startup delay

Hi -

my parts

no oc
msi p965 platinum
2 x 1G corsair valueselect 533
e6600/stock hsf-fan
silverstone st50ef-plus 500w (8-pin cpu pwr)
evga 7600gt pcix16
2 seagate sata hds
2 pata dvd
usb rf mouse/kb
1 12cm chassis fan
coolermaster centurion case (mb firewire, hda audio, 2 usb hooked to chassis front)
verizon fios (using onboard eth connector at rear of mb)

what's happening

I have the side of the chassis off and I can see that when I press the chassis on/off switch to power on, the cpu and chassis fans spin up, come to a stop and about 1-2s later come back on and the system comes up. Nothing comes up on the screen until after this delay; everything seems to run normally from that point (been using the computer for about 3 weeks, linux mostly, a little xp and vmware and nothing whatsoever seems out of whack besides this startup problem).

I only have a few days left to exchange parts at newegg but don't have another psu to swap in to see if that's the problem. It could be the case on/off switch, the psu or the mb (and possibly something else - I have no idea). So I'm hoping you guys can offer some advice.


Originally after building this, EVERY single time I hit power on, this delay would occur. So I talked to msi and they had me clear cmos - that seemed to have fixed the problem but now I'm not so sure it didn't just turn it into a different kind of startup problem.

So after clearing cmos the computer seemed to start normally and I closed everything up and stuck the computer under my desk (I disconnected sys/hdd leds so I have to look at the fans to know if it's powering up correctly). A week after clearing cmos I started thinking the problem had returned, took off the chassis side and sure enough it had.

I talked to msi again and they had me flash to the latest rev bios - didn't fix it.

What's changed is that either after clearing cmos or flashing bios the problem morphed into one where if I have the machine running, power down then turn power back on, there's no fan spin-up/down delay (i.e., computer starts like it should); but after having it off all night then powering up the next day, I get the fan spin-up/down delay.

So the original problem has changed into this (as best as I can tell - it might not be this cut and dried because I haven't done that much experimenting with it). And to test different things I may have to wait hours with it off to be sure the delay problem actually occurs upon restart. Or maybe not - it's all shooting in the dark knowing if the problem would've occured if I had left it off longer, or if whatever test I run actually resolved/pinpointed the problem.

I talked to silverstone and all they said was "quality unit", "never heard of this before"; msi also says they never heard of this before (I find it hard to believe neither has but whatever).

I have no psu to swap in; I'm mostly flying in the dark here.

So can somebody tell me what's going on here? My problem is it's going to get expensive shipping back the psu; mb will be a pita to have to reinstall (I have no idea how to replace the thermal tape intel's stock hsf-fan comes with, let alone undoing my tidy setup and having to redo all the connections); case won't be cheap to ship back either since I'm wondering if the problem is its on/off switch (but since the original problem morphed I'm leaning against it being the problem).

As of right now anything I return is a complete guess and no guarantee an exchanged unit even fixes the problem. And if I return one thing and it turns out to not be the problem, I'll be outside the return dates for new items and get stuck with refurbished crap from the mfgr for whatever else I exchange. And if I just send back case, psu and mb hoping one of the replacements fixes the problem, I'm blowing a stack on return postage expenses.

Any ideas? Does this problem sound terminal? Should I just hope this is the worst it ever gets?

Help? :)
5 answers Last reply
More about startup delay
  1. I have a Gigabyte 965P-DS3 and it does a similiar thing. Did some research and found on the Gigabyte site that this is expected if the main power is dropped (turn off ps or unplug system) and is not there if the system is shut down normally (with ac power still supplied to ps and ps not turned off) and then started. The site reported this is because of the 965 chipset. Don't know if that is true, but it is how mine behaves. Been running it for several months and no problems.
  2. I'm not using the Gigabyte DS3, but rather the Asus P5B Deluxe (based on the same chipset), and I get the same behavior from time to time. BIOS updates seem to have smoothed it out some, but it still happens on occasion.

    If I had to guess, I'd say that it really is part of the chipset power-up process on cold boot, though I have no idea why that might be.

    I would seriously doubt that there is anything wrong with your PSU or case. Silverstone does make "quality units," and there's no way any intermittent fault in the power switch could cause that. The computer would simply power up, then power down ~5 seconds later. Reset switch could possibly do it, if there was a connection loose in just the right way, but the odds of that are pretty slim (I've never seen it in 14 years of servicing PCs).

  3. From a different forum:

    "When I OC my board when I turn it on it turns on for a second shuts down for another second then turns back on."

    and the reply:

    "Its a set feature in all P965 chips. It will shut off your pc to make sure it has all the correct clock speeds. This cannot be disabled so I guess you will have to use other means to power your mach II"


    "I've found it will do the power-off -> power-on -> power-off -> power-on cycle only after I have been fooling around in my bios. At any oter power up & reboot, everything starts up as I would expect (no annoying power cycling). This happens not only on my Commando, but also the P5B Basic & P5B Deluxe I have. My P5W-DH cycles power at every reboot, whether warm or cold. In all cases these boards do this only while overclocked."

    There are many posts in forum reporting this behavior. I do not think you have a worry with this, and an exchanged board will behave the same.
  4. From page 161 of the 965 chipset spec (link )

    5.2.50 PMSTS—Power Management Status

    Size: 32 bits
    This register is reset by PWROK only.

    Bit 8 = Warm Reset Occurred (WRO): This bit is set when a ResetWarn is received, and cleared by PWROK=0.
    0 = No Warm Reset occurred.
    1 = Warm Reset occurred.
    BIOS Requirement: BIOS can check and clear this bit when
    executing POST code. This way BIOS knows that if the bit is set, then
    the PMSTS bits [1:0] must also be set, and if not BIOS needs to
    power-cycle the platform.

    Hope this sheds some light on what is happening
  5. Thanks guys - appreciate it loads. Apparently nothing is broken and I don't need to exchange anything.

    Although I was trying to be as complete as possible about my setup in OP, based on Firemist's replies as it turns out I skipped a super critical piece of information - my psu is plugged into a power strip that has its own on/off switch. When I turn my computer off at night I also turn off the power strip - that is what reveals this peculiarity.

    There are/were two problems because originally the power strip was never turned off when I was building my computer but I would always get the startup delay. After originally clearing cmos it seemed as if I only had the problem if my computer was off for many hours.

    The reason it seemed that way was because the only time it was off for many hours was at night when I turned the computer off; AND turned off the power strip.

    I turned my computer off for ~4 hours but left the power strip on and it started up fine. I turned it off again and turned off the power strip then in a minute or two restarted and voila! The delay.

    So no matter what the root cause of this existing behavior is, it only shows up if the psu is entirely disconnected. cmos must've been corrupted before I reset it and what I thought had morphed into a different problem had actually been fixed - I just didn't know it.

    Thanks again, it's a big relief.
Ask a new question

Read More

Power Supplies Components