HP Pavilion 533w, pooping...

[Grinder]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

 

[DaveW]

Sorry, but it's history.

 

[Rod Speed]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9 leaky caps:

I'm a patient, if not terribly skilled, solderer. Can this be fixed,

Soldering multilayer boards isnt a trivial exercise.

 

[Paul]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

If there is no corrosion of the copper, to the point that a trace is
cut through, then it can be salvaged. The key thing is, did you pull
the board from the computer _before_ the motherboard failed to work ?
In other words, did you catch this problem, before there was
collateral damage to MOSFETs or toroidal inductors ? When a cap
fails to the point of overloading the MOSFET switching stage,
then the MOSFET gets damaged or the coil can get burned. If the cap
leaked but did not fail completely, then surgery can save
the patient.

The tricks with changing caps are:

1) A hot iron. Capacitors are soldered to large copper planes,
and those tend to suck the heat away when you are unsoldering
the caps. But, if you use a high power soldering iron, and
leave it in place for too long, you can delaminate the copper
from the fiberglass.

People may give you some fine ideas, like drilling out the
holes (no!!!) or the like, but resist the temptation. The
holes are plated through, and drilling can damage stuff you
cannot see (inner tracks next to the hole). You need a solder
sucker, a hot iron, and preferably, practice doing it on a
duff board first.

2) The ease of getting a capacitor lead, out of a hole on a
motherboard, depends on the diameter of the hole and the
diameter of the capacitor lead. At one of my former
employers, we used oversized holes, and I could never
figure out why we were doing it. But those oversized holes,
make it dead simple to remove caps for repairs.

At another place, we used to make the holes just a few
thou bigger than the capacitor lead. Which means, even
if there is no solder present in the hole, pulling the
cap out is a bitch. I expect you'll find more designs
like that, than the easy kind.

If you want to get someone else to do this work for you,
try "Homey". His web site appears to have closed up shop,
and his email would be "capman" in the domain of "att"
with a ".net" on the end. He would recap a motherboard
for somewhere around $50, including the caps. He has repaired
a ton of boards, but the bad cap era is probably not as
good a source of work as it used to be (too many upgraders).
It could also be that he is tired of doing it. There is
also a small bit of danger, as breathing the solder
smoke is known to cause lead poisoning. Techs who solder
for a living, get a blood test every three months, to
determine if there is too much lead present in their
blood stream. If the lead is found, it means they've
been careless and breathed too much of the smoke.
(Usually the bench will have a fan and filter, to
pull the solder resin smoke away from the work area,
so the solution is not exactly high tech.) While
repairing one motherboard won't kill you, try not
to inhale

(This is an archived page from Homey's web site)
http://web.archive.org/web/20030816063137/www.motherboardrepair.com/index.php?sec=services

Paul

 

[kony]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

Yes if the problem is only those caps, if the system had
worked up until the point where there was an instability
causing it to go down.

Some till try to suggest that it matters whether a
motherboard is multi-layered or not. It makes no difference
at all. These caps are all plated through-hole mounted and
respond to the same desoldering and resoldering measures as
any basic 2 layer (one per side) board would.

One significant issue is that at least one leg of each cap
will have a lot of copper on the board 'sinking away the
soldering iron heat so you should expect to use at least
40-50W iron for best results, but above 100W is too hot and
even at 100W you'd need a quick technique gained through
practice. You "might" be able to get the job done with a
30W iron with a stubby tip but it'll be more difficult.

Once you get some practice replacing caps it's pretty easy,
but to get practice... well you just have to roll up your
sleeves and do a few.

Make sure the replacement caps are particularly low, very
low ESR. You might buy a 10 pack of these,
http://www.digikey.com/scripts/DkSearch/dksus.dll?PName?Name=P12367-ND&Site=US
assuming they are 10mm diameter as they seem to be in the
picture. If they're only 8mm, then these instead,

http://www.digikey.com/scripts/DkSearch/dksus.dll?PName?Name=P12378-ND&Site=US

They may not be the same voltage or uF value but that is to
be expected to a certian extent, above all else the low ESR
is the important part and by using 16V rated parts you are
covering the 12V as well as the lower voltages, but if you
wanted lower voltage for the post-MOSFET caps, simply
consult the PDF parts list and pick the tallest cap of same
diameter in the 6V range. You'd then need two parts of less
than 10 pieces each though, raising the cost without clear
benefit.

 

[kony]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

You should also consider replacing all caps of the same make
and model, not just those that have already visibly failed.

 

[visions of effty]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

Heh! I threw away a Trigem Glendale the day before yesterday!

Small world. It would have never occured to me to that anyone would want
one, but of course I am wrong.

http://cgi.ebay.com/HP-Imperial-GL-...ryZ44943QQrdZ1QQcmdZViewItem#ebayphotohosting

There's one on ebay for $36 with no RAM or processor!!!

Lord, that's $36 too much.

~e.

 

[badgolferman]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that
has 9 leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed,
or is the mainboard most likely a frisbee at this point?

I am an electronic technician and skilled solderer. I have even gone
through Microminiature Soldering Techniques classes. One of the
hardest soldering jobs I did was removing a PCI connector from a
motherboard.

You can replace the capacitors but it will require a soldering iron
with thin tip and adjustable heat. If you had a Pace station with
vacuumed solder sucker it would be easier to do this job.

I recommend soldering wick with some liquid flux to soak the wick with.
That will make extraction of the old solder easier without damaging the
board with too much heat. One method you might try is to break the
capacitor off and grab hold of the exposed lead on one side with
needle-nose pliers then apply the iron to the opposite side of the
lead. That may be enough to allow you to yank it out. Then you can
clean the hole with the wick or solder sucker.

Those capacitors look like electrolytic capacitors and are likely
polarized. If you manage to get them out make sure you remember the
orientation to put the new ones back in properly.

 

[kony]

People may give you some fine ideas, like drilling out the
holes (no!!!) or the like, but resist the temptation. The
holes are plated through, and drilling can damage stuff you
cannot see (inner tracks next to the hole). You need a solder
sucker, a hot iron, and preferably, practice doing it on a
duff board first.

Actually drilling out holes works fine but it is a more
advanced way to do it and requires practice. The key is
that you are not trying to drill out the entire plated hole,
just a large enough hole in the center for the cap lead
which is about 0.023" or 0.6mm. That leaves a margin all
around the hole.

It also requires a very fine drill bit and great attention
to detail. Given the correct size bit and a centered
indentation to start the hole, it will not risk the inner
tracks... but it all goes back to attention to detail. Once
one has practice doing it, it is actually one of the safest
methods because it does not require continual heating of the
board as with desoldering wick, nor kickbacks from a sucker
(besides the issue that a sucker often does not get
sufficient solder out). A bulb type rather than
spring-loaded plunger type sucker may be a good first
attempt, but overall it seems more fiddly to me than just
using a drill bit, manually.

I seldom mention the drilling method primarily because it
does take more experience, the typical person doing repairs
probably does not have wire-sized bits around #73 to use.

2) The ease of getting a capacitor lead, out of a hole on a
motherboard, depends on the diameter of the hole and the
diameter of the capacitor lead. At one of my former
employers, we used oversized holes, and I could never
figure out why we were doing it. But those oversized holes,
make it dead simple to remove caps for repairs.

Most do use somewhat oversized holes, perhaps PCChips and
their relabels are among the closest to minimally sized
holes. The funny thing is they might be so frugal that they
do it merely to save $0.001 worth of solder.

At another place, we used to make the holes just a few
thou bigger than the capacitor lead. Which means, even
if there is no solder present in the hole, pulling the
cap out is a bitch. I expect you'll find more designs
like that, than the easy kind.

The vast majority of boards I've repaired didn't have holes
nearly that tight, as mentioned above a #73 bit for a ~
0.024" lead left a margin still.

 

[kony]

I am an electronic technician and skilled solderer. I have even gone
through Microminiature Soldering Techniques classes. One of the
hardest soldering jobs I did was removing a PCI connector from a
motherboard.

Make a template the size of the PCI connector leads. That
is, heat-resistant, hole in the middle over the leads area
and it serves as a shield to the rest of the board. Use a
hot air gun to simultaneously melt all of the slot connector
leads simultaneously while pulling it out - of course it
must be suspended on a parts vise to be able to access both
sides simultaneously.

Key is knowing your heat gun, how far away and how long it
will take to reach the melting point. Even more care is
needed now that solder has gone lead-free, as before it was
a fair bet what the melting point was but now at least a
little higher.

You can replace the capacitors but it will require a soldering iron
with thin tip and adjustable heat. If you had a Pace station with
vacuumed solder sucker it would be easier to do this job.

I recommend soldering wick with some liquid flux to soak the wick with.
That will make extraction of the old solder easier without damaging the
board with too much heat. One method you might try is to break the
capacitor off and grab hold of the exposed lead on one side with
needle-nose pliers then apply the iron to the opposite side of the
lead. That may be enough to allow you to yank it out. Then you can
clean the hole with the wick or solder sucker.

Have you ever tried breaking off large caps? I suggest it
may make a mess and should be avoided. In the worst cases
of an unskilled solderer, they can at least be very careful
to not touch the iron to the board and rock the cap out one
lead at a time, then proceed to clear out the hole. Trying
to wick or pump away solder while the cap is still installed
will never work well enough to make much difference, at most
it may reduce the amount of solder that is pulled out of the
hole on the front, component side of the board but once the
cap is out, that is fairly trivial to clean up with wick.

Those capacitors look like electrolytic capacitors and are likely
polarized. If you manage to get them out make sure you remember the
orientation to put the new ones back in properly.

The board is usually marked, or in worst case a picture can
be consulted.

 

[Grinder]

If there is no corrosion of the copper, to the point that a trace is
cut through, then it can be salvaged. The key thing is, did you pull
the board from the computer _before_ the motherboard failed to work ?

No, they watched the system degree for weeks, maybe months, then stopped
using it when the machine started failing a few seconds after startup.

In other words, did you catch this problem, before there was
collateral damage to MOSFETs or toroidal inductors ?

I would be out of my depth if I attempted to answer this. I can with
confidence say, though, that the drivers of this system did not bother
to stop when they heard metal on metal, as it were.

When a cap
fails to the point of overloading the MOSFET switching stage,
then the MOSFET gets damaged or the coil can get burned. If the cap
leaked but did not fail completely, then surgery can save
the patient.

Is is possible/probable that the CPU, a Celeron 2GHZ is fried as well?
When BIOS is running its checks, is it using that CPU such that we can
be optimistic about its state if some of those checks are being done?

The machine will fire up, show an HP logo screen, and is able to spot a
missing floppy drive, or reset the CMOS defaults, but the board will
invariably lose power a few seconds later.

The tricks with changing caps are:

[snip]

Thanks, I'll keep them in mind if I make the leap. Right now, I'm not
feeling courageous about experimenting on someone else's board.

If you want to get someone else to do this work for you,
try "Homey".

[snip]

Added to the list of options ... thanks.

 

[Grinder]

Heh! I threw away a Trigem Glendale the day before yesterday!

Small world. It would have never occured to me to that anyone would want
one, but of course I am wrong.

http://cgi.ebay.com/HP-Imperial-GL-...ryZ44943QQrdZ1QQcmdZViewItem#ebayphotohosting

There's one on ebay for $36 with no RAM or processor!!!

Lord, that's $36 too much.

It's already up to $102!

 

[Grinder]

I have an HP Pavilion 533w, with a Trigem Glendale mainboard, that has 9
leaky caps:

http://www.geocities.com/chastity403/trigem.jpg

I'm a patient, if not terribly skilled, solderer. Can this be fixed, or
is the mainboard most likely a frisbee at this point?

Ok, I'm crying uncle on this endeavor. Although I may take a stab at it
at some time in the future, I'm unwilling to flail away on someone
else's mainboard while they wait for a solution.

How difficult is it to find a board that will fit the existing case, and
processor? How can I tell if the Celeron 2GHZ is operational, or at
least what is your assessment of the odds?

Here's the specs on the existing board:

http://h10025.www1.hp.com/ewfrf/wc/genericDocument?lc=en&cc=us&dlc=en&docname=bph07813

I've specified the following bits at NewEgg trying to find a suitable
replacement:

CPU Socket Type: Socket 478
CPU Type: Intel Pentium 4/Celeron
Form Factor: Micro ATX
Onboard Video: Yes

That nets me 13 mainboards, most of which are under $50. (That beats
the thermal paste out of HP's offer to repair the system for $250.)

http://www.newegg.com/Product/Produ...&description=&srchInDesc=&minPrice=&maxPrice=

Although memory is relatively cheap, every dollar is going to count
here. Can I use my older (PC2100) memory in the boards that
predominately have a DDR standard of PC3200?

 

[kony]

No, they watched the system degree for weeks, maybe months, then stopped
using it when the machine started failing a few seconds after startup.

If it still tried to post, that would tend to indicate a
high probability it can be salvaged.

Is is possible/probable that the CPU, a Celeron 2GHZ is fried as well?
When BIOS is running its checks, is it using that CPU such that we can
be optimistic about its state if some of those checks are being done?

Yes the CPU is required operational to that extent, from
what you have written it seems most likely (though not
guaranteed) that the caps are the only problem. Pulling out
a board and handling, soldering it are additional stresses
themselves so there is even a potential for further damage
unrelated to the initial problem, but in general "IF" you
have a use for the board still it may be a good candidate
for cap replacement.

 

[kony]

Ok, I'm crying uncle on this endeavor. Although I may take a stab at it
at some time in the future, I'm unwilling to flail away on someone
else's mainboard while they wait for a solution.

How much of a rush are they in? I linked Digikey
previously, they usually get orders out the same day if you
place it early enough. Regardless, I'll address the rest of
this post.

How difficult is it to find a board that will fit the existing case, and
processor? How can I tell if the Celeron 2GHZ is operational, or at
least what is your assessment of the odds?

You'll likely want same chipset, or close, since you want
the integrated video. It'll also reduce the potential to
need reinstall Windows. However if the system has restore
CDs rather than full windows disc, there's a potential
software problem right off the bat.

Here's the specs on the existing board:

http://h10025.www1.hp.com/ewfrf/wc/genericDocument?lc=en&cc=us&dlc=en&docname=bph07813

I've specified the following bits at NewEgg trying to find a suitable
replacement:

CPU Socket Type: Socket 478
CPU Type: Intel Pentium 4/Celeron
Form Factor: Micro ATX
Onboard Video: Yes

That nets me 13 mainboards, most of which are under $50. (That beats
the thermal paste out of HP's offer to repair the system for $250.)

http://www.newegg.com/Product/Produ...&description=&srchInDesc=&minPrice=&maxPrice=

Although memory is relatively cheap, every dollar is going to count
here. Can I use my older (PC2100) memory in the boards that
predominately have a DDR standard of PC3200?

Yes, the memory requirement is dictated by the CPU you
use... it's FSB speed, and then if running async mode it's
+33 or (rarely more) above that. The memory you already
have is suitable for any system running that CPU.

Make note of whether the PSU is standard (we'd hope so but I
didn't check) as it applies to a new board being compatible
with the existing psu, and that the rear I/O shield for the
board pops out rather than being stamped into the rear wall
of the case (IF the rear port positions are different).

Beyond that, if the replacement board is wider, make sure
it will fit. This board you have currently is not a full
width mATX board, so if the replacement is wider be sure to
measure the available space in the case and the clearance
between items on the right-edge of the board and the
adjacent drive bays (or whatever else HP might have placed
in that region).

Now the next question, do you want to buy a new board to
support a 2GHz Celeron, or would now be a good time to
upgrade the CPU too since a celeron has little value?

 

[Grinder]

How much of a rush are they in? I linked Digikey
previously, they usually get orders out the same day if you
place it early enough. Regardless, I'll address the rest of
this post.




You'll likely want same chipset, or close, since you want
the integrated video. It'll also reduce the potential to
need reinstall Windows. However if the system has restore
CDs rather than full windows disc, there's a potential
software problem right off the bat.

I don't mind doing a fresh install. In fact, I'll probably recommend it
since the system has gone through four teenagers.

Make note of whether the PSU is standard (we'd hope so but I
didn't check) as it applies to a new board being compatible
with the existing psu, and that the rear I/O shield for the
board pops out rather than being stamped into the rear wall
of the case (IF the rear port positions are different).

The PSU is, I think, and ATX12V 1.x. It has 20-pin keyed connector, as
well as a 4-pin keyed connector and the slew of 4-pin molex connectors.
The I/O shield does pop out of the case.

Beyond that, if the replacement board is wider, make sure
it will fit. This board you have currently is not a full
width mATX board, so if the replacement is wider be sure to
measure the available space in the case and the clearance
between items on the right-edge of the board and the
adjacent drive bays (or whatever else HP might have placed
in that region).

That's a good point. I'll check it.

Now the next question, do you want to buy a new board to
support a 2GHz Celeron, or would now be a good time to
upgrade the CPU too since a celeron has little value?

The 2GHZ Celeron was satisfactory to the owner, so in the interest of
cost I think it should be kept. The only reason to upgrade would be if
the current processor is damaged. How can I assess the risk that the
CPU might have been harmed by the failure of the mainboard?

 

[Grinder]

Beyond that, if the replacement board is wider, make sure
it will fit. This board you have currently is not a full
width mATX board, so if the replacement is wider be sure to
measure the available space in the case and the clearance
between items on the right-edge of the board and the
adjacent drive bays (or whatever else HP might have placed
in that region).

Right-edge as in the picture I posted?

http://www.geocities.com/chastity403/trigem.jpg

ie, do you mean the side opposite the port cluster?

Assuming that's true, I have a health 8cm between that edge and the
front of the case, with a clearance of approximately 3cm. So, I
couldn't stand a board that had the memory slots much further to the
right. (The rightmost one is a little tight as it is close to the
cabling for the floppy drive.)

About 1.5cm off the right edge of my existing mainboard, there is a
couple of mounting points for a wider board, so it looks like the case
was designed to fit a µATX board. I'll work on some of the cheaper
examples from my list to see if the geometry is good.

 

[Grinder]

You'll likely want same chipset, or close, since you want
the integrated video. It'll also reduce the potential to
need reinstall Windows. However if the system has restore
CDs rather than full windows disc, there's a potential
software problem right off the bat.

Just a bit of clarification, please.

If I'm planning on making a fresh operating system install, is there any
advantage to getting the same chipset?

I don't understand why I would want the same (or similar) chipset if I
want integrated video. (The justification for going with integrated
video is to avoid the added cost of purchasing an additional video card.)

 

[Rod Speed]

kony wrote

Just a bit of clarification, please.

If I'm planning on making a fresh operating system install, is there any advantage to
getting the same chipset?
Nope.

I don't understand why I would want the same (or similar) chipset if I want integrated
video. (The justification for going with integrated
video is to avoid the added cost of purchasing an additional video card.)

Correct.

 

[Rod Speed]

I don't mind doing a fresh install. In fact, I'll probably recommend
it since the system has gone through four teenagers.


The PSU is, I think, and ATX12V 1.x. It has 20-pin keyed connector,
as well as a 4-pin keyed connector and the slew of 4-pin molex
connectors. The I/O shield does pop out of the case.


That's a good point. I'll check it.

The 2GHZ Celeron was satisfactory to the owner, so in the interest of
cost I think it should be kept. The only reason to upgrade would be
if the current processor is damaged. How can I assess the risk that
the CPU might have been harmed by the failure of the mainboard?

The risk is minimal that its got damaged.

The only real way to be sure is to put it into a known good motherboard.