I ponder filling in a blank spot by cpu with an inductor from the old
board. Leaving ic controller and vrms all the way they are, just
adding inductor in blank spot. Any advice...
the board is nearly faultless, needs something extra, inductor by
common sense analysis , is about all IO could come up with.
The demise included two vrms that looked like they were shaken off
the board...that means frequency unheard of (pun)...the inductors, as
there are 3, are supposed to help this.
any pro builder/repair advice? oem breaks once, there is an error, I
follow rules to the point of shiny motherboards coming out 23900 hours
later...trying to get another long run, and thinking of extras,
because I have a complete other board identical, I can add things,
spare caps etc. Without question, I found four safe spots to add four
more caps that oe took away for reasons not important enough. The
inductor needs more advice.
First off, there are likely multiple versions and names of boards
based on "6728". This is just one of them. On this particular one,
there is a three phase Vcore regulator (one input inductor, three
output side inductors). There isn't any room here for more
components on Vcore.
http://intenseghetto.com/crap/865pe-neo2-fis2r-board.jpg
You say your old board is failed. And now you're getting
a (new to you) motherboard, which is presumably fully working.
Why not just use the new board the way it is, without mods ?
I recommend:
1) Find the part number of the regulator chip. It could be
the wider chip, which is located underneath the plastic
CPU retention frame. I can't see the number on the chip,
in the above photo. Sometimes you have to shine a light
at just the right angle, to be able to read the number.
2) Look up that part number on the web, and find the datasheet
for the part. Read all about how the circuit works, look
at the math equations that govern its operation. Then
decide whether you're competent to solder extra stuff
on the circuit.
Some multiphase regulator chips, have configuration pins
on the chip, which determine how many phases are running.
If some of the phases are turned off, then soldering
stuff to them won't be helping anything.
There are eight pin DIP chips, used to drive the heavy
capacitive load of the MOSFET gates. A driver chip must
be present for each working phase. Otherwise, there is
nothing to turn the MOSFETs on and off.
This is a sample of a datasheet for a Vcore regulator. It
will show you how complex the subject is. Analog Devices
ADP3180. A sample three phase regulator circuit is on page 11.
HTH,
Paul
> I have an older platinum 6728, and it overloaded on something and
>shut down. I am getting another, new, to replace it, it has been that
>good...
What evidence is there that it shut down? What exactly does
"shut down" mean? Generally, the more common shutdowns are
from excessive temperature of a monitored part like CPU,
sometimes northbridge. Shutdown from excessive VRM
temperature is also sometimes possible from the controller
IC overheating BUT, it is not so likely that adding an
inductor, IF that is what you are seeing as an empty spot,
would do much to resolve that.
>
> I ponder filling in a blank spot by cpu with an inductor from the old
>board. Leaving ic controller and vrms all the way they are, just
>adding inductor in blank spot. Any advice...
It would be very unusual for the board to have all the
stages in the VRM but only be lacking an inductor. I am
wondering if there are also the transistors missing in the
prior stage, OR this is not a spot for an inductor it is a
spot for a capacitor (most likely given past MSI designs
I've seen).
> the board is nearly faultless, needs something extra, inductor by
>common sense analysis , is about all IO could come up with.
... not sure to make of what you wrote, what is the power
requirement of the CPU and are you overclocking it?
If maximum power to the CPU is important for overclocking, a
different board would be the best choice, but if it is
shutting down I doubt that is the problem as mentioned
above.
> The demise included two vrms that looked like they were shaken off
>the board...that means frequency unheard of (pun)...the inductors, as
>there are 3, are supposed to help this.
More details of what happened might help, but if the
inductors were coming off I suspect a bad batch of boards
that weren't soldered properly, or a failure like shorted
mosfet or capacitor that caused enough heat to melt the
inductor solder, meaning having another inductor wasn't a
factor in addition to the other issues mentioned above.
> any pro builder/repair advice? oem breaks once, there is an error, I
>follow rules to the point of shiny motherboards coming out 23900 hours
>later...trying to get another long run, and thinking of extras,
>because I have a complete other board identical, I can add things,
>spare caps etc. Without question, I found four safe spots to add four
>more caps that oe took away for reasons not important enough. The
>inductor needs more advice.
Adding the capactors is a reasonable idea, they will
probably still be in the circuit as they are usually missing
on the final output stage of the VRM, but whether it is
useful depends on which CPU you use, whether you overclock
and how much, etc.
Generally speaking if these things are important a board
with more VRM stages and solid capacitors rather than wet
electrolytic would be helpful. It could be possible that
your last board had a random failure that is not caused by
the design used, just a particular bad specimen of component
used.
If your use seems to cause excessive heat in the VRM,
consider adding a fan. Since there is no way to mount it I
suggest making a bracket to screw the fan to, then either
mounting that on the rear screw holes of the exhaust fan or
drilling a couple new holes in the rear wall of the chassis
and using metal screws or machine bolts/nuts to attach it.
Besides making a bracket, a couple L-brackets from a
hardware store can sometimes work, it depends on the
dimensions present in the particular case you're using and
the heatsink too.