Nearly saw my 503+ go up in smoke

This group is so dead I thought I'd share a saga about my old hardware
reliable 503+ system.

I had some problems with my 503+ system that uses a k6-3+ at 550Mhz. I
thought it was the PS since the PS fan was locked up, so swapped in
another PS, which in fact was a better quality supply made by Delta.
Oddly, it was several days after installing the new PS that the system
crashed/locked, I reset it and walked away to the next room over to
work at another computer. The printer connected to the 503+ was not
accessible over the network, so I walked back into the room where the
503+ system lives, and omg, the smell of something burning permeated
the air.

I killed power and opened the case and did a sniff test, found the CPU
to smell the worst, so I touched the HSF, and it was hotter than a
firecracker. Further analysis revealed the Vcore PS circuit had 2
bulging electrolytic caps. OH NO me sez, thinking the CPU has just
fried. I cycled power just long enuf to get a BIOS boot screen, and
to my amazement, the system tried to run. I quickly power down, and
pulled the CPU, and started checking some voltages. Seems the CPU was
getting 3.8v, that's right, 3.8v, only 1.6 volts above the max spec
for the CPU. I surmise that the bad fan in the old PS may have
caused overheating in the box, and closer inspection revealed the
Vcore switch mode power supply circuit was having problems.

I did a little circuit probing, and replaced the swollen caps, and a
voltage test revealed the CPU socket was still getting 3.8v still (had
the CPU out of the socket during these tests just in case). The
signal from the PWM controller should have the MOSFET turned off, but
it was conducting lots of current to the CPU socket. Sure looked like
the MOSFET in the Vcore circuit had died (it's drain-source impedance
was quite low with no power on the system). I searched the net for
info on the device that failed, it is an n-channel enhanced mode
MOSFET made by Hitachi. I found the data sheet on the 8-pin Vcore PWM
controller chip (it sits right by the MOSFET device), and a typical
circuit schematic was provided by the manufacturer in its data sheet
detailing how such a CPU PS circuit should be built using a MOSFET
device. A brief survey of some old boards I keep around to cabbage
parts out of yielded a good quality Philips MOSFET device of similar
specs (I have to complement Abit in their AX7 design, I've stolen all
the PS caps out of this old socket7 board, and now I'm stealing the
CPU PS components and all were good name brand stuff). Philips has a
nice web site for obtaining data sheets on their semiconductor
components, I tip my hat to their IT department.

The worst part was getting the old MOSFET off the circuit board, some
trace damage occurred since the tab/body of the MOSFET device is
soldered to a large area copper trace on the mobo, and it's darn
difficult to get a enuf heat into such areas to reflow the solder and
separate the device from the mobo without special tools. Fortunately,
enuf of the copper pad remained (after I tore most of the copper off
beneath the MOSFET) to build up some solder onto for soldering to the
replacement MOSFET body tab, which was not actually designed for
surface mount, but bah who cares, a little jerry-rigging, lead
bending, and solder built up on the tab of the replacement device and
I had it attached in place. With fingers crossed, I hit the power
switch (had pulled all boards out of the mobo, and pulled the CPU) and
all was right with the world again as I measured 2.1v in the CPU
socket (I use 2.1v with my k63+, seems to be what it likes). I then
developed a big smile on my face cuz things are back where they should
be, then I drop the CPU and the boards back in, hook up all the
cables, and the old workhorse is back alive again.

I do recall a dead HS fan on an old AMD233 k6 system where I burned HS
fin stripes in my finger when I touched the HS, and it appears a k63+
can also take some serious overvoltage and heat abuse without failing.
It's so satisfying to fix a problem like this w/o even buying a single
component!

 

Hey, Kyle.

Wow. That is quite impressive. Congratulations on a job well done!
Your diagnostic and cannibalism skills are to be commended. I thought
it was too funny reading you rip apart that Abit AX7 (Athlon XP
motherboard) for parts so that you could save the 503+. The graphic
detail reminds me of some of the stories my plastic surgeon friend tells
me about his work. Not for the squeamish.

By the way, I looked at the "Mobile AMD-K6®-III+ Processor Data Sheet"
(
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/23535.pdf
) and saw what you mean about the absolute max Vcore of 2.2 V. I also
saw an absolute max Vio of 3.6 V, and an absolute max Vpin of (Vio + 0.4
V) and <= 3.8 V. The Vpin is the voltage on any I/O pin. It looks like
even with this highest absolute maximum, you were right there. Yikes!!!!!

You are right. It has been dead here for a while. I guess most people
have moved on to more interactive forums. Even so, that was quite an
entertaining read. Thanks!

--Alex

 

| Hey, Kyle.
|
| Wow. That is quite impressive. Congratulations on a job well done!
| Your diagnostic and cannibalism skills are to be commended. I
thought
| it was too funny reading you rip apart that Abit AX7 (Athlon XP
| motherboard) for parts so that you could save the 503+. The graphic
| detail reminds me of some of the stories my plastic surgeon friend
tells
| me about his work. Not for the squeamish.
|
| By the way, I looked at the "Mobile AMD-K6®-III+ Processor Data
Sheet"
| (
|
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/23535.pdf
| ) and saw what you mean about the absolute max Vcore of 2.2 V. I
also
| saw an absolute max Vio of 3.6 V, and an absolute max Vpin of (Vio +
0.4
| V) and <= 3.8 V. The Vpin is the voltage on any I/O pin. It looks
like
| even with this highest absolute maximum, you were right there.
Yikes!!!!!
|
| You are right. It has been dead here for a while. I guess most
people
| have moved on to more interactive forums. Even so, that was quite
an
| entertaining read. Thanks!
|
| --Alex
|
|
|
|
|

Hi Alex:

Glad someone got some reading pleasure out of my long-winded story. I
actually figured the bad part was no more than a $2 part if I had to
buy one, but it was more fun to find a decent replacement in some old
stock.

The real reason for even bothering with the system is it will probably
become the first puter for my daughter, who is nearing 2 years old.
She already has the "grab the mouse" impulse (hehe) so I can tell
she's got some "geek" in her blood, lmao. Just yesterday I let her
sit in my lap and watch me play a bit of Counterstrike (which is
probably not a good thing) and she kept shouting "more more more",
hahahaha. Then she grabs the mouse and gives the scroll wheel a dozen
or so turns and proceeds to pound on the keyboard. What a little
monkey, she does everything we do. Here's a pic for your enjoyment:

http://mywebpages.comcast.net/kylesb/abby-xmas2005.jpg

One small note, the AX7 was a socket 7 mobo that supported up to 233
MHz AMD and Pentium CPUs, in fact, the board lost it's floppy
controller at some point, so I put retired it to the spare parts
stack. It's not a socket A board. The AN7 is a socket A Abit board,
IIRC.

Cya!

 

(snip)

!
Your daughter will be overclocking it before long - and a toddlers
fingers should be small enough to get at those jumpers with ease!

Best wishes

 

Heh, playing Counterstrike with your daughter reminds me of one of my
friends. His daughter was probably about 4 years old, when one day at
nursery school, one of her classmates said something about being scared
of monsters. My friend's daughter told him, "Don't worry. If the
monster comes to get you, you can always kill it!"

I am not sure the teacher was too happy about that. Bah! What does she
know? Video games empower young minds to conquer their fears and the
unknown! (That's my story, and I'm sticking to it.)

Oh, yeah. My friend is a psychiatrist.

 

back in 98 this newsgroup was crazy, and very helpful when folks
were not flaming one another. it was my first taste of what a real
newsgroup was like, and I guess you are right forums have taken over.

good to see some folks still posting, keep it up

 

I have to second the "wow". I once found a Socket 7 mobo where a
bulging cap caused a MOSFET to get hot enough to turn purplish brown
and short its gate. But after replacing those parts, the new MOSFET
(taken from a PSU so bad that PC Power & Cooling used to use it as an
example of a bad PSU) ran at over 100C, even with a 2" x 4" chunk of
aluminum clamped to it with a clothespin. When I later gained access
to a scope I found that the Rockwell PWM chip that drove the MOSFET
had also been damaged, making its output pulses a couple of volts too
low. I was able to fix that by putting an open collecor TTL driver
between them.

I needed a 50W iron to melt the solder on the MOSFET tab, and even then
I had to add fresh solder to make it flow freely. It seems that mobo
manufacturers use solder with a higher melting point.

 

| I have to second the "wow". I once found a Socket 7 mobo where a
| bulging cap caused a MOSFET to get hot enough to turn purplish brown
| and short its gate. But after replacing those parts, the new MOSFET
| (taken from a PSU so bad that PC Power & Cooling used to use it as
an
| example of a bad PSU) ran at over 100C, even with a 2" x 4" chunk of
| aluminum clamped to it with a clothespin. When I later gained
access
| to a scope I found that the Rockwell PWM chip that drove the MOSFET
| had also been damaged, making its output pulses a couple of volts
too
| low. I was able to fix that by putting an open collecor TTL driver
| between them.
|
| I needed a 50W iron to melt the solder on the MOSFET tab, and even
then
| I had to add fresh solder to make it flow freely. It seems that
mobo
| manufacturers use solder with a higher melting point.
|

Aha, your story speak of a higher suffering to restore the mobo to
operational status.

I agree, I think the foreign solder is not the typical 60/40 lead/tin
mix solder, it reflows poorly. I could have borrowed the soldering
iron from hell (my neighbor owns it) with 400w power (he used it for
some massive purpose years ago) but opted to test my skills with a
lesser powered iron (think mine is 25w, it's a dual power jobbie 10
and 25). I should have tried adding solder to assist in the reflow,
failed to think of that in time before the damage was done.

What strikes me as scary is that the mobo traces are the only heatsink
for these surface mounted MOSFETs. The mobo that I cabbaged the
replacement MOSFET off of used a small stand-up HS with the power
devices in the Vcore and I/O voltage switch mode PS circuits. Guess
the need to cut costs to the bare minimum has prompted the designers
to take all kinds of shortcuts. I considered using the HS, but
figured there might be a height problem (the device is in-line with an
ISA slot) and if the old/dead MOSFET had no HS, then maybe the
replacement is OK w/o a HS. I tested the temp of the replacement
MOSFET by touching the device, and remarkably, it is barely warm to
the touch.

 

I haven't noticed much relationship between MOSFET temps and heatsinks,
maybe because MOSFETs come with different resistance ratings. For
example, my ECS K7VTA3 v. 8 with 1.6 GHz Duron has noticeably warmer
MOSFETs (no heatsinks) than my Asus/Asrock K7VT4A Pro does (also no
heatsinks), even though the latter's are in much smaller packages.
OTOH the Asrock has high quality Chemicon capacitors in the CPU voltage
regulator instead of the OSTs found in the ECS.