Socket 370 in Slot 1

  • Thread starter Thread starter Gregory L. Hansen
  • Start date Start date
G

Gregory L. Hansen

Here's a more ambitious upgrade idea for my Kayak XAS -- a socket 370 to
slot 1 converter to put something like a PIII 850 MHz in it, which still
has a 100 MHz bus. The reason I always read that Coppermine processors
can't be used is because they use a different voltage, but if that's the
only reason, a different VRM can be plugged in and the problem will be
solved! I'm trying to find pinouts for HP and others' VRMs to form an
opinion on compatibility.


But are there incompatibilities besides the VRM? A limitation of the
chipset or of the motherboard? If the BIOS is updated for a PIII 600MHz,
will it automatically run the PIII 850MHz, or will it need another update
that can only come from HP, and which HP might not have?
 
On Mon, 13 Oct 2003 16:41:38 +0000 (UTC),
Here's a more ambitious upgrade idea for my Kayak XAS -- a socket 370 to
slot 1 converter to put something like a PIII 850 MHz in it, which still
has a 100 MHz bus.

Yes, that's another alternative, but you might as well shoot for
something higher than an 850, either a 1GHz, 950, Celeron 1.1GHz, or
an Upgradeware type adapter to run a Tualatin Celeron or PIII.

The reason I always read that Coppermine processors
can't be used is because they use a different voltage, but if that's the
only reason, a different VRM can be plugged in and the problem will be
solved!

Just about any ADJUSTABLE voltage regulator (need not be
user-adjustable, only sensing the CPU voltage "automatically") will
support down to 1.8V. This is true of motherboards in general, but
I'm not as confident it applies to VRMs. Even on VRMs that don't
officially support a Coppermine's 1.5V-1.75V Vcore, there might be the
option to use 1.8V... that's low enough to be safe for a Coppermine
CPU, but requires a sloket adapter with a voltage jumper to set 1.8V
(or some CPU pin wire-mods or slotket modifications). This part of it
is certainly doable given the time to work out any issues if/as they
arise.
I'm trying to find pinouts for HP and others' VRMs to form an
opinion on compatibility.

It's quite likely you won't find these pinouts. Once you have a VRM,
if you can provide very detailed pictures of it or the markings on
some of the chips I "might" be able to tell what voltage range it
supports, but of course there's only so much confidence I can have
over usenet, not having the part(s) in front of me. At least by
looking at the controller used on the VRM, the maximum voltage range
possible can be determined.

But are there incompatibilities besides the VRM? A limitation of the
chipset or of the motherboard? If the BIOS is updated for a PIII 600MHz,
will it automatically run the PIII 850MHz, or will it need another update
that can only come from HP, and which HP might not have?

Yes, there are other potentials for incompatibilties. The chipset,
being a BX, isn't one of them, it's known and proven to work with
Coppermine P3/Celerons.

The BIOS is the major issue, above all else. Without a supportive
BIOS the system may not even POST with a Coppermine P3 in it.

It is possible HP added Coppermine support to the BIOS, possible that
the BIOS would work without this specific support, but on the other
hand, it's also possible the BIOS would halt when it couldn't ID the
CPU, would've worked except that it "unnecessarily" halted itself.
There might be a BIOS setting to adjust this halt behavoir, but I
wouldn't count on it, generally such systems have lesser flexibility
when it comes to CPU upgrades.

The last part of what you wrote pretty well sums up the situation,
that it might work with the last BIOS, or might not, and if not, you'd
need a BIOS from HP that did add (allow) Coppermine usage, which may
not exist. If you do try it you should update the BIOS before
swapping the CPU, or just swap the old one back in then update it.


Dave
 
kony said:
On Mon, 13 Oct 2003 16:41:38 +0000 (UTC),


Yes, that's another alternative, but you might as well shoot for
something higher than an 850, either a 1GHz, 950, Celeron 1.1GHz, or
an Upgradeware type adapter to run a Tualatin Celeron or PIII.

I'd just like to point out that there was a thread in a.c.h.o a few months
ago now, maybe last year, where a guy was updating a slot 1 dual-CPU
motherboard to run Tualatins. He found that, using the Upgradeware adapters,
he couldn't even get it to boot. He was using an Asus P2B-DS board. He ended
up modifying slockets himself and ended up running dual PIIIs at 1574MHz
stabily. I guess he overclocked a little too. Strangely, with one modified
slocket and one upgradeware adapter it would boot but wasn't stable so he
modified two slockets.

I have a copy of the webpage he put up to show us how he modified the
slockets but unfortunately not the URL I saved it from. If anyone is
sufficiently interested I can email him for the URL (his email addy is on
the webpage) or email a copy of the page to whoever wants it.

(I no longer use a valid return email in NGs, even munged, as, in one group
I sometimes used to frequent there was a troll who was put in his place. He
took offence and un-munged several peoples addys and signed them up for all
sorts of lists. Of course once this happened the addy got onto spammers
lists and I had to dump it.) If you want a copy of the page let me know
here. Or google a.c.h.o. for the thread. P2B-DS over the last year in
a.c.h.o should find it for you. It was extensive.
 
Gregory L. Hansen said:
Here's a more ambitious upgrade idea for my Kayak XAS -- a socket 370 to
slot 1 converter to put something like a PIII 850 MHz in it, which still
has a 100 MHz bus. The reason I always read that Coppermine processors
can't be used is because they use a different voltage, but if that's the
only reason, a different VRM can be plugged in and the problem will be
solved! I'm trying to find pinouts for HP and others' VRMs to form an
opinion on compatibility.


But are there incompatibilities besides the VRM? A limitation of the
chipset or of the motherboard? If the BIOS is updated for a PIII 600MHz,
will it automatically run the PIII 850MHz, or will it need another update
that can only come from HP, and which HP might not have?

I can only speak from experience Gregory, and I have no experience with a
Kayak.

I'm running a few BX motherboards with coppermines. MS-6163 Pros. They were
running pre-coppermine CPUs when I bought them. I did some research, went to
MSI's site and it was stated that if the board had a blue sticker on the I/O
chip and was past a certain revision then, with a BIOS update, it was
coppermine-compatible. Luckilly mine have the sticker and are sufficiently
late revisions.

I also had an ECS BX mobo that I updated to run a coppermine. With both the
MSI and ECS mobos I read through the changes in each BIOS revision and, at a
certain revision they stated "Added microcode for coppermine CPUs".

These two things suggest to me that you need both a board that supports
coppermines and a BIOS revision that does.
 
Yes, that's another alternative, but you might as well shoot for
something higher than an 850, either a 1GHz, 950, Celeron 1.1GHz, or
an Upgradeware type adapter to run a Tualatin Celeron or PIII.

Yeah, but the 850 was the first one I saw without doing too much work that
had a 100 MHz bus. It's a "for instance".
Just about any ADJUSTABLE voltage regulator (need not be
user-adjustable, only sensing the CPU voltage "automatically") will
support down to 1.8V. This is true of motherboards in general, but
I'm not as confident it applies to VRMs. Even on VRMs that don't
officially support a Coppermine's 1.5V-1.75V Vcore, there might be the
option to use 1.8V... that's low enough to be safe for a Coppermine
CPU, but requires a sloket adapter with a voltage jumper to set 1.8V
(or some CPU pin wire-mods or slotket modifications). This part of it
is certainly doable given the time to work out any issues if/as they
arise.


It's quite likely you won't find these pinouts. Once you have a VRM,
if you can provide very detailed pictures of it or the markings on
some of the chips I "might" be able to tell what voltage range it
supports, but of course there's only so much confidence I can have
over usenet, not having the part(s) in front of me. At least by
looking at the controller used on the VRM, the maximum voltage range
possible can be determined.

I have a 0950-2837 in the box now, but I can't easily get a picture to
you. But it's a two row, forty pin connector, so that's sort of
promising. There's a power transistor package, the kind with three leads
and a heat sink tab sticking out, labeled 9812/FDP7030L. I don't
recognize that number. There's a 20 pin mini DIP labeled
HIP5004ACB/H9810. A sticker that says 5V IN. Ten electrolytic
capacitors, two inductors, and a whole bunch of little tiny diodes,
capacitors, and resistors. And a second power transistor package, I
almost missed it because it's hidden behind capacitors, and I can't read
the label.

For whatever that's worth. Probably not much.
Yes, there are other potentials for incompatibilties. The chipset,
being a BX, isn't one of them, it's known and proven to work with
Coppermine P3/Celerons.

The BIOS is the major issue, above all else. Without a supportive
BIOS the system may not even POST with a Coppermine P3 in it.

It is possible HP added Coppermine support to the BIOS, possible that
the BIOS would work without this specific support, but on the other
hand, it's also possible the BIOS would halt when it couldn't ID the
CPU, would've worked except that it "unnecessarily" halted itself.
There might be a BIOS setting to adjust this halt behavoir, but I
wouldn't count on it, generally such systems have lesser flexibility
when it comes to CPU upgrades.

I thought it would be something like that. Maybe HP supports it, maybe
they don't, maybe the tech support person knows whether they support it,
maybe he doesn't (they haven't gotten back to me on that yet). And since
it's not a standard motherboard, it's less likely that a third party
wrote a BIOS to support it. But maybe they did. PowerLeap's web site
has some links to third-party BIOSes, maybe I can get something from a
non-HP source.
 
On Tue, 14 Oct 2003 01:47:13 +0000 (UTC),
Yeah, but the 850 was the first one I saw without doing too much work that
had a 100 MHz bus. It's a "for instance".

I suppose the primary question I don't remember seeing asked, is what
the purpose, role of this system will be. If it's only going to be a
filesever for example, it matters very little what CPU is in it,
wouldn't have better performance with any CPU upgrade, not realized
across a LAN.
I have a 0950-2837 in the box now, but I can't easily get a picture to
you. But it's a two row, forty pin connector, so that's sort of
promising. There's a power transistor package, the kind with three leads
and a heat sink tab sticking out, labeled 9812/FDP7030L. I don't
recognize that number. There's a 20 pin mini DIP labeled
HIP5004ACB/H9810. A sticker that says 5V IN. Ten electrolytic
capacitors, two inductors, and a whole bunch of little tiny diodes,
capacitors, and resistors. And a second power transistor package, I
almost missed it because it's hidden behind capacitors, and I can't read
the label.

For whatever that's worth. Probably not much.

Are you certain the "HIP5004ACB" above isn't "HIP6004"?
I see very little mention of the 5004 on the 'net, nothing on the
manufacturer's website. On the other hand, the 6004 has some data
available, like a VRM datasheet (w/schematics), briefs, datasheets:
http://www.intersil.com/design/parametric/deviceinfo.asp?pn=HIP6004


The HIP6004, ACB version seems to only support down to 1.8V, while the
BCB supports down to 1.3V, not that this is reliable extended to a
5004, but suggests it might be less likely to support under 1.8V.

Anyway, it's likely the 5004 would support at least down to 1.8V, so
in that case you'd need a slotket with voltage adjustment jumpers on
it, or one of a few other possible mods to the slotket or CPU to get
the correct VID signal sent to the VRM (or I suppose you could modify
the VRM instead, but not knowing how crowded it is with
parts/traces...

I thought it would be something like that. Maybe HP supports it, maybe
they don't, maybe the tech support person knows whether they support it,
maybe he doesn't (they haven't gotten back to me on that yet). And since
it's not a standard motherboard, it's less likely that a third party
wrote a BIOS to support it. But maybe they did. PowerLeap's web site
has some links to third-party BIOSes, maybe I can get something from a
non-HP source.

At this point I wouldn't expect much from HP, likely even the
engineers who worked on it don't remember all the details, and you
won't likely be in contact with them. Any other support person(s)
response is most likely to be that it "Isn't supported", which of
course could mean almost anything.

The odds that a 3rd party made a bios for it are incredibly low, not
even worth looking for one, IMHO.

I get the feeling that the only way you're going to find out if a
Coppermine will work, is to actually try one, with the voltage set to
1.8V. If that works, then you know the board will run a Coppermine
and can try lower voltages. It might still be a good idea to update
the bios to the last released version if it hasn't been, regardless of
whether there is mention of Coppermine support.


Dave
 
On Mon, 13 Oct 2003 16:41:38 +0000 (UTC),
Here's a more ambitious upgrade idea for my Kayak XAS -- a socket 370 to
slot 1 converter to put something like a PIII 850 MHz in it, which still
has a 100 MHz bus. The reason I always read that Coppermine processors
can't be used is because they use a different voltage, but if that's the
only reason, a different VRM can be plugged in and the problem will be
solved! I'm trying to find pinouts for HP and others' VRMs to form an
opinion on compatibility.
But are there incompatibilities besides the VRM? A limitation of the
chipset or of the motherboard? If the BIOS is updated for a PIII 600MHz,
will it automatically run the PIII 850MHz, or will it need another update
that can only come from HP, and which HP might not have?

should run it after replacing the VRM (probably you will get a garbage
at post on your screen, but it should work if the clock generator can
handle 8,5 multiplier (very probably yes even if there is not
mentioned in the manual) But maybe there would even not be a necesity
to replace a VRM if it can go lower than 1,8V (check its markings &
find a datasheet on the net)

if it goes to 1,8Vcore low in bios or jumpers (on the slotket also if
used), maybe the chip Would not need replacement, but a bit better
cooling & than for god sake, OC it than even further :-) !

see what I run! the MoBo was also not ment to run this Cpu .. :-)





-- Regards, SPAJKY
& visit site - http://www.spajky.vze.com
Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
E-mail AntiSpam: remove ##
 
Spajky said:
On Mon, 13 Oct 2003 16:41:38 +0000 (UTC),



should run it after replacing the VRM (probably you will get a garbage
at post on your screen, but it should work if the clock generator can
handle 8,5 multiplier (very probably yes even if there is not
mentioned in the manual) But maybe there would even not be a necesity
to replace a VRM if it can go lower than 1,8V (check its markings &
find a datasheet on the net)

if it goes to 1,8Vcore low in bios or jumpers (on the slotket also if
used), maybe the chip Would not need replacement, but a bit better
cooling & than for god sake, OC it than even further :-) !

see what I run! the MoBo was also not ment to run this Cpu .. :-)





-- Regards, SPAJKY
& visit site - http://www.spajky.vze.com
Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"

I like the changes to your hame page Spajky. You've done a good job, both
with the site and with the computer. I'm amazed you ever got that Mendicino
to go that fast. I had five 400Mhz CPUs to play with a while ago and the
fastest of them will only do 545MHz.

I have a question about the MS6163-Pro. I'm running three of them. One of
them counts the RAM jerkily at start-up, hard to explain but it sort of
counts a block, pauses for a quarter-second or so, then counts some more. It
pauses about three or four times. It does this regardless of the RAM
modules, I've swapped them for others and it's still the same. It's running
a Coppermine Celly 900, not overclocked as it won't even POST overclocked.
(Standard 100Mhz FSB)

The other two boards are running Celly Coppermine 600s. One cB0 stepping
running at 855Mhz and one cC0 stepping running at 900MHz. They both count
the RAM smoothly. The other board used to as well, it just started going
jerky after I tried 110MHz FSB (It's PC133 RAM) and it wouldn't POST, then
stayed that way when I dropped it back to 100.

Any ideas? It runs fine otherwise, I just wondered about it.

Thanks.
 
I like the changes to your hame page Spajky. You've done a good job, both
with the site and with the computer. I'm amazed you ever got that Mendicino
to go that fast. I had five 400Mhz CPUs to play with a while ago and the
fastest of them will only do 545MHz.
It was from Malaysia B plant boxed one ´99 43 week; I could get from
the same box opened few of them (all did more than 600 default Vcore,
tried) :-) , looks like was a good wafer !
I have a question about the MS6163-Pro. I'm running three of them. One of
them counts the RAM jerkily at start-up, hard to explain but it sort of
counts a block, pauses for a quarter-second or so, then counts some more. It
pauses about three or four times. It does this regardless of the RAM
modules, I've swapped them for others and it's still the same. It's running
a Coppermine Celly 900, not overclocked as it won't even POST overclocked.
(Standard 100Mhz FSB)

The other two boards are running Celly Coppermine 600s. One cB0 stepping
running at 855Mhz and one cC0 stepping running at 900MHz. They both count
the RAM smoothly. The other board used to as well, it just started going
jerky after I tried 110MHz FSB (It's PC133 RAM) and it wouldn't POST, then
stayed that way when I dropped it back to 100.

Any ideas? It runs fine otherwise, I just wondered about it.

it looks like chipset gone partially south! (or some ram slot having
problems affecting all others too!
My experience, avoid using the middle ram slot...
or maybe some "cold" soldering somewhere, who knows; happens ...
maybe some cap is bad also or something there around ...

Maybe was hit by a electrical surge & by time started having problems
(recently had with some mobo such a case here...)

have you tried changing slotkets? could be problem also here ...

-- Regards, SPAJKY
& visit site - http://www.spajky.vze.com
Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
E-mail AntiSpam: remove ##
 
I suppose the primary question I don't remember seeing asked, is what
the purpose, role of this system will be. If it's only going to be a
filesever for example, it matters very little what CPU is in it,
wouldn't have better performance with any CPU upgrade, not realized
across a LAN.

The purpose of the system will be hobby. Maybe some GIMP, maybe some
number crunching, some dicking around. A processor upgrade probably
isn't even really necessary for what I'll be doing, but hardware is also
part of the hobby.

At any rate, I think I've decided that if I'm going to stick with HP
approved CPUs I'll just buy a 0950-2837 from them for $45 including
shipping. And if I decide to go beyond that I'll go to 1 GHz with upgrade
cards from PowerLeap which include an on-board voltage regulator.
PowerLeap is actually looking pretty attractive considering 550 and 600MHz
Katmais are still going on eBay for more than $50 each and I'll save the
cost of a VRM, so it won't be that much more money spent. So the VRM
issue is essentially solved, one way or the other.
At this point I wouldn't expect much from HP, likely even the
engineers who worked on it don't remember all the details, and you
won't likely be in contact with them. Any other support person(s)
response is most likely to be that it "Isn't supported", which of
course could mean almost anything.

The odds that a 3rd party made a bios for it are incredibly low, not
even worth looking for one, IMHO.

I get the feeling that the only way you're going to find out if a
Coppermine will work, is to actually try one, with the voltage set to
1.8V. If that works, then you know the board will run a Coppermine
and can try lower voltages. It might still be a good idea to update
the bios to the last released version if it hasn't been, regardless of
whether there is mention of Coppermine support.

Yep. The VRM issue is settled, that leaves the BIOS. PowerLeap has a 30
day return policy, so some day I'm just going to have to try it and see.

Thanks for the help.
 
Spajky said:
It was from Malaysia B plant boxed one ´99 43 week; I could get from
the same box opened few of them (all did more than 600 default Vcore,
tried) :-) , looks like was a good wafer !


it looks like chipset gone partially south! (or some ram slot having
problems affecting all others too!
My experience, avoid using the middle ram slot...

OK, I'll try that. At the moment it's got 320MB, Two 128s and a 64. I'll
take the 64 out of the third slot and put the 128 from slot two in there.
or maybe some "cold" soldering somewhere, who knows; happens ...
maybe some cap is bad also or something there around ...

Maybe was hit by a electrical surge & by time started having problems
(recently had with some mobo such a case here...)

have you tried changing slotkets? could be problem also here ...

Yep, tried that but no luck. It's currently running the 'best' slocket I
have, an MSI Master.

Thanks for the help.
 
CPU upgrade kits
www.newwaveupgrades.com

~misfit~ said:
OK, I'll try that. At the moment it's got 320MB, Two 128s and a 64. I'll
take the 64 out of the third slot and put the 128 from slot two in there.


Yep, tried that but no luck. It's currently running the 'best' slocket I
have, an MSI Master.

Thanks for the help.
 
Back
Top