Re: Power Supply Cause of Crashes? (Review)

From: W. Watson (wolf_tracks_at_invalid.inv)
Date: 03/20/05 

Date: Sun, 20 Mar 2005 22:08:17 GMT

w_tom wrote:

> First, if you ever read anything pejorative in anything I
> post, then you have missed the intent of that post. I am
> blunt, honest, and quite technical - which means nothing
> special is posted to appease human emotions. Those emotions
> simply get in the way of problem solving. Problems - as evil
> as satan.
>
> I am suggesting you back up and first establish some
> fundamental baselines. For example, nowhere do I read
> anything that says the power supply 'system' is acceptable.
> One cannot prove a good power supply 'system' by just
> replacing on of the 'system' components - the power supply.
> Posted on Saturday are hyperlinks to two previous discussions
> that define how to verify power supply 'system' integrity in
> only two minutes. Background information unique to your
> situation is also in that previous 19 Mar 2005 post.
>
> Second, you are trying to solve OS (software) and hardware
> questions simultaneously. As unknowns increase, then the
> complexity of that problem increase exponentially. We must
> break this problem down into parts. Then one by one establish
> the integrity of that part. Above is the first thing that
> creates so many strange and unrelated problems - the power
> supply system. Next are the other hardware components. This
> is why better computer manufacturers provide comprehensive
> hardware diagnostics for free - either on their web site on on
> the hard drive. Comprehensive diagnostics test hardware
> without any use of the OS. If you computer manufacturer does
> not provide those diagnostics, then you must download them
> individually from the various hardware component manufacturers
> or from third parties.
>
> For example, one reason for your strange behavior could be
> questionable memory. Posted earlier today is how you verify
> integrity of that memory - that also included using heat to
> aggravate intermittent (soft) problems. Third party sources
> for memory diagnostics include Memtst86 and Docmem.
>
> Other 'usual' suspects include video controller and
> soundcard.
>
> Once we have established hardware integrity, only then is
> the 'nuke and pave' option safe to use. However as Floyd
> Davidson noted, much to be gained by actually first finding
> the OS problem.
>
> Again, swapping part such as using the Antec does not
> confirm power supply 'system' integrity. The procedure was
> posted previously and definitely does not call for using an
> oscilloscope - although the scope does provide even better
> information. Still there are numbers that only a meter can
> provide with sufficient accuracy. (if your motherboard also
> has a voltage monitor, a meter is necessary to calibrate that
> voltage monitor).
>
> One tool that everyone doing things electrical has is the so
> inexpensive and so ubiquitous multimeter. Even sold in Sears,
> Home Depot, Radio Shack and Lowes. $20 for this essential
> tool. A tool so complex that even junior high school students
> use it. It looks complex. Its simple. Scary at first which
> is why so many computer assemblers don't understand it. But
> it is simple. Select the reading. Touch probes to two
> wires. Read number. That simple. And again, in but two
> minutes, the power supply 'system' is verified.
>
> In summary, forget for now, the OS. Just concentrate on
> establishing the integrity of hardware. Once hardware
> integrity is established, then OS problems can be addressed.
> But your symptoms are more synonymous with intermittent
> hardware problems.
>
>
> "W. Watson" wrote:
>
>>What specifically do you suggest I do? Am I missing something in
>>your post? Are you asking me to measure voltages or pull out an
>>oscillioscope and observe the wave form? Note that I replaced the
>>power supply with an antec, and got the same results. Did you
>>notice the review in the subthread in which I summarized
>>matters, and the prior post in which I discussed the PSU
>>replacement.
>>
>>Again pejorative. Don't misunderstand me. I appreciate the
>>methodical problem solving technique, but you are assuming an awful
>>lot about my experience and availability of technical tools.
>>
>>What I find is that I get to listen to everyone, and I get to make the
>>decisions. In group problem solving, the key to it is friendly cooperation and
>>positiveness. I sincerely appreciate suggestions, but I get to decide what to
>>do. You may be critical of my methods, but I know where I'm headed with the
>>information I have and I will take the risks.
>>
>>I suspect there about 7-8 different people who have responded to my message.
>>Each one has a different view. This is typical of what happens on a problem. I
>>find it useful for possible solutions. However, ultimately NGs become cumbersome
>>on a particular point. Phone conversations, personal meetings, etc. ususally
>>work faster to clear my misconceptions. This is an imperfect medium.
>>
>>At the risk of being pejorative myself, I can suggest many good sources on such
>>matters. One can start with Aristotle. "A life unexamined is not worth living."
>>A little severe, but he's on the right track--self examination. See. I followed
>>his advice. I thought it might be Plato, and just looked it up. It was Plato. ~

Since I've go the time, and interest, I'll give this a fair shake. First, we
need to clarify a few things.

Tell me what is a 3.5 digit multimeter? I know what a multimeter is. I happen to
have a Fluke digital mulitmeter, but I'm curious what you think a 3.5 is.

What is to be found at <http://www.hardwaresite.net/faqpowersupply.html>? I see
no chart there.

It looks to me the heart of what you say is:
=========start===========
  If motherboard control circuits have power: motherboard
control circuit orders power supply on and off via green
wire. Power supply should be off when voltage is above 2.4
volts. Power supply on when voltage drops to below 0.7 volts
(when power switch is pressed). If motherboard control
circuit does not responde, then multimeter can verify that
power switch functions as a pushbutton (is not getting stuck
in the on position).

   Next are the output voltages from power supply. Red (+5 V),
orange (+3.3 V), and yellow (+12 V) wires must see a rising
voltage. If any of these voltages is not available, then
power supply cuts off other voltages. IOW with multimeter,
monitor each voltage as power switch is pressed. Do they rise
up and meet limits in that above table?
==========end====================
or is it?
==========start==================
Procedure: Motherboard turns power supply on and off. But
controller circuit requires power. When power supply is
plugged into wall receptacle, then purple wire (from power
supply to motherboard) provides 5.0 volts (within limits of
chart).

   When power button is pressed, then motherboard controller
tells power supply to power on. Green wire is above 2.4 volts
when power supply should be off. Green wire should measure
less than 0.7 volts when power button is pressed - to command
power supply on.

   Three components - power supply, motherboard controller, and
power switch. Using meter, discover which is or is not
working. Only then were you ready to replace something.
Voltage Wire Color Min V Max V
  +5 V Red 4.75 V 5.25 V
  -5 V White -4.75 V -5.25 V
  +12 V Yellow 11.4 V 12.6 V
  -12 V Blue -11.4 V -12.6 V
  +3.3 V Orange 3.135 V 3.465 V
  +5VSB Purple 4.75 5.25
  !Power On Green 0.8 2.0
  Power OK Gray >2.4 when power is good

   BTW, if you did not pull plug from wall receptacle before
removing supply, then motherboard failure is a new
possibility.
==========end====================

What is to be gained by looking at:
=================start==========
" ... bean counter' supply cannot provide sufficient power on one
voltage. Latter problem discussed in:
  http://firingsquad.gamers.com/guides/power/default.asp
and so often demonstrated by:

http://www6.tomshardware.com/howto/02q4/021021/powersupplies-15.html"
==================end=================

Once we get it clear what you want me to do and have it assembled in one
document, then I'll do it. When we get through with the use of the digital
multimeter and if we have found nothing, then we'll proceed to the next step.
Can we agree on that? 

-- 
              Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
                  (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
                   Obz Site:  39° 15' 7" N, 121° 2' 32" W, 2700 feet
              "I know that defies the law of gravity, but, you see, I never
               studied the law of gravity." -- Bugs Bunny
                         Web Page: <home.earthlink.net/~mtnviews>

Relevant Pages

  • Re: Dead motherboard?
    ... Green wire voltage is too low when powered off. ... However grey wire voltage never tells motherboard that voltage is ...
    (microsoft.public.windowsxp.perform_maintain)
  • Re: P5Q wont boot after long stable service
    ... The motherboard grounds PS_ON#, when it wants to turn on the power supply. ... (The voltage should be the same ... each pin is crimped to the wire. ... from the meter to the motherboard. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: P5Q wont boot after long stable service
    ... The motherboard grounds PS_ON#, when it wants to turn on the power supply. ... (The voltage should be the same ... each pin is crimped to the wire. ... from the meter to the motherboard. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: CPU Core Voltage Too Low -> Crash? - Important Addition
    ... > it to always 'stay on top' so I could see the CPU Core Voltage at all times. ... > is on the motherboard - hence the conclusion is that the motherboard is the ... This particular motherboard ALWAYS gives a low reading for the +12V and +5V ... Open Device Manager and disable any hardware you've temporarily removed. ...
    (microsoft.public.windowsxp.help_and_support)
  • Re: CPU Core Voltage Too Low -> Crash? - Important Addition
    ... > it to always 'stay on top' so I could see the CPU Core Voltage at all times. ... > is on the motherboard - hence the conclusion is that the motherboard is the ... This particular motherboard ALWAYS gives a low reading for the +12V and +5V ... Open Device Manager and disable any hardware you've temporarily removed. ...
    (microsoft.public.windowsxp.general)