Return of "What is JS214A?"

[alpha_uma]

[Sorry for a long post.]

I'm looking for replacement parts for the MOSFET JS214A from an unknown
manufacturer.

I posted this question before early July this year, so the message is
already gone from the newsgroup server, but an archive of it may be found
here:
http://www.electronics-forum.net/components/What_is_JS_214A_91320.html. At
that time Franc Zabkar suggested that JS214A might be a clone of 2SJ214 by
Hitachi as suggested by the following site:
http://www.rf-china.com/dzqj/2sje.asp.

Since then, I've searched Hitachi's site and many other sites (including
NEC, Panasonic, Toshiba, Sanyo, etc). The amazing thing is that many of them
carry their own "2SJ series", but as far as I can tell, they don't let their
numbers overlap! I wonder if they have "carved" up all the 2SJ numbers into
commercial "districts".

Here is what I have found so far:

(a) No conclusive findings to confirm (or deny) if JS214A is the same as
Hitachi's 2SJ214.
(b) At Hitachi's site, there is no listing of 2SJ214 or 2SJ214A among its
many other 2SJ numbers. (Perhaps it has been discontinued. If so, I'd wish
they leave the listing there instead of just removing it.)
(c) According to International Rectifier (IRF.com), their cross reference
search lists three of their own MOSFETs as possible replacement for 2SJ214:
IRF9Z24NL, IRF9Z24NS, IRF9Z24S
(d) IRF.com's cross reference database DO suggest that 2SJ214 is made by
Hitachi.
(e) There is a big discrepancy between the values of I_D (max continuous
drain current) as reported by "rf-china.com" and "irf.com". A partial
listing (with rows shortened) of the numbers given by "rf-china.com" is as
follows (see
http://www.rf-china.com/dzqj/2sje.asp):

Model, Make, Uses, V_DSS, I_D
2SJ213, NEC, SW, -100V, +/-0.5A
2SJ214, Hitachi, Motor/Relay-D, -60, -105A
2SJ215, Hitachi, Motor/Relay-D, -60, -35A

The drain current of "105A" looks awfully high to be realistic! I wonder if
that is a typo or not. But since Hitachi's own site does not even have a
listing for 2SJ214, I have no way to check. In contrast, the replacement
MOSFET "IRF9Z24NL" as suggested by IRF has an I_D rating of only minus 12A
(with V_GS at -10V and T_C at 25degC).

As it stands, the following questions are still unanswered:
(1) Is JS214 the same (i.e., clone of) 2SJ214 or not?
(2) Who makes JS214A?
(3) Is/Was Hitachi the maker of 2SJ214 and 2SJ214A?
(4) What is the maximum (continuous) drain current of 2SJ214?

So if you have worked with either 2SJ214 or JS214, and happen to have the
documentation for 2SJ214 (or better yet, JS214), any information will be
much appreciated. Thanks

Al-U

 

[Watson A.Name - \"Watt Su]

"alpha_uma" <no_one@nonesuch.com> wrote in message
news:MlmRc.48061$M95.28354@pd7tw1no...

I'm looking for replacement parts for the MOSFET JS214A from an
unknown
manufacturer.

I posted this question before early July this year, so the message is
already gone from the newsgroup server, but an archive of it may be
found
here:

http://www.electronics-forum.net/components/What_is_JS_214A_91320.html.

At

that time Franc Zabkar suggested that JS214A might be a clone of
2SJ214 by
Hitachi as suggested by the following site:
http://www.rf-china.com/dzqj/2sje.asp.

Since then, I've searched Hitachi's site and many other sites
(including
NEC, Panasonic, Toshiba, Sanyo, etc). The amazing thing is that many
of them
carry their own "2SJ series", but as far as I can tell, they don't let
their
numbers overlap! I wonder if they have "carved" up all the 2SJ numbers
into
commercial "districts".

No, the numbers are each individually registered with a central
registry, just like JEDEC in the U.S. One company might register a
block of 2 or more number such as a series of the various devices in
that product line.

Here is what I have found so far:

(a) No conclusive findings to confirm (or deny) if JS214A is the same
as
Hitachi's 2SJ214.
(b) At Hitachi's site, there is no listing of 2SJ214 or 2SJ214A among
its
many other 2SJ numbers. (Perhaps it has been discontinued. If so, I'd
wish
they leave the listing there instead of just removing it.)

Hitachi is now Renesas, along with Mitsubishi, I believe. Check
Renesas.com.

Some surface mount power FETs, smaller than TO-220, believe it or not,
have currents of a hundred amps or more because they have very low
Rds(on).

It would be best to analyze the circuit to figure out what would
actually be needed as far as specs go.

(c) According to International Rectifier (IRF.com), their cross
reference
search lists three of their own MOSFETs as possible replacement for
2SJ214:
IRF9Z24NL, IRF9Z24NS, IRF9Z24S
(d) IRF.com's cross reference database DO suggest that 2SJ214 is made
by
Hitachi.
(e) There is a big discrepancy between the values of I_D (max
continuous
drain current) as reported by "rf-china.com" and "irf.com". A partial
listing (with rows shortened) of the numbers given by "rf-china.com"
is as
follows (see
http://www.rf-china.com/dzqj/2sje.asp):

Model, Make, Uses, V_DSS, I_D
2SJ213, NEC, SW, -100V, +/-0.5A
2SJ214, Hitachi, Motor/Relay-D, -60, -105A
2SJ215, Hitachi, Motor/Relay-D, -60, -35A

The drain current of "105A" looks awfully high to be realistic! I
wonder if
that is a typo or not. But since Hitachi's own site does not even have
a
listing for 2SJ214, I have no way to check. In contrast, the
replacement
MOSFET "IRF9Z24NL" as suggested by IRF has an I_D rating of only minus
12A
(with V_GS at -10V and T_C at 25degC).

As it stands, the following questions are still unanswered:
(1) Is JS214 the same (i.e., clone of) 2SJ214 or not?
(2) Who makes JS214A?
(3) Is/Was Hitachi the maker of 2SJ214 and 2SJ214A?
(4) What is the maximum (continuous) drain current of 2SJ214?

So if you have worked with either 2SJ214 or JS214, and happen to have
the
documentation for 2SJ214 (or better yet, JS214), any information will
be
much appreciated. Thanks

Al-U

 

[Franc Zabkar]

On Mon, 09 Aug 2004 00:04:29 GMT, "alpha_uma" <no_one@nonesuch.com>
put finger to keyboard and composed:

I want JS214A. But I have not been able to see any mention of this chip
anywhere when I "googled". The closest I got was "2SJ214" as suggested by
Franc Zabkar's tip.

Perhaps you have a fake MOSFET with a botched part number?

FWIW, my vrt databook lists the specs for a 2SJ214 MOSFET as:
P-channel, enhancement type, V-MOS, 60V, 10A, 40W

No cross reference is listed.

Sorry if this has been asked before, but what is your part's package
style? Can you see a logo?


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Franc Zabkar]

On Mon, 09 Aug 2004 10:19:27 GMT, "alpha_uma" <no_one@nonesuch.com>
put finger to keyboard and composed:

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:gg8eh01v3nn9b351lrsdj3v5stm5mul7i2@4ax.com...
snip

Perhaps you have a fake MOSFET with a botched part number?

OK, why couldn't they fake the entire part number as well, Looking for
a non-existent part number is not exactly what I would call fun.

FWIW, my vrt databook lists the specs for a 2SJ214 MOSFET as:
P-channel, enhancement type, V-MOS, 60V, 10A, 40W

Can you confirm that it is "10A" in your databook, not "105A" as listed at
the site
http://www.rf-china.com/dzqj/2sje.asp? If it is really 10A, then it is
great because it agrees with the (minus) 12A listed for IRF9Z24NL that
IRF.com gave as their cross-reference for 2SJ214.

It really is 10A.

No cross reference is listed.
What vrt databook are you using? Is there an online version?

The "vrt" databook is a paper version published by Tech Publications
Pte Ltd, Singapore.

If the JS214A
on my motherboard is really a "clone" of Hitachi's 2SJ214, then I would
simply go with IRF.com's suggested replacements. But the questions as to who
makes this JS214A and whether they are clones of 2SJ214 still remain.

Sorry if this has been asked before, but what is your part's package
style? Can you see a logo?

Unlike the other three NEC MOSFETs (K3296) which have "NEC" printed on them,
these other three MOSFETs do not have logos on them. But their package style
is just like those of the three NEC MOSFETs. The six ICs are all of the type
where the "fin" part of the IC is soldered onto the motherboard, and the
middle leg of the IC is clipped short. [It's in the style of what IRF.com
called their D^2 paks.] The only markings on the JS214A that I can see are
"JS214A B550S03" followed by the symbol of two opposing diodes.

This symbol could be that of a twin diode pack. The suffix "A" may
indicate that the diodes are joined at the anode, while a "C"
indicates common cathode. Is there a circuit reference on the
motherboard? For example, a "Dxxx" would indicate a diode, while a
"Qxxx" would indicate a transistor.

In any case, your device's package style is not like that pictured in
the vrt databook for the 2SJ214.

I'll take a picture of it tonight, and post a link to it so that I can show
you what it looks like. But my digital camera cannot take close-ups, so it
probably won't show the markings clearly.

See if you can identify which component(s) connect(s) to each pin of
your device. That may give a clue as to its function.


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Walter Harley]

"alpha_uma" <no_one@nonesuch.com> wrote in message
newscIRc.58166$M95.3182@pd7tw1no...

Unlike the other three NEC MOSFETs (K3296) which have "NEC" printed on
them,
these other three MOSFETs do not have logos on them. But their package
style
is just like those of the three NEC MOSFETs. The six ICs are all of the
type
where the "fin" part of the IC is soldered onto the motherboard, and the
middle leg of the IC is clipped short. [It's in the style of what IRF.com
called their D^2 paks.] The only markings on the JS214A that I can see
are
"JS214A B550S03" followed by the symbol of two opposing diodes.

Might be Shindengen? I think I've seen that logo. It shows up at
http://freespace.virgin.net/matt.waite/resource/logos/logo4.htm, but I think
I've seen a simpler version of the same logo elsewhere.

I didn't get any hits on their site in a quick search for JS214, but I
didn't try hard.

 

[Watson A.Name - \"Watt Su]

The only markings on the JS214A that I can see are
"JS214A B550S03" followed by the symbol of two opposing diodes.

Ah-HAH!! That tells you all you need to know. This is _not_ a MOSFET.
The two diodes mean that it is a dual rectifier in a transistor package.

I'll take a picture of it tonight, and post a link to it so that I can
show
you what it looks like. But my digital camera cannot take close-ups,
so it
probably won't show the markings clearly.

For closeups, put a magnifying glass in front of the lens. Then
autofocus should make it somewhat clear. You should be able to get
within a few inches of the lens with this method.

> Al-U

 

[alpha_uma]

"Walter Harley" <walterh@cafewalterNOSPAM.com> wrote in message
news:cf90fu$2m0$0$216.39.172.65@theriver.com...

"alpha_uma" <no_one@nonesuch.com> wrote in message
newscIRc.58166$M95.3182@pd7tw1no...
[...]
The only markings on the JS214A that I can see
are "JS214A B550S03" followed by the symbol
of two opposing diodes.

Might be Shindengen? I think I've seen that logo. It shows up at
http://freespace.virgin.net/matt.waite/resource/logos/logo4.htm, but I
think
I've seen a simpler version of the same logo elsewhere.

I didn't get any hits on their site in a quick search for JS214, but I
didn't try hard.

It could very well be! Thanks!

However, the markings on these three MOSFETs do NOT have the circle
enclosing the pair of diodes nor the capacitor symbol in the middle as
depicted in the logo at the link you gave above. Also, on my MOSFETs, there
is a visible "dot" connecting the opposing diode symbols. I'll try searching
for Shindengen, and see if I can find anything on them.

 

[Walter Harley]

"alpha_uma" <no_one@nonesuch.com> wrote in message
news:iC4Sc.68784$J06.7879@pd7tw2no...

OK, the markings on the three ICs actually looks like this:

JS 214A
B55QS03 (the earlier B550S03 was my typo)
[symbols of two opposing diodes]

Where these three ICs are located, they had printed the labels/legends
"D16", "D18" and "D19" on the motherboard.

So, you are saying that the symbols of two opposing diodes are not the
logo
(of Shindengen) as suggested by Walter. The D16, D18 and D19 legends
appear
to be confirming your suggestion that each of these three JS214 ICs is
actually "a dual rectifier in a transistor package".

Should I look for "JS 214" in the DIODE sections of databooks or somewhere
other than the MOSFET sections?

Personally, I'm skeptical, although Watson's probably looked at more old
components than I have. Firstly, I see no reason to believe that the
back-to-back-diode logo means the contents are diodes: after all, do you
know of any other components that are marked that way? The MOSFETs weren't
marked with a MOSFET logo, were they? And secondly, although it would be
commonplace to have three MOSFETs in parallel, it would be rather uncommon
to have three corresponding back-to-back diodes. Seems much more reasonable
to have three complementary MOSFETs. And you already know that there is
(/was) a 2SJ214 MOSFET; the problem is just that you can't find a datasheet
for it.

But if they are diodes, it's unlikely that they're all fried; diodes are
pretty tough critters, a lot harder to damage than MOSFETs. Why not
unsolder one or two of them and see if they behave like diodes? They would
probably be two diodes per package, with the case being the common anode or
common cathode.

Or, can you trace enough of the circuit to make an educated guess as to
function?

 

[Franc Zabkar]

On 10 Aug 2004 18:23:45 GMT, "Walter Harley"
<walterh@cafewalterNOSPAM.com> put finger to keyboard and composed:

Personally, I'm skeptical, although Watson's probably looked at more old
components than I have. Firstly, I see no reason to believe that the
back-to-back-diode logo means the contents are diodes: after all, do you
know of any other components that are marked that way?

Good point, but I've seen many twin diodes marked this way.

See http://www.wagner.net.au/Catalogue/01_19.pdf


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Walter Harley]

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:aldih01b5vi5k60rbjc5irnq2h2fe6com0@4ax.com...

Good point, but I've seen many twin diodes marked this way.

See http://www.wagner.net.au/Catalogue/01_19.pdf

Okay, you win

 

[Watson A.Name - \"Watt Su]

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:aldih01b5vi5k60rbjc5irnq2h2fe6com0@4ax.com...

On 10 Aug 2004 18:23:45 GMT, "Walter Harley"
walterh@cafewalterNOSPAM.com> put finger to keyboard and composed:

Personally, I'm skeptical, although Watson's probably looked at more
old
components than I have. Firstly, I see no reason to believe that the
back-to-back-diode logo means the contents are diodes: after all, do
you
know of any other components that are marked that way?

Good point, but I've seen many twin diodes marked this way.

See http://www.wagner.net.au/Catalogue/01_19.pdf

Here's one dual diode in the same packages that power transistors and
power FETs come in. I couldn't see the scrambled package, so the OP
will have to determine which package it's in.

http://www.st.com/stonline/books/pdf/docs/10182.pdf

My guess is that the JS214 is the date code - 14th week of 2002, and the
real key to the part number is the B55Q whatever.

- Franc Zabkar
--

 

[Franc Zabkar]

On Wed, 11 Aug 2004 00:59:36 -0700, "Watson A.Name - \"Watt Sun, the
Dark Remover\"" <NOSPAM@dslextreme.com> put finger to keyboard and
composed:

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:aldih01b5vi5k60rbjc5irnq2h2fe6com0@4ax.com...
On 10 Aug 2004 18:23:45 GMT, "Walter Harley"
walterh@cafewalterNOSPAM.com> put finger to keyboard and composed:

Personally, I'm skeptical, although Watson's probably looked at more
old
components than I have. Firstly, I see no reason to believe that the
back-to-back-diode logo means the contents are diodes: after all, do
you
know of any other components that are marked that way?

Good point, but I've seen many twin diodes marked this way.

See http://www.wagner.net.au/Catalogue/01_19.pdf

Here's one dual diode in the same packages that power transistors and
power FETs come in. I couldn't see the scrambled package, so the OP
will have to determine which package it's in.

http://www.st.com/stonline/books/pdf/docs/10182.pdf

My guess is that the JS214 is the date code - 14th week of 2002, and the
real key to the part number is the B55Q whatever.

I suspect the B55QS03 is a 55A device. I believe it is configured as a
flywheel diode in the Vcore regulator circuit(s).

IIRC, the OP's motherboard is the earlier K7S5A which uses a KA7500B
PWM controller.

http://www.overclockers.com.au/pic.php?pic=techstuff/r_ecs_k7s5a/k7s5a.jpg
http://www.fairchildsemi.com/ds/KA/KA7500B.pdf

In the OP's case, I suspect the Vcore circuit(s) look something like
this (part of an M571 socket 7 motherboard):

ferrite L20
beads FET coil
L18 |---| ===== to JP8
+5V o--|--[]-[]--|---------|Q9 |----|---/\/\/---|--o core jumper
+_|_ +_|_ |-|-| _|_ +_|_
EC18 ___ ___ EC22 | / \ Q8 ___ EC27
EC35 | 4000uF | 1500uF | /-|-\ diode | EC32 3500uF
EC36 _|_ _|_ | _|_ _|_ EC24
EC21 = = ___|___ = =
| |
|KA7500B|
|_______|

FWIW, my ECS L7S7A2 motherboard uses 3 pairs of MOSFETs (NEC 2SK3296 +
Philips PHB55N03LTA) rather than the traditional configuration of 3
MOSFETs plus 3 flywheel diodes. Coincidentally (?) the second MOSFET
part number is similar to the OP's twin diode. It is rated at 55A,
25V. The function of the MOSFET pairs is described in the following
datasheet.

RT9237A Multi-Phase DC/DC Controller for CPU Core Power Supply:
http://www.richtek.com/ch/count.asp?catalog=product&item=datasheet&id=70&file=../Product/Docs/DS9237A-00P.pdf


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Franc Zabkar]

On Wed, 11 Aug 2004 11:32:46 GMT, "alpha_uma" <no_one@nonesuch.com>
put finger to keyboard and composed:

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:5onjh013bsqmrb7skil752mkannf0enfuh@4ax.com...
On Wed, 11 Aug 2004 00:59:36 -0700, "Watson A.Name - \"Watt Sun, the
Dark Remover\"" <NOSPAM@dslextreme.com> put finger to keyboard and
composed:


"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:aldih01b5vi5k60rbjc5irnq2h2fe6com0@4ax.com...
On 10 Aug 2004 18:23:45 GMT, "Walter Harley"
walterh@cafewalterNOSPAM.com> put finger to keyboard and composed:

Personally, I'm skeptical, although Watson's probably looked at more
old
components than I have. Firstly, I see no reason to believe that the
back-to-back-diode logo means the contents are diodes: after all, do
you
know of any other components that are marked that way?

Good point, but I've seen many twin diodes marked this way.

See http://www.wagner.net.au/Catalogue/01_19.pdf

Here's one dual diode in the same packages that power transistors and
power FETs come in. I couldn't see the scrambled package, so the OP
will have to determine which package it's in.

http://www.st.com/stonline/books/pdf/docs/10182.pdf

My guess is that the JS214 is the date code - 14th week of 2002, and the
real key to the part number is the B55Q whatever.

I suspect the B55QS03 is a 55A device. I believe it is configured as a
flywheel diode in the Vcore regulator circuit(s).

IIRC, the OP's motherboard is the earlier K7S5A which uses a KA7500B
PWM controller.

http://www.overclockers.com.au/pic.php?pic=techstuff/r_ecs_k7s5a/k7s5a.jpg
http://www.fairchildsemi.com/ds/KA/KA7500B.pdf

In the OP's case, I suspect the Vcore circuit(s) look something like
this (part of an M571 socket 7 motherboard):

ferrite L20
beads FET coil
L18 |---| ===== to JP8
+5V o--|--[]-[]--|---------|Q9 |----|---/\/\/---|--o core jumper
+_|_ +_|_ |-|-| _|_ +_|_
EC18 ___ ___ EC22 | / \ Q8 ___ EC27
EC35 | 4000uF | 1500uF | /-|-\ diode | EC32 3500uF
EC36 _|_ _|_ | _|_ _|_ EC24
EC21 = = ___|___ = =
| |
|KA7500B|
|_______|

FWIW, my ECS L7S7A2 motherboard uses 3 pairs of MOSFETs (NEC 2SK3296 +
Philips PHB55N03LTA) rather than the traditional configuration of 3
MOSFETs plus 3 flywheel diodes. Coincidentally (?) the second MOSFET
part number is similar to the OP's twin diode. It is rated at 55A,
25V. The function of the MOSFET pairs is described in the following
datasheet.

RT9237A Multi-Phase DC/DC Controller for CPU Core Power Supply:

http://www.richtek.com/ch/count.asp?catalog=product&item=datasheet&id=70&file=../Product/Docs/DS9237A-00P.pdf



You guys are amazingly resourceful! Thanks for helping. I'm still digesting
the last few messages that you guys posted, especially the latest one from
Franc.

I was reading Fairchild's Application Note 558 titled "Introduction to Power
MOSFETs & Their Applications":

http://www.fairchildsemi.com/an/AN/AN-558.pdf#page=1

Many of the MOSFET driver techniques mentioned in the article employ diodes
and zener diodes. See, for example, Figures 22, 24 and 25 in that article.
So, I'm beginning to lean toward the suggestion that those three "JS214A"
ICs are some kind of dual diode in D^2-style paks.

Here is a link to a picture of my motherboard:
http://members.shaw.ca/ra11dec50/my_mosfets.jpg

OK, it's not a K7S5A, but the Vcore circuit looks similar.

Any suggestion of what kind of driver technique they might be using on my
motherboard?
Al-U

The circuit is probably similar to that in my diagram above, except
that yours probably has three MOSFET/diode/coil/cap sets wired in
parallel. Each of the twin diode packs are probably wired as a single
larger capacity diode by connecting together the two outside pins.
IIRC, your previous testing confirms this.

FWIW, here are my notes for an M571/M571lmr socket 7 motherboard which
uses a Vcore circuit similar to yours:
http://members.optusnet.com.au/~fzabkar/LMRvcore.txt
http://members.optusnet.com.au/~fzabkar/pwrnotes.txt
http://members.optusnet.com.au/~fzabkar/v32corev.txt
http://members.optusnet.com.au/~fzabkar/voltcalc.txt
http://members.optusnet.com.au/~fzabkar/voltcore.txt

The LMR board uses BIOS to control the Vcore (probably similar to
yours), the other uses jumpers.


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Franc Zabkar]

On Wed, 11 Aug 2004 11:32:46 GMT, "alpha_uma" <no_one@nonesuch.com>
put finger to keyboard and composed:

Here is a link to a picture of my motherboard:
http://members.shaw.ca/ra11dec50/my_mosfets.jpg

On 3/8/04 you wrote that "the three electrolytic caps around these six
MOSFETs are already confirmed bad out-of-circuit, and I replaced them
with good ones." However, in your picture the three caps still look
swollen, or is the poor focus playing tricks with my eyes?


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Watson A.Name - \"Watt Su]

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:4a2lh0let9lqr3dmmo3vintapn8rncmam9@4ax.com...

On Wed, 11 Aug 2004 11:32:46 GMT, "alpha_uma" <no_one@nonesuch.com
put finger to keyboard and composed:
[snip]

FWIW, here are my notes for an M571/M571lmr socket 7 motherboard which
uses a Vcore circuit similar to yours:
http://members.optusnet.com.au/~fzabkar/LMRvcore.txt
http://members.optusnet.com.au/~fzabkar/pwrnotes.txt
http://members.optusnet.com.au/~fzabkar/v32corev.txt
http://members.optusnet.com.au/~fzabkar/voltcalc.txt
http://members.optusnet.com.au/~fzabkar/voltcore.txt

The LMR board uses BIOS to control the Vcore (probably similar to
yours), the other uses jumpers.

This is all fine and well. But meddling with the Vcore means you're
presumably also going to meddle with the CPU or clock speed. And that's
a whole 'nother story.

- Franc Zabkar
--

 

[alpha_uma]

"Franc Zabkar" <fzabkar@optussnet.com.au> wrote in message
news:so3mh010m0e8u49k3s48fqfbemqsmd3d51@4ax.com...

On Wed, 11 Aug 2004 11:32:46 GMT, "alpha_uma" <no_one@nonesuch.com
put finger to keyboard and composed:

Here is a link to a picture of my motherboard:
http://members.shaw.ca/ra11dec50/my_mosfets.jpg

On 3/8/04 you wrote that "the three electrolytic caps around these six
MOSFETs are already confirmed bad out-of-circuit, and I replaced them
with good ones." However, in your picture the three caps still look
swollen, or is the poor focus playing tricks with my eyes?


- Franc Zabkar

Franc, you have good eyes. Those three 3300uF caps are indeed the swollen
ones. The picture was taken before I replaced them with good ones. In fact,
I also replaced the following caps even though they were not swollen:

(a) The group of four caps at the top of the picture. These were also 3300uF
caps, but they were later found to be "good" (meaning that they passed the
CRUDE testing routine with my DMM).

(b) A 1000uF cap near the ATX power socket. That is the one right below the
ATX power socket in the picture. It was found to be "good" later.

But these cap replacements did not solve the problem.

I thought about replacing the four green 1500uF caps (two of them next to
the memory socket on the right side of the picture and the other two right
above the AGP socket). That was when I started to question my basic
assumption that it was only a "bad caps" problem, and started suspecting the
MOSFETs and began posting on this newgroup and the "repair" newsgroup
seeking advice on MOSFET testing.

I think I will take a few days to read up on typical power regulation
techniques on motherboards (including those information you and others have
posted) to learn something. Meantime, I'm feeling adventurous, so I may
decide to go ahead and order half-a-dozen or so replacement 2SK3296 MOSFETs
(plus other stuffs to make up a minimum order) from DigiKey and see if I
have any luck.

Al-U