How to overclock Celery566 w/ MSI mobo?

~misfit~ wrote:
David Maynard wrote:

I take the direct approach and simply break the appropriate pin off
the
chip carrier but you're screwed if you go cross-eyed and break off the
wrong one, although P2B in here seems to have mastered some form of
voo-doo that enabled him to actually solder pins back on one of his.
Seems unlikely
to me that mere mortals should count on replicating that feat, though.


Hey! That was me who managed to solder a pin back on a CPU (Celly 600).
Unless P2B has done it as well. :-)
--
~misfit~

My apologies for inadvertently omitting you from the voo-doo master roster

David Maynard wrote:
~misfit~ wrote:
David Maynard wrote:

I take the direct approach and simply break the appropriate pin off
the
chip carrier but you're screwed if you go cross-eyed and break off
the wrong one, although P2B in here seems to have mastered some
form of voo-doo that enabled him to actually solder pins back on
one of his. Seems unlikely
to me that mere mortals should count on replicating that feat,
though.


Hey! That was me who managed to solder a pin back on a CPU (Celly
600). Unless P2B has done it as well. :-)
--
~misfit~

My apologies for inadvertently omitting you from the voo-doo master
roster

:-)

Cheers, it's just that I'm bloody pleased I managed it. I'd never heard of
it being done before but figured I had nothing to lose. T'wasn't easy but
it's stood up to being taken out of the socket and put into a different one
ok.
--
~misfit~

David Maynard wrote:
noob wrote:
What do you recommend for isolating the pin? Any ideas?

Well, there are a number of 'ideas' that have been promulgated. One
is to 'paint' it with either nail polish or super glue.

Cyanoacrylat is a moderately bad conductor, so superglue is probably not a
good idea. The gases it releases while drying might also damage the chip.

For anyone reading this who doubts that plastics can conduct electricity,
note that the 2000 Nobel Price in chemistry was awarded to the chemists who
could first explain this phenomenon.

Nail polish is probably better, but I would check it out first.

The one I
remember seeing was to use the bottom end of a ball-point pen, the
guts removed, fill it
with the 'paint' and then slide it over the pin so it gets coated.

Few inks are good electrical insulators -- many of them contain large
amounts of conducting metals and metal oxides.

Regards,
--
*Art

the problem with altering the voltage using the wire trick is that the
'good' voltages (those between 1,65V - 1,85V) aren't available. often the
overclock will work without a vcore hike; and if it won't, then i always
found it was safe to go up to 2,05V without serious issues. having said
that, experience showed that anything above 1,95V got you into the realm of
diminishing returns: a higher voltage produces a clearer signal, but also
causes more heat - which in turn muddies up the signal again etc.


"noob" wrote in message
news:ZZnMb.61161$PK3.21579@okepread01...
"David Maynard" wrote in message
...
Got it, thanks!! I found a web reference here
http://www.dualcpu.jp/fcpga-pin.html
There doesn't seem to be much material on the web about this.....

It's on old 'trick' that originated with the slot-1 celerons, except it
was
easier to tape the connector pin than it is to isolate the same thing on
the PPGA and FC-PGA chip carrier. It's not 'common' now because most
third
party 'modern' motherboards provide some way to do it, with either the
BIOS
or jumpers, and I'm a bit surprised that the MSI doesn't.


I wonder if I should go ahead and raise the vCore while I'm at it?

It's likely you'll need a bit of a Vcore boost.

Odds are it'll run 112 MHz FSB too, assuming you have memory that'll do
it.
Mine made it to 120Mhz FSB for 1020.

What voltage did you run yours at?


I'm running a spare Vantec Aeroflow on it, so it should have plenty of
cooling.

What do you recommend for isolating the pin? Any ideas?

Well, there are a number of 'ideas' that have been promulgated. One is
to
'paint' it with either nail polish or super glue. The one I remember
seeing
was to use the bottom end of a ball-point pen, the guts removed, fill it
with the 'paint' and then slide it over the pin so it gets coated. I've
never tried that so lord knows if it works, or simply slops goop all
over
everything. Another is to wallow out the CPU socket, so there's more
room,
and then slide a real thin insulation, like stripped from a wire wrap
wire,
over the processor pin before inserting it into the socket. Another I
saw
was to remove the socket top, cut a small piece of tape and insert it in
the contact in the socket, and then put the top back on. The idea with
these is to leave the processor 'like new' to save the warrantee and
avoid
a 'boo-boo' destroying it (although if one 'boo-boos' on wallowing out
the
wrong socket pin you've ruined the motherboard so it's not 'risk free'
either).

I take the direct approach and simply break the appropriate pin off the
chip carrier but you're screwed if you go cross-eyed and break off the
wrong one, although P2B in here seems to have mastered some form of
voo-doo
that enabled him to actually solder pins back on one of his. Seems
unlikely
to me that mere mortals should count on replicating that feat, though.

A variation on that would be to buy an adapter socket and break the pin
off
of it rather than the processor but with FC-PGA celerons, up to 700Mhz,
going for under 25 bucks it's not much cheaper to replace in the event
of
a
'boo-boo' and it adds the up front cost of the socket even without a
'boo-boo'. On the other hand, maybe one of the adapters has a jumper
inside
it to isolate the pin. I haven't checked for that since, as I mentioned,
I
do the direct pin break.

Thanks! I think I'll go for the pin break. I don't think I'd go wrong
with it, and if I did, I'd prolly save a lot of stress & mess w/ mods by
just getting another Celly for under $25, like you say.

"atwifa" wrote in message
...
the problem with altering the voltage using the wire trick is that the
'good' voltages (those between 1,65V - 1,85V) aren't available. often the

I'd think they were available, according to this link:
http://www.hyperformance-pc.com/intel_vid.htm

overclock will work without a vcore hike; and if it won't, then i always
found it was safe to go up to 2,05V without serious issues. having said
that, experience showed that anything above 1,95V got you into the realm
of
diminishing returns: a higher voltage produces a clearer signal, but also
causes more heat - which in turn muddies up the signal again etc.

noob wrote:

"atwifa" wrote in message
...

the problem with altering the voltage using the wire trick is that the
'good' voltages (those between 1,65V - 1,85V) aren't available. often the


I'd think they were available, according to this link:
http://www.hyperformance-pc.com/intel_vid.htm

They are, just not by using wire tricks only.

To change a 1 to a 0, connect the pin to Vss as noted on that page.
To change a 0 to a 1, insulate the pin from the socket - or break it off
if you're *certain* you want it changed to a 1 permanently.

overclock will work without a vcore hike; and if it won't, then i always
found it was safe to go up to 2,05V without serious issues. having said
that, experience showed that anything above 1,95V got you into the realm

of

diminishing returns: a higher voltage produces a clearer signal, but also
causes more heat - which in turn muddies up the signal again etc.

Arthur Hagen wrote:
David Maynard wrote:

noob wrote:

What do you recommend for isolating the pin? Any ideas?

Well, there are a number of 'ideas' that have been promulgated. One
is to 'paint' it with either nail polish or super glue.


Cyanoacrylat is a moderately bad conductor, so superglue is probably not a
good idea. The gases it releases while drying might also damage the chip.

For anyone reading this who doubts that plastics can conduct electricity,
note that the 2000 Nobel Price in chemistry was awarded to the chemists who
could first explain this phenomenon.

Nail polish is probably better, but I would check it out first.

Yes, well, we're not talking about insulating the thing from a lightning
strike, just to keep it from pulling the BSel0 line low.


The one I
remember seeing was to use the bottom end of a ball-point pen, the
guts removed, fill it
with the 'paint' and then slide it over the pin so it gets coated.


Few inks are good electrical insulators -- many of them contain large
amounts of conducting metals and metal oxides.

The ball point pen idea had nothing to do with using it's ink. It was just
to use the body of it as a paint holder to insert the processor pin into.
The theory, of course, being that surface tension keeps whatever coating
material one is using from 'leaking' out the small hole on the writing end.

As I said, I have no idea if that works as I've never tried it.

Regards,

atwifa wrote:
the problem with altering the voltage using the wire trick is that the
'good' voltages (those between 1,65V - 1,85V) aren't available.

It depends on which celeron. If it's a 1.7 volt core then you're in hog
heaven as you can select anything from there on up. If it's a 1.65 volt
celeron then you're screwed: the only choice by simply pulling a VID pin
low (VID3 being the only one high) is 2.05. If it's a 1.5 volt core then
pulling VID3 low gives 1.9. But then you can also get 1.65 by pulling VID 0
and Vid 1 low (leaving VID3 high).

often the
overclock will work without a vcore hike; and if it won't, then i always
found it was safe to go up to 2,05V without serious issues. having said
that, experience showed that anything above 1,95V got you into the realm of
diminishing returns: a higher voltage produces a clearer signal, but also
causes more heat - which in turn muddies up the signal again etc.



"noob" wrote in message
news:ZZnMb.61161$PK3.21579@okepread01...

"David Maynard" wrote in message
...

Got it, thanks!! I found a web reference here
http://www.dualcpu.jp/fcpga-pin.html
There doesn't seem to be much material on the web about this.....

It's on old 'trick' that originated with the slot-1 celerons, except it

was

easier to tape the connector pin than it is to isolate the same thing on
the PPGA and FC-PGA chip carrier. It's not 'common' now because most

third

party 'modern' motherboards provide some way to do it, with either the

BIOS

or jumpers, and I'm a bit surprised that the MSI doesn't.



I wonder if I should go ahead and raise the vCore while I'm at it?

It's likely you'll need a bit of a Vcore boost.

Odds are it'll run 112 MHz FSB too, assuming you have memory that'll do

it.

Mine made it to 120Mhz FSB for 1020.

What voltage did you run yours at?



I'm running a spare Vantec Aeroflow on it, so it should have plenty of
cooling.

What do you recommend for isolating the pin? Any ideas?

Well, there are a number of 'ideas' that have been promulgated. One is

to

'paint' it with either nail polish or super glue. The one I remember

seeing

was to use the bottom end of a ball-point pen, the guts removed, fill it
with the 'paint' and then slide it over the pin so it gets coated. I've
never tried that so lord knows if it works, or simply slops goop all

over

everything. Another is to wallow out the CPU socket, so there's more

room,

and then slide a real thin insulation, like stripped from a wire wrap

wire,

over the processor pin before inserting it into the socket. Another I

saw

was to remove the socket top, cut a small piece of tape and insert it in
the contact in the socket, and then put the top back on. The idea with
these is to leave the processor 'like new' to save the warrantee and

avoid

a 'boo-boo' destroying it (although if one 'boo-boos' on wallowing out

the

wrong socket pin you've ruined the motherboard so it's not 'risk free'

either).

I take the direct approach and simply break the appropriate pin off the
chip carrier but you're screwed if you go cross-eyed and break off the
wrong one, although P2B in here seems to have mastered some form of

voo-doo

that enabled him to actually solder pins back on one of his. Seems

unlikely

to me that mere mortals should count on replicating that feat, though.

A variation on that would be to buy an adapter socket and break the pin

off

of it rather than the processor but with FC-PGA celerons, up to 700Mhz,
going for under 25 bucks it's not much cheaper to replace in the event

of

a

'boo-boo' and it adds the up front cost of the socket even without a
'boo-boo'. On the other hand, maybe one of the adapters has a jumper

inside

it to isolate the pin. I haven't checked for that since, as I mentioned,

I

do the direct pin break.

Thanks! I think I'll go for the pin break. I don't think I'd go wrong
with it, and if I did, I'd prolly save a lot of stress & mess w/ mods by
just getting another Celly for under $25, like you say.

P2B wrote:


noob wrote:

"atwifa" wrote in message
...

the problem with altering the voltage using the wire trick is that the
'good' voltages (those between 1,65V - 1,85V) aren't available.
often the



I'd think they were available, according to this link:
http://www.hyperformance-pc.com/intel_vid.htm


They are, just not by using wire tricks only.

To change a 1 to a 0, connect the pin to Vss as noted on that page.
To change a 0 to a 1, insulate the pin from the socket - or break it off
if you're *certain* you want it changed to a 1 permanently.

Or, if you're really ambitious you could break them all off and wire the
motherboard socket pins to a dip switch and set it to anything you like.

Which is also a way to 'get it back' if you wanted to after breaking one
off. Either the switch or a simple wire jumper on the motherboard.


overclock will work without a vcore hike; and if it won't, then i always
found it was safe to go up to 2,05V without serious issues. having said
that, experience showed that anything above 1,95V got you into the realm


of

diminishing returns: a higher voltage produces a clearer signal, but
also
causes more heat - which in turn muddies up the signal again etc.

~misfit~ wrote:
David Maynard wrote:

~misfit~ wrote:

David Maynard wrote:


I take the direct approach and simply break the appropriate pin off
the
chip carrier but you're screwed if you go cross-eyed and break off
the wrong one, although P2B in here seems to have mastered some
form of voo-doo that enabled him to actually solder pins back on
one of his. Seems unlikely
to me that mere mortals should count on replicating that feat,
though.


Hey! That was me who managed to solder a pin back on a CPU (Celly
600). Unless P2B has done it as well. :-)
--
~misfit~

My apologies for inadvertently omitting you from the voo-doo master
roster


:-)

Cheers, it's just that I'm bloody pleased I managed it. I'd never heard of
it being done before but figured I had nothing to lose. T'wasn't easy but
it's stood up to being taken out of the socket and put into a different one
ok.

I understand. That's why I humorously referred to it as voo-doo

--
~misfit~