s5393 power up

Discussion in 'Tyan' started by jim, Jun 9, 2008.

  1. jim

    jim Guest

    i am having trouble getting the s5393 tempest i5400pl to power up.
    when you turn it on it will spin the fans for a few seconds then shut
    off. sometimes it boots fine. it does not reliably power up. is anyone
    having this problem??
    jim, Jun 9, 2008
    1. Advertisements

  2. jim

    The Doctor Guest

    How are the CPU fans?
    The Doctor, Jun 10, 2008
    1. Advertisements

  3. jim

    jim Guest

    they spin with the other fans for a few seconds. they also run fine if
    it decides to boot. i have already had two different psu on it so i
    don't think its that.
    jim, Jun 10, 2008
  4. jim

    The Doctor Guest

    I know what 2 CPU heatsinks and fans were not working on my S2727 the
    MB would shut down right away because the feedback was incorrect.
    The Doctor, Jun 10, 2008
  5. jim

    jim Guest

    jim, Jun 11, 2008
  6. jim

    midol Guest

    midol, Jun 11, 2008
  7. jim

    Paul Guest

    The CMOS RAM is housed in the Southbridge chip. It is powered from a
    separate rail on the Southbridge chip, and special isolation techniques
    are used between the area where the CMOS and RTC (real time clock)
    are located. That cuts down on leakage, when the rest of the computer
    is not powered. (It is referred to as the "CMOS well", because it
    can be at a different potential than the rest of the Southbridge chip.)

    The CMOS RAM runs from the logical ORing of the battery and a voltage
    made from +5VSB. If you regularly turn off the computer at the back
    of the machine every day (i.e. the switch on the PSU is switched off),
    then the CMOS battery (likely a CR2032) is called on to power the
    CMOS and RTC. (Check the manual or look at the battery itself, to
    verify it is a CR2032. Years ago, there were more different battery
    formats used.)


    If you leave a computer unpowered (switched off at the back or unplugged),
    a CR2032 lasts about 3 years. If +5VSB is available all the time (such
    as what happens if the computer is left in suspend to RAM), then the
    battery can last for its shelf life, perhaps 10 years. In other words, if
    +5VSB is available, the battery is not supposed to be used at all. All
    the microamps come from a linear regulator tied to +5VSB.

    The Southbridge has a minimum voltage spec, for continued operation of
    the CMOS area. Say it is 2.0 volts, for example. The battery is connected
    to the Southbridge via a Schottky diode, which wastes another 0.4V
    at the tiny current flow level involved. (The diode prevents accidental
    attempts to recharge the battery, which could cause it to burst.) So
    the minimum battery voltage might be 2.4V, to have the thing continue to
    remember the CMOS settings while the power is off. New batteries likely
    read 3.0V or higher. A multimeter can be used to check the voltage
    level, and no artificial loading is needed to evaluate the battery.

    When the BIOS starts up, it computes the checksum of the info in the
    CMOS RAM, as a mechanism to detect corruption. It is possible the
    password field in the CMOS, is protected by its own checksum. Simple
    checksums are what the BIOS uses, to judge the contents of the CMOS


    Some users have claimed, that if you pull the old battery, and install
    the new one without a lot of fumbling about, that the settings can be
    maintained. It all depends on whether there are any capacitors present,
    which hold a small amount of charge, while you're changing the battery.
    It helps to have a paper list handy, with any custom settings that
    are needed to get the machine to boot, for occasions where the
    settings are lost, and default settings are being used. Alternately,
    a series of digital camera photos of the BIOS screens, can be used
    as a reference for later (if you store the pictures on another

    Since the machine seemed to run, while using the defaults, it could
    be that there are no custom settings being used at all. Normally,
    something like the boot order, would have to be restored in the
    BIOS setup screens, to get any work done.

    Paul, Jun 11, 2008
  8. jim

    jim Guest

    so is it possible i have had two bad psu on it?? the third one should
    arrive in a few days. it is a 1200watt thermaltake W0133RU that is
    eps12v compliant. any thoughts??
    jim, Jun 11, 2008
  9. jim

    Paul Guest

    The power supply itself, will not run unless +5VSB is functional. +5VSB
    is also the backup power source for the CMOS. The CR2032 and the
    +5VSB sourced sustaining voltage, power the CMOS. For the CMOS to
    entirely lose it settings, and for there to be a checksum complaint
    for the CMOS, you need a bad CR2032 battery, and also to have shut
    off the power supply completely.

    The BIOS code in its EEPROM, is also protected by a checksum, but
    the error message should be different, if that was the problem. A BIOS
    checksum failure can be caused by accidental overclocking of the LPC
    clock used with the LPC bus to the EEPROM. (Trying to read bytes from
    the EEPROM, before they've settled.)

    I don't have enough information, to tell you what exactly is up with
    your original symptoms. The power supply itself has protection, and
    some of the high end ones have more protection features than the cheap
    stuff. Protection on the power supply, should start to be applied, about
    35 to 50 milliseconds after the power supply starts - that gives time
    for it to stabilize. After that, the power supply may be able to detect
    a problem with current draw. Or, the power supply could have a failure
    condition internally, leading to it shutting off. (Like rectifier
    overtemperature or overcurrent detection.)

    If the power supply shuts off, at least some of them have a latch-off
    feature. They won't recover, until you flip the switch on the back, to
    the off position, and then back to the on position.

    Motherboards also have sophisticated protections. The Vcore regulator has
    an overcurrent protection feature, and it too can be designed with
    latch-off as a feature. Pressing the power button on the front of the
    computer case, may not be enough to re-cycle the power, and it might
    still need to be shut off at the back, to recover.

    If the BIOS detects a CPU fan with no RPM signal, that can be a reason
    for shutoff. That usually takes a few seconds. The BIOS tells the
    motherboard to shut off the power in that case.

    The time constant of the shutoff, can sometimes shed some light on the
    failure mechanism. For example, if pressing the power button on the
    front of the computer, only elicits a "twitch" from the CPU fan, then
    you know that is just enough time for the 35 to 50 millisecond interval
    on the power supply to have passed - then you know the power supply
    is what detected the problem. As the interval gets longer, things like
    motherboard or CPU temperature (THERMTRIP) issues, could be at fault.

    The ones that are hard to figure out, involve the inability of a
    new motherboard, to reliably detect initial power up. Sometimes, a
    source of leakage current (like down the monitor cable), prevents the
    motherboard power monitoring circuits, from properly determining that
    the board is off. Then, when the power button is pressed on the
    front of the computer, it won't start up properly. You can experiment with
    the reset button, but I have at least one motherboard in my
    possession, that chooses to ignore the reset button, whenever it
    gets into a foul mood. (And *no* chip should ignore reset. That was
    a shooting offense at the place I used to work. A design should *always*
    be resettable.)

    With so many different ways for a motherboard to fail to initialize properly,
    you can't always isolate the source of the problem to one exact thing.
    For example, there were cases in the past, where certain Antec power
    supplies, mixed with some Asus motherboards, resulted in problems.
    And yet, if the components were mixed with others, perhaps they'd
    be perfectly happy. That one seemed to be related to a timing problem,
    like the power supply being too slow to get started.

    It could always be a motherboard problem - in which case you want to
    find a private forum with enough feedback on the product, to see whether
    there is a known issue or not. For server motherboards, the best place
    I know of is forums.2cpu.com . The reviews on Newegg, are good for
    detecting problems with desktop motherboards, but for a server
    motherboard, it is harder to find enough customer reviews, to get
    a clear picture.


    Paul, Jun 12, 2008
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.