Dimm issue?

Discussion in 'Computer Information' started by Diana BB, Mar 3, 2008.

  1. Diana BB

    Diana BB Guest

    Hi All,
    I am not all that computer savvy when it comes to what under bonnet, if
    you know what I mean.

    I live in an area that has many short sharp power outages. I have
    EVERYTHING I own on power surge boards.

    The other day after one such power failure, the computer rebooted with
    the message that goes something like this.

    BIOS has detected a
    non-optimal memory configuration Dual Channel operation requires identical
    paired DIMMS installed in both memory channels. Changing the existing DIMM
    config. can improve systems memory Bus Bandwith. Channel A DIMM 0 should
    match Channel B DIMM 0
    Channel A DIMM 1 should match Channel B DIMM 1. Press<F4> For setup.

    When I looked at the info on in the setup it tells me that
    Channel B DIMM 1 is not installed.

    Now I do not understand what that means....I did have an SDRAM
    installed. Is this the one that is having the problem?

    When it happened the first time I didnt do anything at the setup, just
    exited. But when Windows(Vista) booted the date and time was set to 1st
    of Janusary 2002, 12 midnight.
    It has just happened again this morning....no time when it happens to
    get over and switch the power off at the wall.....this time I had a look
    and saw the above Channel B info.......

    Now I do n ot get this message if I shut down and restart the
    computer...it is only after a power failure.

    Can anyone help me with info or a website that may be able to tell me
    exactly what has happened and if I can repair it or it will need to go
    to the computer repairers.
    Diana in Aus
    Diana BB, Mar 3, 2008
    1. Advertisements

  2. Diana BB

    Kevin Guest

    Does Windows show the correct amount of RAM in your system at present? If
    you have 512 MB of RAM, does Windows show this or does it only show that you
    now have 256 MB of RAM? If your system reports that it now has half the RAM
    it should have, then that RAM stick is probably damaged. You have two
    channels, A and B. They each have two slots for chips, numbered 1 and 2.
    You need matching chips in A1 and B1 and in A2 and B2. You seem to have two
    chips in the A1 and B1 slots, as they should be. You might try resetting
    those chips in their slots.

    It's not difficult. You can find a diagram of the inside of your machine by
    going to the website of the manufacturer, like Dell, HP or whoever, and
    searching for an owners manual. If you have your owners manual, all the
    better. Find the RAM slots, you'll see them as soon as you open the case,
    because there are four of them, of equal length, about four inches (100 mm)
    long, and all in one place on the motherboard. NOTE: unplug everything
    from the case -- especially the power plug before opening the case.

    The RAM chips are locked in place by two fastening levers, on each end of
    the chips. Push both locking levers down at the same time. The RAM chip
    will pop up slightly when you do. Using even, steady pressure on the top
    edge of the chip, push it back down into the slot, making sure it is lined
    up. As you do, note that the locking levers move with the chip to lock it
    in place. Make sure they engage snugly. They will "click" into place. They
    will fit firmly, so you may need to use some pressure. Increase the
    pressure of your fingertips slowly until the chip snaps into place.

    Button up the machine and see if that solves your problem. Double check
    those surge suppressors you are using. They can be damaged by voltage
    spikes and rendered useless. You can get new ones for a few dollars. In
    fact, this might be a good idea anyway if they have taken several of these
    outages you have experienced.
    Kevin, Mar 3, 2008
    1. Advertisements

  3. Diana BB

    Diana BB Guest

    Hi Kevin,
    Thanks for the quick and detailed response.I know exactly what you are
    talking about.I have seen the diagram on the inside of my machine so
    will be able to follow it. MAy do as you suggest though and get a manual
    BEFORE I do anything.

    According to the properties my computer has 1534 MB of RAM!
    I will do as you say and open her up and reset them.
    Thanks a million.
    Diana in Aus
    Diana BB, Mar 3, 2008
  4. Diana BB

    Neil Green Guest

    Not a good idea.
    Leave the power lead plugged in and switch it off at
    the wall, then before you handle any components touch
    the power supply with both hands momentarilly.
    This will discharge any static in your fingers to
    earth or ground.
    Neil Green, Mar 3, 2008
  5. Diana BB

    Diana BB Guest

    We usually short ourselves out on the case. Is it better to do it on the
    power cord?
    Diana BB, Mar 3, 2008
  6. Diana BB

    Neil Green Guest

    Not the cord!!!!
    With the power switched off at the wall but the lead
    still connected open the case then touch the power
    supply unit inside, which unlike the case is
    unpainted, before you handle any components.
    Touching the back of the case will give you the same
    result as it's normally bare steel and well earthed.
    Neil Green, Mar 3, 2008
  7. Diana BB

    w_tom Guest

    Leaving power cord connected to receptacle is bad. But many who
    don't even understand how static electricity damage computers
    recommend leaving the power cord connected. They believe a myth that
    earth ground somehow provided protection. Others who heard this
    'power cord' myth should learn about static electric protection in a
    previous discussion that details the science in alt.comp.hardware.pc-
    homebuilt on 4 Feb 2008 entitled "Unplug the power supply?" at:

    That discussion provides details. Diana BB is doing what will also
    provide protection if (for example) feet are not moved after touching
    the chassis. But a best solution is using a wrist strap attached to
    the computer's chassis and power cord disconnected.

    Returning to the original problem - BIOS boots, first locates then
    tests various hardware, then boots the OS. Diana BB's BIOS test does
    not see correct memory SIMMs. Did it see correct memory before? Was
    anything since changed?. BIOS test functions that only work on power
    up now see failure. What changed since the last time those functions
    operated only during power up?

    Assumed: your computer has two memory SIMMs. Memory boards mounted
    in adjacent slots. Both memory should have same part numbers. BIOS
    reads an ID code from each board. If ID codes are not identical or if
    BIOS cannot read one of the memory boards, then that error message may
    result. This ID code is not used during normal operation; only read
    when booting. Why does BIOS this time read a different ID number?

    So, is memory not identical, or has hardware that reads memory code
    now failed, or did something change in BIOS power up procedures,
    or ... Well, what has changed or was moved? Was power cord
    attached when anything was changed (that can also create failures)?

    Solve this problem by first establishing what is known good and
    what changed between power ups. Do not swap parts until after all
    facts are collected. For example, do both memory boards have same
    part numbers, properly seated, and not touched by static electrically
    charged hand? Just a few obvious questions to first answer - to
    establish what is good.

    Do not leave chassis connected to wall receptacle if removing or
    installing any hardware. Using a static wrist strap attached to
    computer chassis is recommended.
    w_tom, Mar 3, 2008
  8. Diana BB

    Paul Guest

    Based on the time changing to 1st January 2002, you need to replace the
    CMOS battery. Look on the motherboard. A typical battery is CR2032,
    available even at Radio Shack.

    The CMOS battery is only used by the computer, if you switch off the
    computer at the back, or unplug the computer. The current drain is
    tiny, and in the microamp region. On my computers here, if unplugged,
    they last for three years running the CMOS clock from the CR2032.
    (I've had a couple batteries fail here, on stored computers. I leave
    the battery in them.)

    If the computer is left in sleep mode (S3 suspend to RAM), then the
    +5VSB from the power supply, supplies the current for the clock.
    In theory, if alternating between "sleep" and running, without
    switching off the computer at the back, the CMOS battery could
    last for 10 years (or whatever the listed shelf life of the battery is).

    What I don't understand about your symptoms, is why the BIOS did
    not enumerate the memory properly during POST. But change the
    battery, and see if the symptoms change for the better or not.

    Paul, Mar 3, 2008
  9. Diana BB

    Baron Guest

    I disagree, particularly if it carries a protective earth. If it is
    only two wire then an alternative earth connection should be used.
    I agree !
    Ensuring that the machine case remains at ground potential is good
    practice whether It is done by leaving the power cord connected,
    assuming that it actually carries an earth connection, or by using an
    earthing clip.
    Your example is not relevant, however a wrist strap is !

    The sole purpose of the wrist strap is to ensure that there is no
    voltage differential between the human body and the machine that could
    cause any damage to static sensitive components.

    The purpose of earthing the computer is to ensure that any charges on
    the machine case are dissipated.

    But even taking these precautions does not rule out the possibility of
    damage due to static discharge through bad handling.
    Baron, Mar 3, 2008
  10. Diana BB

    w_tom Guest

    Anybody can disagree. But reasons 'why' define which is a
    technically accurate post. Protective earth has no place in the
    static electric discharge circuit as defined in posts in http://tinyurl.com/2565rq
    .. Baron says otherwise but does not even say why; does not even
    define the circuit. Baron either ignored or did not comprehend a
    static electric circuit explained in http://tinyurl.com/2565rq .

    Purpose of a wrist strap is to eliminate a potential difference
    between hand and machine. But not the sole purpose. Various
    functions of a wrist strap were defined in http://tinyurl.com/2565rq
    .. The wrist strap may also discharge those static charges. Wrist
    strap also includes an essential safety function to eliminate a
    potential 'electric shock' hazard.

    But this remains the number one point. Those who only know from
    popular believe (did not learn the science) assume earth ground is
    somehow involved in static electric discharging. One reason for this
    popular myth are so many computer techs who are A+ Certified without
    learning how electricity works. A serious problem for a nation that
    intends to be a world technical leader. A common misconception among
    computer assemblers is that wire is a perfect conductor. Electricity
    is always different at two ends of a wire. No reason to continue. A
    benchmark defines why so many feel earth ground provides static
    electric protection.

    The electrical concepts in http://tinyurl.com/2565rq demonstrate why
    earth ground is not part of a static discharge circuit.

    Diana BB equalizes potential between a human arm and a computer by
    touching the chassis before touching anything else. Baron says that
    is not relevant. Baron says "sole purpose of the wrist strap is to
    ensure that there is no voltage differential between the human body
    and the machine". But what Diana BB is doing when touching the
    chassis (assuming feet don't move after touching the chassis) DOES
    ensure no voltage differential exists between human body and machine.

    Bad handling can also cause static electric damage. See a reference
    to the "EPROM handshake" in that previous discussion; another example
    of good safety practices. Also see references to room humidifier;
    essential for static discharge protection.

    Leaving the power cord connected is a popular myth among computer
    assemblers who believe earth ground is part of a static electric
    discharge circuit. Leaving the power cord connected only creates an
    unnecessary risk to computer hardware. Anyone who works on any
    electronics always disconnects a power cord before touching any
    circuits. A standard practice for both human safety and for
    electronics safety. Reasons 'why' provided in http://tinyurl.com/2565rq
    w_tom, Mar 3, 2008
  11. Diana BB

    Diana BB Guest

    Durrr well thats what I get for not reading it properly.
    Diana BB, Mar 4, 2008
  12. Diana BB

    Diana BB Guest


    You are right I do have two memory SIMMS. The computer was originally
    set up for movie editing with two drives, The second was specifically
    for the movie editing. Since then I gave added more memory.

    I have the program Everest and it tells me that I have the following.....

    [ DIMM1: Kingston K ]

    Memory Module Properties:
    Module Name Kingston K
    Serial Number 05153C00h
    Manufacture Date Week 29 / 2003
    Module Size 256 MB (1 rank,
    4 banks)
    Module Type Unbuffered
    Memory Type DDR SDRAM
    Memory Speed PC3200 (200 MHz)
    Module Width 64 bit
    Module Voltage SSTL 2.5
    Error Detection Method None
    Refresh Rate Reduced (7.8
    us), Self-Refresh

    Memory Timings:
    @ 200 MHz 3.0-3-3-8
    @ 166 MHz 2.5-3-3-7
    @ 133 MHz 2.0-2-2-6

    [ DIMM2: 128M64K-40C ]

    Memory Module Properties:
    Module Name 128M64K-40C
    Serial Number 040413C3h
    Module Size 1024 MB (2
    ranks, 4 banks)
    Module Type Unbuffered
    Memory Type DDR SDRAM
    Memory Speed PC3200 (200 MHz)
    Module Width 64 bit
    Module Voltage SSTL 2.5
    Error Detection Method None
    Refresh Rate Reduced (7.8
    us), Self-Refresh

    Memory Timings:
    @ 200 MHz 3.0-3-3-8
    @ 166 MHz 2.5-3-3-7
    @ 133 MHz 2.0-2-2-6

    [ DIMM3: Kingston K ]

    Memory Module Properties:
    Module Name Kingston K
    Serial Number 07153E00h
    Manufacture Date Week 29 / 2003
    Module Size 256 MB (1 rank,
    4 banks)
    Module Type Unbuffered
    Memory Type DDR SDRAM
    Memory Speed PC3200 (200 MHz)
    Module Width 64 bit
    Module Voltage SSTL 2.5
    Error Detection Method None
    Refresh Rate Reduced (7.8
    us), Self-Refresh

    Memory Timings:
    @ 200 MHz 3.0-3-3-8
    @ 166 MHz 2.5-3-3-7
    @ 133 MHz 2.0-2-2-6

    [ Memory Devices / J5G1 ]

    Memory Device Properties:
    Form Factor DIMM
    Type Detail Synchronous
    Size 256 MB
    Speed 400 MHz
    Total Width 64-bit
    Data Width 64-bit
    Device Locator J5G1
    Bank Locator CHANNEL A DIMM0
    Manufacturer Manufacturer1
    Serial Number SerNum1
    Asset Tag AssetTagNum1
    Part Number PartNum1

    [ Memory Devices / J5G2 ]

    Memory Device Properties:
    Form Factor DIMM
    Type Detail Synchronous
    Size 1024 MB
    Speed 400 MHz
    Total Width 64-bit
    Data Width 64-bit
    Device Locator J5G2
    Bank Locator CHANNEL A DIMM1
    Manufacturer Manufacturer2
    Serial Number SerNum2
    Asset Tag AssetTagNum2
    Part Number PartNum2

    [ Memory Devices / J5H1 ]

    Memory Device Properties:
    Form Factor DIMM
    Type Detail Synchronous
    Size 256 MB
    Speed 400 MHz
    Total Width 64-bit
    Data Width 64-bit
    Device Locator J5H1
    Bank Locator CHANNEL B DIMM2
    Manufacturer Manufacturer3
    Serial Number SerNum3
    Asset Tag AssetTagNum3
    Part Number PartNum3

    [ Memory Devices / J5H2 ]

    Memory Device Properties:
    Form Factor DIMM
    Type Detail Synchronous
    Speed 400 MHz
    Total Width 64-bit
    Data Width 64-bit
    Device Locator J5H2
    Bank Locator CHANNEL B DIMM3
    Manufacturer Manufacturer4
    Serial Number SerNum4
    Asset Tag AssetTagNum4
    Part Number PartNum4

    I had one of the drives crash about 12 months ago after a root kit
    virus. It was the SATA drive. I replaced this with a PATA...I was told
    that was okay.
    But the only time I get this message is when it reboots itself after the
    last two power outages. They are very brief outages and there is no time
    to turn the computer off at the wall. When I shut down per Windows
    shutdown and restart the computer I do not get this message. Everything
    appears to be normal.
    As you say we always unplug everything from the computer before opening
    the case and stand on tiles and short ourselves out before we touch
    I do have it set for fast boot( I think that is what it is called) in
    the BIOS setup. Is that why I do not get the message in a normal start up?

    Thanks for your input and advice.

    Diana In Aus
    Diana BB, Mar 4, 2008
  13. Diana BB

    Diana BB Guest

    Thanks Paul,
    I will give that a go first.
    That is what I couldnt understand about the whole thing. Why was it
    reverting back to that date?
    And I cannot understand why I am not getting this message during a
    normal shutdown/start up. Everything is normal under those conditions.
    The time and date dont change.

    I do leave my computer going almost all the time and do use sleep mode.
    Although that can be a pain as it seems the longer it is in sleep mode
    the harder it is to 'wake' it up.....mostly over night. I do believe
    that may be a problem with my graphics card though...remember reading it
    somewhere sometime.....before it started being a problem for me.
    Diana BB, Mar 4, 2008
  14. Diana BB

    Paul Guest

    I probably didn't explain it clear enough.

    The CMOS clock is a thing like a digital watch, and is located in
    the Southbridge. It keeps time for the computer.

    Most of the time, the CMOS clock is running from your power supply.
    When the power supply is not available, a tiny bit of power is
    drawn from the CMOS battery instead. And that is why they put
    the battery there.

    Now, let's try another scenario.

    1) The A.C. power just failed. Now there is no power from the PSU available.
    2) The CMOS battery (CR2032 coin cell) is flat, and offering below 2.4V
    (Normally it is about 3V or so.)

    If all power is gone from the computer, the CMOS clock stops. The
    BIOS settings are stored in a tiny RAM located in the CMOS clock.
    The RAM loses its contents, because there is no power. The RAM is
    volatile, and needs a source of power.

    Now, the power comes back on.

    The clock resets itself, to a well known, but visually bogus date
    and time. Old Macintoshes, for example, reset themselves to a year of
    1980. Other more modern computers, seem to be using a more recent, but
    still out of date, year at reset. I think the machine I'm typing on,
    uses 2002 or so.

    After the A.C and battery failure, the CMOS RAM is empty. The BIOS
    consults it, and notices the corruption. What the BIOS is supposed to
    do, is use all "Auto" settings, re-enumerate the hardware (notice all
    your RAM), set everything up, and carry on. Sometimes, the computer
    won't boot, because the default boot order is not what the user was
    previously using.

    If you used any custom settings in the BIOS screens, like boot
    order, enabled RAID for drives, that sort of thing, then if
    the CMOS loses all power, all the good settings are lost.
    You have to put them back. (A real PITA, if you haven't looked
    in the BIOS lately.)

    If the CMOS battery is flat, then every time the power fails,
    you'll be looking at "2002" date again, and your custom settings
    will be lost. That is why people replace the battery when it
    goes flat. So that, in the event of a power failure, custom
    CMOS settings will be maintained.

    It still doesn't explain why all RAM would not be found. Even
    if the BIOS has no recollection of what happened yesterday, it
    should still be able to detect all the hardware present in
    the machine. If some hardware was busted, that would be a reason
    for not being able to detect it.

    Paul, Mar 4, 2008
  15. Diana BB

    Baron Guest

    w_tom wrote:

    I would suggest that you go and read exactly what my post said !

    I would also read what the posts in your quoted link say !
    Baron, Mar 4, 2008
  16. Diana BB

    w_tom Guest

    OK. Now I am confused. Your computer BIOS can read from one pin the
    information on each SIMM. But that report says you have 4 memory
    SIMMS - not two. For example, Simm 1 and 3 have two different
    Kingston serial numbers: 05153C00h and 07153E00h.

    One solution is to back the system down to a known good
    configuration. For example, only install both Kingston memories.
    Test. If it works, then replace those Kingston memories with other
    1024 Mb memories and test again.

    This to establish integrity of both memory sets. If both memory
    pairs check out, then install all to learn if or what the BIOS
    complains about. We know this much. BIO clearly reads and identifies
    what memories are installed where. So any error message is not
    because memories are misidentified or hardware failure has destroyed
    that 'ID the part' hardware that operates during power up.

    According to memory specs, your motherboard should be PC3200 meaning
    that the faster 1024 MB memory is even faster than required - maybe
    more stable. Just an observation; probably not a useful conclusion.

    Returning to the 'surprise' failures created by static electricity:
    we once identified tiles that created static electric charges after a
    wax was applied. Never did learn what wax created the problem since
    my part was to eliminate a hardware design defect. Static electric
    charges are created by numerous materials when room humidity is too
    low. Static electric discharges can even create a failure that does
    not appear for months. Just another reason why we take precautions
    such as static wrist strap or touch the chassis case.

    Wish I could provide more information on your (intermittent?)
    problem. If it happens again, well, you are better informed to
    identify the failure before it again goes away. Of course, there is an
    execution difference between a cold boot (no problem seen) and a warm
    boot (boot without power cycling). Just another pertinent fact of how
    hardware gets reset; information to collect if the problem happens

    Hopefully this is useful should future problems occur.

    w_tom, Mar 5, 2008
  17. Diana BB

    Diana BB Guest

    Thanks again Paul,

    Wasnt your fault I didnt understand. I dont know a lot about the
    internal workings of my computer. Usually learn when something goes wrong...

    I am wondering though if I actually have two separate issues at the moment.
    One being the memory and the other the CMOS battery as you suggest.

    Still dont understand that I dont get this message in a normal start up.
    But I did find something interesting today as I was looking around
    trying to find out about the CMOS/BIOS message.

    Found it on my motherboard manufacturers website.....although I did look
    there before but didnt have as much knowledge as I do now.

    It was referring to a memory mismatch error(that is the message I am
    getting isnt it? )

    Says that the new size will be detected automatically but I would still
    need to intervene in order to avoid the error message.

    Said to go into the CMOS/BIOS setup utility, hit ESC and choose to save
    when prompted. When it came up at the power failure I didnt know what to
    do so I just exited without saving.
    I did as it said above this afternoon. Dont know if it fixed anything
    guess time and another outage will tell.....

    I am not sure but it says on the BIOS page the following
    Memory Channel A Slot 0 256 MB DDR400
    Memory Channel A Slot 1 1024MB DDR400
    Memory Channel B Slot 0 256 MB DDR400
    Memory Channel B Slot 1 Not Installed.

    From my understanding of the error message this does not appear to be

    If it does happen again I guess I will just hit the esc key and save.
    But I did find that I couldnt hit the esc key and get the save and exit
    option. It asked if I wanted to discard and exit. I actually hit the F10
    key...assuming that is just the way this motherboard wants to do it.....

    And if that doesnt stop this problem I will refer back to w_toms suggestion.

    As for the time/date thing I will replace as you suggested and see if
    that fixes that.
    Thanks for explaining it easy enough for me to understand. Both you and
    w_tom have been great. Sometimes the answer can leave you more confused
    than you were when you asked the question.
    Thanks again.

    Diana in Aus
    Diana BB, Mar 5, 2008
  18. Diana BB

    Diana BB Guest

    Thanks w_tom,
    It has been very helpful. As I wrote to Paul,being more aware of what
    the problem is I was able to find some info with regards to the brand of
    motherboard I have. It seems that when I added the extra memory card it
    was detected automatically but I needed to intervene in order to not
    receive a memory mismatch error.
    I have done as I was instructed and will now have to wait and see if I
    get the error message again.....
    I am still not sure about the info in the BIOS with regards to the slots
    but I will see what happens.
    The computer was set up with the two Kingstons I do know that much. As I
    said one was designated to the drive that ran the editing program.I am
    assuming that if I went back to that then the error message would stop
    if that is the original config?
    When I read the info in the BIOS it says......

    Memory Channel A Slot 0 256 MB DDR400
    Memory Channel A Slot 1 1024MB DDR400
    Memory Channel B Slot 0 256 MB DDR400
    Memory Channel B Slot 1 Not Installed.
    That seems mismatched doesnt it? But not sure how to fix it.......hope
    you havent answered that and I have not understood what you were saying:

    You wrote......
    I am not sure what you mean by this. If you have time to explain I would
    like to know.

    Thanks so much for making it easy enough to understand. I do appreciate
    that and your time to help.

    Diana in Aus
    Diana BB, Mar 5, 2008
  19. Diana BB

    Paul Guest

    To tell if something is actually wrong with the RAM, you'd start
    by taking the side off the computer case, and verifying how many
    sticks are physically present. Based on the thread so far,
    you seem to have four slots. A four slot machine is typically a
    dual channel capable machine, meaning there can be a performance
    advantage by meeting certain memory matching criteria.

    The reason we'd want to open the case, is to count the sticks.
    If four sticks were present, and one of them just died, then
    your symptoms would be a bit easier to understand. If there
    were three sticks, and there had always been three sticks,
    then you might have seen that "mismatch" message before.

    Depending on the computer and the quality of its documentation,
    it might explain somewhere, how "matched" the memory has to be.

    Some Intel chipsets, use "A1+A2=B1+B2" to get dual channel
    operation. What that means, is the total memory in each channel,
    has to be the same quantity. For example, it would be sufficient
    to use 512MB+0MB = 256MB+256MB. That would be three sticks,
    where one of the sticks is twice the size of the other sticks.
    And yet it could run in a dual channel mode.

    Some other chipsets use A1=B1 and A2=B2. To meet that, you'd perhaps
    buy pairs of DIMMs packaged as such (for convenience). You could
    use 2x256MB and 2x1GB, for a total of 2.5GB of memory. A pair would
    go into A1,B1 and a pair would go in A2,B2. The dual channel slots
    may have identically colored DIMM slots. (A1,B1 white plastic,
    A2,B2 blue plastic etc. Color is arbitrary, and contrasting colors
    are used to make figuring out the matching easier.)

    Since you say you currently have 256+256+1024, that set of
    three sticks cannot meet either of the above stated cases.
    That means the motherboard would run in single channel mode.
    That may reduce memory bandwidth, and cost about 10% application
    performance (just a rough guess).

    You can test the memory sticks, one at a time, even at the BIOS
    level. You could power off and install just one stick at a time,
    and verify it is working OK (like seen in the BIOS, or test with
    memtest86+ on a floppy disk).

    If you are satisfied that the three sticks are being identified
    correctly, and you really do own 256+256+1024, then there is no
    need to test the sticks individually (unless you suspect
    one of them is bad or is throwing errors or the machine
    is crashing).

    To make the sticks match, the easiest addition would be another
    1024MB stick. That would give enough memory, to suit either kind
    of chipset behavior described in the two examples above.

    Or you can just ignore the issue entirely :) Plenty of people
    run systems with single channel configurations, and are none the
    wiser. This issue doesn't have to affect stability, and if you
    aren't looking for every ounce of performance (like this is your
    email machine for example), in fact there might not be an incentive
    to even take the side off the machine. If, on the other hand,
    the machine crashes a lot, then memtest86+ and Prime95 are
    good tools for detecting the problem (and perhaps even
    fixing it). To work around memory problems, you'd need to
    know a bit more about adjusting the BIOS. And not every
    motherboard has the necessary settings (microATX boards,
    the ones that are 9.6"x9.6", don't tend to have very
    good BIOS in this regard).

    If you want to go further on this, either indicate the
    make and model of computer, or the make and model of
    motherboard. And also state what actual sticks are
    installed in the slots (like if you have four sticks
    and only three are being detected).

    Before touching the sticks, you do want the switch on
    the back of the computer to be in the OFF position. On
    my computer, I unplug the power cord, mainly as a fail
    safe way of assuring myself there is no power present
    in the chassis. Standby power keeps the RAM sockets
    powered, which is why I unplug for safety. You shouldn't
    insert or remove stuff from live sockets, as it can get
    damaged. (Yes, we can have a long discussion about
    static electricity, but I don't want to go into that
    now :) )

    Paul, Mar 5, 2008
  20. Diana BB

    w_tom Guest

    Before fixing anything, you want to know if that battery is good or
    bad. This is but another in a long list of reasons why a 3.5 digit
    multimeter is such as powerful tool. Do not remove the battery. Use
    the meter in 20 VDC setting to read that battery voltage. For
    example, if that coin cell battery is a CR2032, then voltage should be
    something like 3.0 volts. If only 2.8 volts, then battery should be
    replaced sometime in the next year. If lower, then battery may be
    part of your problem.

    Best battery voltage measurement is with battery always attached to
    the motherboard. Now you know without any confusion whether the
    battery would explain your problem. If battery is above 2.8 volts,
    then your problem is elsewhere.

    Multimeter is an inexpensive tool sold in Radio Shack, Sears, Lowes,
    any hardware store, Home Depot, or even K-mart. The tool is that
    ubiquitous and that simple. Best price is often in Wal-mart for about
    $16. Multimeters look complex. But a cell phone is far more complex.

    Paul has described how memory works. Also useful is to record part
    numbers on each memory board and in which slots each memory is
    located. This could confirm what the BIOS reports and better explain
    which part of Paul's post applies to your configuration.

    Usually, the relevant number is on a paper pasted on that board.
    If no paper label, then recording the part number on duplicated on
    each IC and counting the number of ICs could also physically identify
    that Simm.

    Meanwhile, taking that battery voltage (and identifying the battery
    - is it a 2032?) would immediately terminate any further battery
    speculations AND suggest how many years before a new battery is
    w_tom, Mar 5, 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.