Triple failure. Software? Hardware? Please help

Paul wrote:
Bill in Co wrote:
Ken Springer wrote:
On 7/7/11 9:38 AM, Bob F wrote:
GlowingBlueMist wrote:
A low battery can keep a system from booting. Fortunately it is
easier to swap the battery than it is to determine if it really needs
replacing.
Not if you have a voltmeter. If it reads 3V, it's fine.


It's always smart, IMO, to check a battery while it's under a load, not
just sitting out on your desk. :-)

True enough! Sometimes the open circuit voltage can be close to
normal,
but it's still bad (when tested under load, due to excessive internal
resistance in the cells).

Don't tell anyone, but I sometimes test these dinky batteries directly
(but
briefly!) with an ammeter (i.e., for their short circuit current value).

Caveat: I said dinky battery. Don't even think about trying this on a
car
battery (at least not with most meters). :-)


But the load on the CMOS cell is 10 microamps, meaning
the open circuit voltage is very close to the voltage when
a 10 microamp load is present. A load test is unnecessary.
Just using a multimeter reading, and the curve from the datasheet,
is enough to predict lifespan.

Even with a high internal cell resistance, 10 microamps times
that resistance is nothing.

As a matter of fact, there is a 1K ohm resistor in series
with the CMOS coin cell, as a current limiter. That should give
you some idea how little they care about cell resistance. A
ceramic cap is placed at the end of the circuit, where it
joins to the Southbridge, to provide good transient response,
which happens mainly when the computer is running. There are
next to no transients when the battery is being called on
to deliver the 10 microamps.

Paul

Well, maybe you're right for testing these CMOS/BIOS batteries. 10 ua is
practically nothing, I agree (and it would take a really high internal
resistance to drop much of anything, as you said).

But for flashlight batteries, and the like, I find the load test often
necessary. :-)
But I stand corrected on the former. Thanks. :-)

On Thu, 7 Jul 2011 08:38:00 -0700, "Bob F"
wrote:

GlowingBlueMist wrote:
A low battery can keep a system from booting. Fortunately it is
easier to swap the battery than it is to determine if it really needs
replacing.

Not if you have a voltmeter. If it reads 3V, it's fine.

Yes, my battery does read 3V. Easy to measure, since any part of the
the case is the ground, and the battery face is pretty big.


3V was under the tiny load of the CMOS, but if any of you think I
should, I'll measure again while the computer is on. I thought the
load was no greater then, since at that point, the power supply powers
everything.

A mobo battery doesn 't face the load that many batteries do. In fact
it's not much more than a digital watch to keep the clock ticking and
a little more data saved than in a watch. But less load than a
digital watch because it's not used to display the time, change the
time display, or do the alarm beeping. And the battery is several
times as big as most digital watch batteries.

micky wrote:
On Thu, 7 Jul 2011 08:38:00 -0700, "Bob F"
wrote:

GlowingBlueMist wrote:
A low battery can keep a system from booting. Fortunately it is
easier to swap the battery than it is to determine if it really needs
replacing.

Not if you have a voltmeter. If it reads 3V, it's fine.

Yes, my battery does read 3V. Easy to measure, since any part of the
the case is the ground, and the battery face is pretty big.


3V was under the tiny load of the CMOS, but if any of you think I
should, I'll measure again while the computer is on. I thought the
load was no greater then, since at that point, the power supply powers
everything.

A mobo battery doesn 't face the load that many batteries do. In fact
it's not much more than a digital watch to keep the clock ticking and
a little more data saved than in a watch. But less load than a
digital watch because it's not used to display the time, change the
time display, or do the alarm beeping. And the battery is several
times as big as most digital watch batteries.

I'd be surprised if it reads any different in circuit, in that case.

If it did, it could indicate an excessive leakage problem somewhere in the
circuit connected to the battery.

On Tue, 05 Jul 2011 12:51:01 -0700, "Ken Blake, MVP"
wrote:

On Mon, 04 Jul 2011 22:39:32 -0400, micky
wrote:

On Mon, 04 Jul 2011 09:19:13 -0500, GlowingBlueMist
wrote:


One other thing worth mentioning is that the life of the motherboard
battery is usually only rated for 5 years from date of manufacture. I

I think I've replaced it once, but that could be 5 years ago already!

have seen many strange things happen when one finally dies or gets too
low to fully drive the motherboard, including symptoms like you describe.

Does it need the mobo battery when the computer is plugged and
running?


No. It not only doesn't need it, it doesn't use it. That's the reason
why I've posted the following many times:

Before anyone whose clock is running slow rushes out to buy a new
battery, he should first take note of whether he is losing time while
the computer is running or while it's powered off. If it's while
powered off, the problem *is* very likely the battery. But if it's
while running, it can *not* be the battery, because the battery isn't
used while the computer is running.

Good to know.

I havent' had any problems with computer batteries, unless you count
twice in 27 years that the battery got used up and had to be replaced
(Actually iirc there was no battery in the PCJr and one had to put the
date and time in every time you logged in.)

But I do have major problems with my DVDR-HDD. It needs to know what
time it is to start recording at the right time, yet the time is wrong
soon after it's set. Is this because there's a bad piece of
quartz in the built in clock and if, when this breaks, I buy another
of the same model, will I likely get a better clock?

It has three settings for "Automatic Time Set"
Off
Automatic
Manual.

Whjen it's on Off, the time runs slow (or maybe fast) and the
recording starts late (or ends early**)

When it's on Auto, I think the time jumps around by a minute or two,
and even when it's just "jumped" its not very likely to be right.

When it's on Manual, I guess that means that's Manual Time Set, not
automatic Time Set like the whole box is called., the next screen lets
me set it. I don't want to have set it every few days. I don't even
wear my watch to the bedroom and no clock in the room has a second
indicator.


Because the time is usually wrong, I set it to record starting a
minute or two early and ending a minute late. But if I want to record
or watch something in time period before or after that, it can't
record both at the same time, so I lose part of one show, and when the
clock is wrong in the other direction, I lose part of the other.
Ugh.

On Tue, 05 Jul 2011 04:31:13 -0400, Paul wrote:


Both kinds of readings are valuable. Your low readings, could
indicate that the computer needs better power balancing (moving
some loads to another cable). If you can arrange to measure
a "no load" voltage, that would tell us how stressed the
supply is, inside the PSU casing.

Those numbers I gave you, 10.9 for one 12 volt pin and 11.5 for the
other** were indeed no load readings. (All the other readings were
very close to nominal, 3 and 5.)

**I didn't note positive or negative. I assumed those were correct)

But an A7M266 system, shouldn't be loading the 12V much.
The processor runs from the 5V rail, at least on some
of them. The single 12V wire on the main power cable,
would be barely enough to run Vcore, if they did that.
So +5V is a more likely source, if no ATX12V cable is
present.

Thanks. I have a speaker now. no beeps at all!!


Perhaps you'll hear something when you test another power
supply in there. The supply you swap in, must have a decent
+5V ampere rating.

Well, a friend gave me a bunch of PSUs, and the first two were 235
watts instead of the 275 Silencer (PC Power and Cooling) that I've
been using, but I figured I should test them before installing them.
First one tested dead and then the other. I don't mind him giving me
his junk, but I'm surprised he let it sit around his own apartment
without marking it as bad.

The next one I tool out of a young friend's junk computer and all the
pin voltages were wrong.** Not dead but worng. One was over 12v,
16v iirc.

Finally I found one from the first friend that was still in the
wrapping bag. Also a Silencer, Brand new, and 275 watts.

I tested it and the votlages were all good except the 12 volt pins
were actually a little lower than the one I removed!!!

Rather than install it, I just plugged in the 20-pin conector and
ignored alll the plugs to the CDs, the floppies, harddrrive,, the
extra fan.

That should be okay, shouldn't it?

Turned it on, green light goes on, red light goes on, and still
displays nothing on the monitor.

Doesn't beep either. (I may have unplugged the USB keyboard, but I
don't remember right now.)


Question about the AGP video card in the AGP slot. At the bottom
corner of the card, the blank circuit board has an L-shaped part that
I thought was supposed to hook onto a tab at the end of the slot.
But it seems to just slide into the slot at the same time all the
contacts and the rest of the board slide into the slot. I couldn't
manage to hook it onto anything. I installed ti right, didn't I?


**I don't remember if it was a Dell, but I guess so. I'm going to
call him to make sure. No indication on the Delta Electronics case
that it was meant for a Dell!! .DPS-200PB-101
With a suffix F it's used in a Compaq

On 7/8/11 2:37 AM, micky wrote:

3V was under the tiny load of the CMOS, but if any of you think I
should, I'll measure again while the computer is on.

Sort of out of the box here, but if you've got a computer that is
working right, switch batteries.

But, unless you're also looking for the knowledge, by now you could have
bought a new battery and known for sure. :-)

--
Ken

Mac OS X 10.6.8
Firefox 5.0
Thunderbird 3.1.11
LibreOffice 3.3.2

On Tue, 05 Jul 2011 16:27:55 -0400, Paul wrote:


The simplest alternative, if you own a multimeter, is to take
a reading off the top surface of the coin cell battery, with
respect to chassis ground. You can pick up a ground connection,
using a metal screw on an I/O connector in the I/O place area
of the computer. I clip on there with one lead of the
multimeter, and then use the red lead to make a voltage
reading from the top of the coin cell. Slightly above +3V,

Mine was exactly 3. IIRC. Two days ago and I'm starting to forget
details.

is a good battery. Below 2.3V is a bad battery. The "knee"
of the battery is relatively sharp, so if the battery is
"on the decline" and near end of life, it'll be flat after
three or four weeks or so.

If you don't own a multimeter, you can remove the battery
and take it to your local Radio Shack. And they can test it
with a meter. But removing the battery is just a PITA,

Yes it is. Amazingly, some times the battery learns to fly.

and I'd just replace it on the spot. If I was going to
go to the trouble of getting it out of that damn socket,
I'd want to resolve the issue immediately, instead of
wasting the gas on a trip to Radio Shack for a test.

Thanks.

micky wrote:
On Tue, 05 Jul 2011 04:31:13 -0400, Paul wrote:

Both kinds of readings are valuable. Your low readings, could
indicate that the computer needs better power balancing (moving
some loads to another cable). If you can arrange to measure
a "no load" voltage, that would tell us how stressed the
supply is, inside the PSU casing.

Those numbers I gave you, 10.9 for one 12 volt pin and 11.5 for the
other** were indeed no load readings. (All the other readings were
very close to nominal, 3 and 5.)

**I didn't note positive or negative. I assumed those were correct)

But an A7M266 system, shouldn't be loading the 12V much.
The processor runs from the 5V rail, at least on some
of them. The single 12V wire on the main power cable,
would be barely enough to run Vcore, if they did that.
So +5V is a more likely source, if no ATX12V cable is
present.

Thanks. I have a speaker now. no beeps at all!!

Perhaps you'll hear something when you test another power
supply in there. The supply you swap in, must have a decent
+5V ampere rating.

Well, a friend gave me a bunch of PSUs, and the first two were 235
watts instead of the 275 Silencer (PC Power and Cooling) that I've
been using, but I figured I should test them before installing them.
First one tested dead and then the other. I don't mind him giving me
his junk, but I'm surprised he let it sit around his own apartment
without marking it as bad.

The next one I tool out of a young friend's junk computer and all the
pin voltages were wrong.** Not dead but worng. One was over 12v,
16v iirc.

Finally I found one from the first friend that was still in the
wrapping bag. Also a Silencer, Brand new, and 275 watts.

I tested it and the votlages were all good except the 12 volt pins
were actually a little lower than the one I removed!!!

Rather than install it, I just plugged in the 20-pin conector and
ignored alll the plugs to the CDs, the floppies, harddrrive,, the
extra fan.

That should be okay, shouldn't it?

Turned it on, green light goes on, red light goes on, and still
displays nothing on the monitor.

Doesn't beep either. (I may have unplugged the USB keyboard, but I
don't remember right now.)


Question about the AGP video card in the AGP slot. At the bottom
corner of the card, the blank circuit board has an L-shaped part that
I thought was supposed to hook onto a tab at the end of the slot.
But it seems to just slide into the slot at the same time all the
contacts and the rest of the board slide into the slot. I couldn't
manage to hook it onto anything. I installed ti right, didn't I?


**I don't remember if it was a Dell, but I guess so. I'm going to
call him to make sure. No indication on the Delta Electronics case
that it was meant for a Dell!! .DPS-200PB-101
With a suffix F it's used in a Compaq

One word of warning, is Dell computers had a period of product years,
where the wiring on the main power connector was non standard.

If you're going to pick through a junk pile, it's a good idea
to verify wire color. As long as the manufacturers of the supplies
stick to the standard color scheme (i.e yellow = 12V), then you can
detect a "Dell specific" supply from an "ATX standard" supply. For
a reference on color codes to expect, you can use an ATX spec.

http://www.formfactors.org/developer...X12V_1_3dg.pdf (page 30)

You would at least, compare the wire color of your (previously working)
ATX supply, to any new supply you want to test out.

In terms of power rating, you want to check the label on the supply,
and see if it has adequate +5V amps. On my Nforce2 Athlon system,
that number was around 25 amps, which left a bit of room for
the +5V current my video card was using. If those 250 or 275 watt
supplies have relatively low 5V ratings, they might not be the
best choice for a substitute.

*******

To complicate matters, supplies come in two types. One type is the
"honest" type. The label contains both maximum and minimum current
ratings. The "minimum" value tells you that the supply may not
regulate the rail properly, under "no load" conditions. For
example, if you have a supply with a 12V 1A minimum spec, then
if you disconnect all 12V loads, the output won't read 12V. No
damage should result, but the +/-5% spec in the ATX standard
may be violated. That might be viewed as an admission they
don't meet crossloading specs, and by putting the information
on the label, there should be no surprises.

The other kind of supplies, only have maximum current values
listed. With those, we *assume* the supply stays in regulation
all the way down to zero.

When probing your existing power supply, you want to do as follows.

ATX ------- +5V/+12V ------ disk#1 ----- disk#2 ----- Disk#3
Supply ------- +5V/+12V --- Cable with no loads connected

You take a reading off both cables. The "cable with no loads connected"
gives a "true" reading of current output voltage. Taking a reading
off the end of the first cable, would show the impacts of R * I
voltage drop.

If you read the second cable, and it reads 12V, then you check the first
cable, and it reads 11V, you might conclude from that, that there
is too much resistance in the cabling to support that amount of
peripheral loading.

If both cables read 10.6, especially the no-load cable, then you might
conclude the power supply is out of spec. If the offered load is
below the "minimum" current specified on the label, then you
have your explanation. 10.6 volts is too low to run a hard drive,
so the hard drive may refuse to spin up, or may spin up and then
spin down, over and over again.

A standard supply works roughly as follows. There is a transformer
in the supply. The primary side has high voltage chopped DC on it.
The output voltage is established by "turns ratio" of the transformer
windings. Rectifiers at the output, convert the AC from the
transformer, back into DC.

Core
||
Primary_winding ||3.3V winding
||
||5V winding
||
||12V winding

The power supply observes the outputs, and tried to compensate for
loading. If we heavily load the +5V, the power supply immediately
notices this, and "turns up" the primary. All output voltages rise.
The 5V rises, until it's close to 5V again. The 12V has now
risen above the 12V value, and might be 12.6V. You can actually
tell how the supply is being loaded, by watching the direction
the output rails are moving in. If I saw the five volt rail at
4.8V, and the twelve volt rail at 12.6V, then I conclude from
that, that the five rail is "heavily loaded".

*******

PC Power and Cooling, did make replacement supplies for Dell computers,
which means some of their products will have the "alternate wire
color" pattern and can't be used on your computer. The documentation
may have been removed from their site, as to which models were
Dell substitutes.

*******

You can use the main power connector only, if you want. As long
as the motherboard places some loading on each rail, it should
stay in spec enough to be usable. Your 12V rail will still be
used to power fans.

*******

The L-shaped AGP "heel" has two potential functions. It
can be used to keep the card in the slot, as in this example.
This shows the "slide lock" style retainer, on the AGP slot.
But there was also the spring-loaded side button heel lock,
which is more annoying to disengage.

http://www.interfacebus.com/agp-expansion-card.png

But in an AGP Pro motherboard, the heel would fit right into
the slot, where some extra power pins are located. The heel
has no electrical contacts on this, so it doesn't matter.
The heel may provide some mechanical support in that case,
but not be as secure as a card with a heel lock. AGP Pro
adds a section before and after the regular slot area.

http://www.motherboards.org/files/te.../apro_r11a.pdf (Fig.6 page 12)

If you own an AGP Pro card, and have an AGP Pro motherboard,
then the heel area has electrical contacts on it. Not too
many people own AGP Pro cards.

http://upload.wikimedia.org/wikipedi...ria_II_Pro.jpg

Paul

On Sun, 03 Jul 2011 03:58:50 -0400, Paul wrote:

micky wrote:
Triple failures? Hard Drive Data? Hard drive? Video Card? CPU?
Power Supply?

All these are good except, I think, the CPU.

I took the heat sink off and the thermal compound looks pretty bad.
10 or 20% of surface has no compound, on either the sink or the chip.

The rest is dry although I don't know if that matters.

CPU's of this vintage are pretty cheap now, under 20, some under 10,
shipping included.

I have 800 MHz now but I can get up to 1.1 GHz, which would be faster
I suppose.


The question is, The 800 MHz I had overheated, but does that mean I
shouldn't use a 1.1 GHz???

It will have new Arctic Silver Paste, and I will use a new heat sink
just like the old heat sink but clean.


Thanks.

"micky" wrote in message
...
On Sun, 03 Jul 2011 03:58:50 -0400, Paul wrote:

micky wrote:
Triple failures? Hard Drive Data? Hard drive? Video Card? CPU?
Power Supply?

All these are good except, I think, the CPU.

I took the heat sink off and the thermal compound looks pretty bad.
10 or 20% of surface has no compound, on either the sink or the chip.

The rest is dry although I don't know if that matters.

CPU's of this vintage are pretty cheap now, under 20, some under 10,
shipping included.

I have 800 MHz now but I can get up to 1.1 GHz, which would be faster
I suppose.


The question is, The 800 MHz I had overheated, but does that mean I
shouldn't use a 1.1 GHz???

It will have new Arctic Silver Paste, and I will use a new heat sink
just like the old heat sink but clean.

Should work OK with the 1.1 Mhz, just make sure that the fan works
properly. That might be why the 800 packed up.