Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computer Certification > A+ Certification > Re: PC will only boot sometimes?

Reply
Thread Tools

Re: PC will only boot sometimes?

 
 
w_tom
Guest
Posts: n/a
 
      04-20-2004
Until you get some basic numbers and fact, then you have not
a clue as to what to suspect. For example, immediately, what
does the Power Good signal from power supply report to
motherboard power supply controller. You did, of course,
start right off by measuring power supply voltages. Then
confirming both Power On and Power Good signals.

If you suspect memory, then, of course, you ran the
manufacturer's comprehensive diagnostics (provided free). And
you heated that memory with hairdryer on high so that the
intermittent failure would occur more frequently or even be
detected.

But why suspect memory? Memory is not even involved in the
BIOS startup.

In short, what did you do to first identify the problem.
Instead I read how you spent much time fixing things that were
not broken - wildly speculating as to what the problem might
be. First collect facts which is why your training noted the
importance of a 3.5 digit multimeter.

usp8riot wrote:
> I am going for my A+ certification this winter, still need some time.
> But I am repairing my Aunt's PC just for practice and her psu was
> fried. I replaced the psu but now the PC will only correctly boot
> every one in 7-8 boots. It boots other times but I get no screen.
> And when that happens everything else seems to be working but the
> one fan in the rear. I've tried it with 2 monitors, and 2 video
> cards, so it's not a display problem. It's strange. I'm thinking
> it's the memory but have no other compatible memory around to test.
> I'm thinking the psu may have fried something more when it went
> out, perhaps the cd drive because it won't open and the light
> stays on all the time. I just want to know what anyone else
> thinks about it.
>
> Thanks

 
Reply With Quote
 
 
 
 
w_tom
Guest
Posts: n/a
 
      04-21-2004
Take memory out. BIOS will startup just fine - and should
even report missing memory. One function a BIOS does only
after startup is to identify memory - to learn what kind of
memory is installed. It does all this without using any
stacks or using SIMM memory.

But if BIOS sometimes does not start; not even perform a
memory identification, then something else has failed. Memory
is not required until after BIOS startup and after other
things have executed. A PC that does not boot will not even
perform a basic memory check. Memory completely irrelevant to
original problem. His problem: sometimes will not boot - will
not even begin a BIOS startup. Memory and video are then
irrelevant. Stacks are irrelevant.

First things to check are power supply's Power Good and
Power On signals. Functions that must work even before BIOS
startup. Until power supply operation has been verified -
power supply and other components - then other testing is
relevant. Power supply and associated voltages must be well
in spec before even BIOS can execute properly, before memory
can run properly, and before video subsystem works properly.
IOW the first tool is a 3.5 digit multimeter and a voltage
chart for those limits.

Barry Watzman wrote:
> ...
> That is not true. Anyone who knows how a processor works knows
> that the stack must exist in memory (DRAM memory, not the BIOS),
> and that without a valid stack in valid memory, you won't get
> ANYTHING, in most cases not even the POST single beep.
>
> w_tom wrote:
>> Until you get some basic numbers and fact, then you have not
>> a clue as to what to suspect. For example, immediately, what
>> does the Power Good signal from power supply report to
>> motherboard power supply controller. You did, of course,
>> start right off by measuring power supply voltages. Then
>> confirming both Power On and Power Good signals.
>>
>> If you suspect memory, then, of course, you ran the
>> manufacturer's comprehensive diagnostics (provided free). And
>> you heated that memory with hairdryer on high so that the
>> intermittent failure would occur more frequently or even be
>> detected.
>>
>> But why suspect memory? Memory is not even involved in the
>> BIOS startup.
>>
>> In short, what did you do to first identify the problem.
>> Instead I read how you spent much time fixing things that were
>> not broken - wildly speculating as to what the problem might
>> be. First collect facts which is why your training noted the
>> importance of a 3.5 digit multimeter.

 
Reply With Quote
 
 
 
 
w_tom
Guest
Posts: n/a
 
      04-22-2004
All those other systems I designed and programmed don't
work? You are only speculating that a BIOS startup must use
subroutines. What happens when no subroutines are used in
BIOS startup? No problem. Welcome to a world long ago when
computer people wrote this stuff. Entire setup was simply
execution from ROM code - no read/write memory and no
subroutines - until after essential functions had been
verified and setup.

In the meantime, that is also what a BIOS does. Take out
the memory. BIOS will startup properly, identify missing
memory, and inform human accordingly with an error code.
Memory identification does not occur until after other basic
BIOS startup code has been executed.

In the PC, one needs no subroutine calls to put simple text
information on screen. Complex subroutines come later when
more complex video functions are enabled.


Barry Watzman wrote:
> You can't do subroutine calls without a stack (which is where
> the return address is stored), and you can't have a stack
> without RAM. And there's no way to get as far as video out
> without doing subroutine calls.

 
Reply With Quote
 
AG
Guest
Posts: n/a
 
      04-22-2004

"SBFan2000" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Every motherboard I've seen will not display anything on the screen if
> memory is missing, it will only give beep codes.
>

Me too. I've gotten a bad board to show error codes on a POST card but
that's all. I don't know anything about subroutines or such. I just know
what happens when I put bad RAM into a good board. And what happens is no
picture and beeps.
With no RAM in the board at all I just get beeps.
AG


> "w_tom" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
> > All those other systems I designed and programmed don't
> > work? You are only speculating that a BIOS startup must use
> > subroutines. What happens when no subroutines are used in
> > BIOS startup? No problem. Welcome to a world long ago when
> > computer people wrote this stuff. Entire setup was simply
> > execution from ROM code - no read/write memory and no
> > subroutines - until after essential functions had been
> > verified and setup.
> >
> > In the meantime, that is also what a BIOS does. Take out
> > the memory. BIOS will startup properly, identify missing
> > memory, and inform human accordingly with an error code.
> > Memory identification does not occur until after other basic
> > BIOS startup code has been executed.
> >
> > In the PC, one needs no subroutine calls to put simple text
> > information on screen. Complex subroutines come later when
> > more complex video functions are enabled.
> >
> >
> > Barry Watzman wrote:
> > > You can't do subroutine calls without a stack (which is where
> > > the return address is stored), and you can't have a stack
> > > without RAM. And there's no way to get as far as video out
> > > without doing subroutine calls.

>
>



 
Reply With Quote
 
w_tom
Guest
Posts: n/a
 
      04-22-2004
BIOS execution can be and has been done both ways - with or
without video message. If computer beeps an error code, then
BIOS startup code has executed - without subroutines. Beep
code means that BIOS is executing without any memory
installed. And that is the point. OP has a problem where
sometimes the BIOS startup code does not even execute. Trying
to fix it by changing memory? Do you now understand the
absurdity of that solution? He is trying to fix something
that is not even involved in his problem.

There is no complicated initialization for Display Mode 7
text display. Complication only occurs later. BIOS even
introduces itself early on - typically at end of BIOS startup
- because a simple text display is trivial. BIOS does not
require complex program with numerous subroutines to
initialize video controller. Video Controller's computer
initializes for trivial text display on powerup.

However back to the original point that is getting
overlooked due to speculation. BIOS startup code executes
regardless of whether DRAM is or is not installed. If BIOS
startup code does not execute (sometimes), then DRAM is not
reason for problem.

Does the A+ certification even bother to teach how BIOS
works?

Back to the OP's problem. Before BIOS even executes, a 3.5
digit multimeter reports basic signals such as Power On and
Power Good. Simple information that any trained A+ technician
would know. If those signals do not 'handshake' accordingly,
then we can suspect a problem in one of three computer
components (and definitely not involving memory). Nothing
complex here. The most difficult part appears to get one to
start looking at where problems may exist.

SBFan2000 wrote:
> Every motherboard I've seen will not display anything on the
> screen if memory is missing, it will only give beep codes.

 
Reply With Quote
 
AG
Guest
Posts: n/a
 
      04-23-2004
You have a point but there are only so many hours in a day. Is it necessary
for me to know subroutines to diagnose a bad motherboard? Is it really
going to make any difference in the way that I treat a malfunctioning
computer?
The truth is that 95% of the problems that I run up against in my job, which
I will admit is mostly techsupport for ISP customers, are software related.
I hand most of the hardware issues off to other local techs who have more
time to hand hold and hardware in stock.
The only hardware problem that I've had recently was a bad RAM stick and
that was easy to diagnose by looking up the beep code in my little blue
book.
If a motherboard is bad the truth is that I'm going to replace it. It
doesn't matter why it's bad, I'm not going to take the time to "fiddle" with
it, I'm going to run to the nearest store or get one shipped in next day air
if I'm only in a slight hurry.
If I'm working on something mission critical, something we try to avoid, I
don't have time to mess with subroutines or diagnosing exactly what part of
the machine is at fault. As soon as I know what is at fault I'm going to
replace it.
I could go on but understand, I do understand what you are saying. Too many
people don't know what causes problems and what they can do about it. Too
many people can't navigate in command line for instance and there are still
things that can only be done there.
AG
"Barry Watzman" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> In my view, the fact that "I don't know anything about subroutines or
> such" [and certainly that includes things like stacks, segement
> registers, etc.] is a serious problem in this industry.
>
> We are creating a an entire industry of "experts" who, very literally,
> don't know how a computer works [by that I mean, what goes on INSIDE the
> CPU itself].
>
> I started in computers in 1967, and in microcomputers in 1975 with an
> early IMSAI. We had to do our own programming in machine code (switches
> and lights] and Assembly language.
>
> I don't want to be so far out as to suggest a return to that. But, on
> the other hand, when certified experts literally don't know what a
> "stack" or subroutine call or segment register are, how they are used,
> and the various implications of all of this, then there is something
> wrong. [Many don't even know what a "register" is at all, in terms of
> the internals of a CPU.]
>
> This stuff can be taught fairly easily, it takes a few weeks to learn,
> but it really dramatically increases your understanding of this stuff
> (and your apprecitation of it's complexity, as well). It bothers me
> that it isn't taught or expected of ANY of the major certifications (A+,
> MCSE, any of the network certs. In my view, if you don't know this
> stuff, you can't legitimately call youself a computer expert, anymore
> than an auto mechanic could call himself an "expert" if he didn't know
> what a piston or valve were or did (and that's actually a pretty good
> analogy).
>
>
> AG wrote:
> > "SBFan2000" <(E-Mail Removed)> wrote in message
> > news:(E-Mail Removed)...
> >
> >>Every motherboard I've seen will not display anything on the screen if
> >>memory is missing, it will only give beep codes.
> >>

> >
> > Me too. I've gotten a bad board to show error codes on a POST card but
> > that's all. I don't know anything about subroutines or such. I just

know
> > what happens when I put bad RAM into a good board. And what happens is

no
> > picture and beeps.
> > With no RAM in the board at all I just get beeps.
> > AG
> >
> >
> >
> >>"w_tom" <(E-Mail Removed)> wrote in message
> >>news:(E-Mail Removed)...
> >>
> >>> All those other systems I designed and programmed don't
> >>>work? You are only speculating that a BIOS startup must use
> >>>subroutines. What happens when no subroutines are used in
> >>>BIOS startup? No problem. Welcome to a world long ago when
> >>>computer people wrote this stuff. Entire setup was simply
> >>>execution from ROM code - no read/write memory and no
> >>>subroutines - until after essential functions had been
> >>>verified and setup.
> >>>
> >>> In the meantime, that is also what a BIOS does. Take out
> >>>the memory. BIOS will startup properly, identify missing
> >>>memory, and inform human accordingly with an error code.
> >>>Memory identification does not occur until after other basic
> >>>BIOS startup code has been executed.
> >>>
> >>> In the PC, one needs no subroutine calls to put simple text
> >>>information on screen. Complex subroutines come later when
> >>>more complex video functions are enabled.
> >>>
> >>>
> >>>Barry Watzman wrote:
> >>>
> >>>>You can't do subroutine calls without a stack (which is where
> >>>>the return address is stored), and you can't have a stack
> >>>>without RAM. And there's no way to get as far as video out
> >>>>without doing subroutine calls.
> >>
> >>

> >
> >

>



 
Reply With Quote
 
w_tom
Guest
Posts: n/a
 
      04-23-2004
You are assuming that because video output is traditionally
done via interrupt (subroutine) calls, then such calls are
only way to output video. Wrong. I wrote just a program
that did video output with subroutine calls so that program
could even run faster on slower 386 and 486 processors.
Written this way for faster video displays. IBM did same
(without subroutines) in their AT BIOS. So yes, text can be
displayed on video without calling subroutines. Code even
demonstrated in video subsystems book.

Assumed was that only interrupt calls (subroutines) can
perform a video display. And yet is not true. Video
subsystem can be written to directly. Text video is easily
accomplished without calling subroutines as long as video is
only a text mode display.

But again returning to the OPs problem - if BIOS does not
even start execution, then DRAM has absolutely nothing to do
with the problem. Furthermore BIOS will execute even with no
RAM in SIMM sockets. It will complain, but it will execute.

Not necessary for a technician to learn the subroutines.
But it is necessary to learn what, for example, the BIOS does
before repairing things. More often, repair technicians only
know the shotgun method. Don't even know how to use or read a
meter.

That brings us back to the original post. He wanted to
blame something without taking basic information. He did not
even first verify system voltages are stable or in spec. He
wildly speculated without facts - and that is just too common
among computer repairmen.

A long list of replaced items is only a customer's problem
and expense. Instead, a technician should know enough about
the machine as to first collect basic facts and then know
exactly which part is (could be) defective - all before
replacing any parts. This usually only takes minutes.
Computers today are so easy to understand - which is why
shotgunning is not done by properly trained repairmen.

Barry Watzman wrote:
> I agree that you can do a beep without any RAM at all. But you
> can't do video output. Video output REQUIRES subroutine calls
> between the system bios on the motherboard and the video bios on
> the video card. Subroutien calls require a stack. A stack
> requires RAM. End of discussion.
>
> [There might be an exception for motherboards with video on the
> motherboard as part of the chipset. Doing video output without
> subroutine calls would be convoluted and almost impossibly difficult
> even in those cases, but there is a difference between "impossibly
> difficult" and physically impossible.]

 
Reply With Quote
 
Martin
Guest
Posts: n/a
 
      04-24-2004
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah
blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
blah blah blah


 
Reply With Quote
 
Tom MacIntyre
Guest
Posts: n/a
 
      04-24-2004
On Fri, 23 Apr 2004 09:50:08 -0500, "AG" <(E-Mail Removed)> wrote:

>You have a point but there are only so many hours in a day. Is it necessary
>for me to know subroutines to diagnose a bad motherboard? Is it really
>going to make any difference in the way that I treat a malfunctioning
>computer?
>The truth is that 95% of the problems that I run up against in my job, which
>I will admit is mostly techsupport for ISP customers, are software related.
>I hand most of the hardware issues off to other local techs who have more
>time to hand hold and hardware in stock.
>The only hardware problem that I've had recently was a bad RAM stick and
>that was easy to diagnose by looking up the beep code in my little blue
>book.
>If a motherboard is bad the truth is that I'm going to replace it. It
>doesn't matter why it's bad, I'm not going to take the time to "fiddle" with
>it, I'm going to run to the nearest store or get one shipped in next day air
>if I'm only in a slight hurry.
>If I'm working on something mission critical, something we try to avoid, I
>don't have time to mess with subroutines or diagnosing exactly what part of
>the machine is at fault. As soon as I know what is at fault I'm going to
>replace it.
>I could go on but understand, I do understand what you are saying. Too many
>people don't know what causes problems and what they can do about it. Too
>many people can't navigate in command line for instance and there are still
>things that can only be done there.
>AG


Good points. Getting the job done means getting it done. If you are
well-educated and can't produce, an uneducated producer is more
valuable than you are.

I have a diploma in electronics engineering technology. I worked for
17+ years in radio/TV broadcasting, and 5+ in consumer electronics
repair. I rarely had to use the true depth of my original knowledge in
either of those jobs, but it certainly went a long way in helping me
to understand, conceptualize, etc., in most situations that came up.
As well, when those situations arose that required more in-depth
knowledge (training on new equipment/technology, dealing with
manufacturers' tech support), it was good to be able to walk the walk
and talk the talk.

It all depends on what you need/want to be able to do. It depends on
how much time (we're talking years for high-level technological
education) you have, your family situation, and other factors. The big
thing is to get the job done. The one thing I do recommend for those
lesser-educated producers...if you're not sure why something worked,
do a de-brief, find out, ask the right person. Most positive results
have logical reasons why they happened.

Tom

>"Barry Watzman" <(E-Mail Removed)> wrote in message
>news:(E-Mail Removed)...
>> In my view, the fact that "I don't know anything about subroutines or
>> such" [and certainly that includes things like stacks, segement
>> registers, etc.] is a serious problem in this industry.
>>
>> We are creating a an entire industry of "experts" who, very literally,
>> don't know how a computer works [by that I mean, what goes on INSIDE the
>> CPU itself].
>>
>> I started in computers in 1967, and in microcomputers in 1975 with an
>> early IMSAI. We had to do our own programming in machine code (switches
>> and lights] and Assembly language.
>>
>> I don't want to be so far out as to suggest a return to that. But, on
>> the other hand, when certified experts literally don't know what a
>> "stack" or subroutine call or segment register are, how they are used,
>> and the various implications of all of this, then there is something
>> wrong. [Many don't even know what a "register" is at all, in terms of
>> the internals of a CPU.]
>>
>> This stuff can be taught fairly easily, it takes a few weeks to learn,
>> but it really dramatically increases your understanding of this stuff
>> (and your apprecitation of it's complexity, as well). It bothers me
>> that it isn't taught or expected of ANY of the major certifications (A+,
>> MCSE, any of the network certs. In my view, if you don't know this
>> stuff, you can't legitimately call youself a computer expert, anymore
>> than an auto mechanic could call himself an "expert" if he didn't know
>> what a piston or valve were or did (and that's actually a pretty good
>> analogy).
>>
>>
>> AG wrote:
>> > "SBFan2000" <(E-Mail Removed)> wrote in message
>> > news:(E-Mail Removed)...
>> >
>> >>Every motherboard I've seen will not display anything on the screen if
>> >>memory is missing, it will only give beep codes.
>> >>
>> >
>> > Me too. I've gotten a bad board to show error codes on a POST card but
>> > that's all. I don't know anything about subroutines or such. I just

>know
>> > what happens when I put bad RAM into a good board. And what happens is

>no
>> > picture and beeps.
>> > With no RAM in the board at all I just get beeps.
>> > AG
>> >
>> >
>> >
>> >>"w_tom" <(E-Mail Removed)> wrote in message
>> >>news:(E-Mail Removed)...
>> >>
>> >>> All those other systems I designed and programmed don't
>> >>>work? You are only speculating that a BIOS startup must use
>> >>>subroutines. What happens when no subroutines are used in
>> >>>BIOS startup? No problem. Welcome to a world long ago when
>> >>>computer people wrote this stuff. Entire setup was simply
>> >>>execution from ROM code - no read/write memory and no
>> >>>subroutines - until after essential functions had been
>> >>>verified and setup.
>> >>>
>> >>> In the meantime, that is also what a BIOS does. Take out
>> >>>the memory. BIOS will startup properly, identify missing
>> >>>memory, and inform human accordingly with an error code.
>> >>>Memory identification does not occur until after other basic
>> >>>BIOS startup code has been executed.
>> >>>
>> >>> In the PC, one needs no subroutine calls to put simple text
>> >>>information on screen. Complex subroutines come later when
>> >>>more complex video functions are enabled.
>> >>>
>> >>>
>> >>>Barry Watzman wrote:
>> >>>
>> >>>>You can't do subroutine calls without a stack (which is where
>> >>>>the return address is stored), and you can't have a stack
>> >>>>without RAM. And there's no way to get as far as video out
>> >>>>without doing subroutine calls.
>> >>
>> >>
>> >
>> >

>>

>


 
Reply With Quote
 
Tom MacIntyre
Guest
Posts: n/a
 
      04-24-2004
On Sat, 24 Apr 2004 09:28:53 GMT, "Martin" <(E-Mail Removed)>
wrote:

>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah
>blah blah blah
>


Jingle Bells, as sung by the Grinch?

Tom
 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
boot of "cisco2-C1600" using boot helper "eprom:c1600-boot-r.111-10.AA" failed Pete.Rudolph@gmail.com Cisco 2 06-14-2007 03:41 AM
2970 3750 3760 & Family What do these boot switches do in the boot loader: -n -j -p -c -f -i -m -notnd -notni -post -skip_post -reserve_mem -no_esp_keepalive mickrussom@yahoo.com Cisco 0 11-07-2005 10:23 PM
Multi-boot system; Win98SE won't boot Frank J. Fedel Computer Support 14 10-16-2003 02:55 AM
RAID + Dual Boot = Slow or No Boot mhicaoidh Computer Support 0 08-28-2003 03:15 AM
Boot error from copied boot partition Peter Fournia Computer Support 5 07-09-2003 03:00 PM



Advertisments