to build an absolutely stellar WHS-2011 server box. I discovered that I had in storage a (practically brand-new) ASUS Striker Extreme 680i board and a (practically brand-new) Conroe E6600. I also seem to have plenty of DDR2 RAM kits. So -- "I built it!" [Try and make a joke about "You didn't build that!" -- but I did.]


680i was nVidia chipset, and ASUS stopped producing driver updates of any significance for it. Nothing there on their download site for either Win 7 or Win-2008-R2 (core for WHS2011). Have to rely on Microsoft's own chipset or nForce drivers.


So I find I'm having trouble implementing either S3 or S1&S3 hybrid sleep.
Systems will go to sleep, but they fail to wake up.


But I've got a 780i system with Win 7 installed -- also relied on the MS chipset driver versions. Hybrid sleep works great on that system. Wake-on-LAN works great, too.


And somewhere today -- I found a trouble-shooting page on a Win 7 forum, which suggests that one (of many) sources of the problem could be the PSU.


Currently all the systems in the house which wake and sleep properly are fitted with Seasonic 650 and 750 PSUs. The ones causing trouble have an older OCZ PowerStream and Antec NeoPower, respectively. Those PSUs are older than the Seasonics.


Is there any empirical experience among you [member-readers] that would sustantiate my suspicions? I have a spare Seasonic PSU -- fairly recent and brand new.

 

[BonzaiDuck]

OK . . . I see several -- maybe a hundred people -- looked at this thread, but no responses.

I found this at Tom's HW:

http://www.tomshardware.com/forum/334784-31-motherboard-fail

I'm going to take the time and trouble to replace the old Antec NeoPower with a new 650W Seasonic. The Tom's forum poster RAIDDinn seems so adamant about this as a possible problem, I'll have to eliminate it before anything else. Hell -- I'll even report back on it.

 

[mikeymikec]

I once had a problem with two PCs (same model board) waking out of S3 sleep - about 50% of the time, trying to wake them (keyboard, power button or mouse) would do almost nothing, a click but then the light carried on flashing. It turned out to be both CX430 PSUs. As soon as I swapped them out for other models, no further problems.

- edit - the CX430 and that board worked with S1 sleep though.

 

[Kodongo]

BonzaiDuck, you have no idea how much this thread relates to me.

I have an OCZ ModXStream 700W, released in 2008 I think, and this power supply refuses to wake up from sleep with any kind of significant overclock.

If I take my 2600k above 4.0GHz then the sleep function ceases to work.

I too have ordered a Seasonic X-660 (the European equivalent) from Amazon UK and am planning to retire the OCZ from service when the X arrives.

With regards to your system which I assume is overclocked, have you tried going back to stock timings/settings for everything?

That is the only way to get my system to wake from sleep. For me, having sleep work properly is not worth giving up 4.4GHz on the 2600k and OCed memory so I just manage the computer by shutting down instead of sleeping.

Hopefully my new X-660 will remedy my problem.

 

[BonzaiDuck]

I once had a problem with two PCs (same model board) waking out of S3 sleep - about 50% of the time, trying to wake them (keyboard, power button or mouse) would do almost nothing, a click but then the light carried on flashing. It turned out to be both CX430 PSUs. As soon as I swapped them out for other models, no further problems.

- edit - the CX430 and that board worked with S1 sleep though.

Yeah . . . . I just didn't want to believe that a system with stock settings and this Antec Neopower would have drawbacks. But the Antec PSU had been in my brother's system for at least the last five years.

I swapped in a Seasonic 650W unit I got last week from the Egg with something like a two-part 50% markdown.

Like the sargeant says in "Private Ryan" after the beach misery subsides: "We're in Bid-nis!! Def-i-nitely!!"

It sleeps; it wakes; it walks; it talks. I'm happier than a pig . . . in . . . . s***!

 

[BonzaiDuck]

BonzaiDuck, you have no idea how much this thread relates to me.

I have an OCZ ModXStream 700W, released in 2008 I think, and this power supply refuses to wake up from sleep with any kind of significant overclock.

If I take my 2600k above 4.0GHz then the sleep function ceases to work.

I too have ordered a Seasonic X-660 (the European equivalent) from Amazon UK and am planning to retire the OCZ from service when the X arrives.

With regards to your system which I assume is overclocked, have you tried going back to stock timings/settings for everything?

That is the only way to get my system to wake from sleep. For me, having sleep work properly is not worth giving up 4.4GHz on the 2600k and OCed memory so I just manage the computer by shutting down instead of sleeping.

Hopefully my new X-660 will remedy my problem.

Just caught your post. The OTHER machine I have which will only function to wake from an S1 state (versus S3 or S1&S3-hybrid) -- has an OCZ 550W PowerStream. It cost me a pretty penny . . . back in 2005. Sucker is 7 years old! How could I not suspect something like this?

Point is, though. None of the sleep/wake-enabled systems is over-clocked. My main workstation is a 2600K @ 4.60 Ghz and PLL Overvoltage Enabled. That one -- I'm not even going to worry about now.

But if you want to enable sleep mode and have it function properly -- check first with stock settings, and suspect an old PSU when things aren't working the way they should.

 

[VirtualLarry]

Very interesting. I would have never guessed that the PSU would prevent waking from sleep.

Not enough +5vSB current, perhaps?

 

[BonzaiDuck]

Very interesting. I would have never guessed that the PSU would prevent waking from sleep.

Not enough +5vSB current, perhaps?

I think that's what I read. I'll check my IE browser history as soon as I can get home to my main workstation, and post the link. And in fact . . . now I remember -- I sent that link the other night to a friend in e-mail, so -- no challenge to find it.

In any case, the PSU has to supply a small amount of power to the board, CPU and RAM to keep these sleep-states stable. If it's insufficient, it won't wake up. I think we had a discussion some time back about S1 versus "deeper" modes? An aging PSU may be adequate for S1, but now we've discovered something about old parts.

 

[mikeymikec]

Re: Old parts - the two machines in question in my situation were brand-new, including brand-new CX430s (v1).