Asus P2b + Tualatin Celeron CPU

[Sune H.]

Hello.

Is it posible to get a Tualatin Celeron CPU to work in my old Asus P2b rev.
1.02 ?

I already have a slotkey konverter with volt adjustment and uses the latest
bios (1014.beta2 or something like that). Have been running a Celeron
600@900 so far, but that CPU has to go for an other system and now I need a
new for this one.

Thanks in advance for any answer.


Regards.
Sune
Denmark

 

[[BnH]]

Last time I muck around with Tualatin technology .. you can't
as Celeron 600 @ 900 is 1.5v ? and Tualatin core uses 1.3v ?
can your Slotket go that low?

=bob=

 

[P2B]

Hello.

Is it posible to get a Tualatin Celeron CPU to work in my old Asus P2b rev.
1.02 ?

I already have a slotkey konverter with volt adjustment and uses the latest
bios (1014.beta2 or something like that). Have been running a Celeron
600@900 so far, but that CPU has to go for an other system and now I need a
new for this one.

Thanks in advance for any answer.


Regards.
Sune
Denmark

Yes, but you will not be able to use the slot adapter you already have
as it was more than likely designed for Coppermine processors and there
are several pinout changes on the Tualatin processors. Coppermine slot
adapters are incompatible with Tualatin processors, although they can
usually be modified to work.

The main issue is the voltage regulator on your version 1.02 mainboard -
the minimum CPU voltage it can supply is 1.8v, whereas Tualatin
processors require 1.5v. The voltage jumpers on the adapter allow you to
override the voltage requested from the mainboard by the CPU, but the
voltage regulator will shut down if the voltage request is less than
1.8v, and a Tualatin processor is unlikely to last very long if supplied
with 1.8v.

Therefore you need a Tualatin compatible slot adapter and a 1.5v CPU
voltage supply. There are two possible approaches:

1. Change the voltage regulator chip on the motherboard to one that can
supply 1.5v, and use a Slot-T adapter from Upgradeware (about $20). This
solution requires surface-mount soldering skills, and a replacement
voltage regulator chip - which in practice must be removed from another
mainboard because new chips are very difficult to obtain, especially in
quantity 1. I can supply a used one if you decide to go this route.

2. Purchase a Tualatin slot adapter with a built-in CPU voltage
regulator. AFAIK the Powerleap PL-iP3/T is the only suitable product
available. This solution is plug and play, but more expensive -
currently $120 (including 1.4Ghz Celeron) from Powerleap, possibly
cheaper on eBay.

HTH

P2B

 

[Paul]

Hello.

Is it posible to get a Tualatin Celeron CPU to work in my old Asus P2b rev.
1.02 ?

I already have a slotkey konverter with volt adjustment and uses the latest
bios (1014.beta2 or something like that). Have been running a Celeron
600@900 so far, but that CPU has to go for an other system and now I need a
new for this one.

Thanks in advance for any answer.


Regards.
Sune
Denmark

According to Roland's web page, the cutoff for Tualatin compatible voltage
regulators is P2B rev 1.12 -

http://homepage.hispeed.ch/rscheidegger/p2b_procupgrade_faq.html

This means the owner of a P2B rev 1.12 can use an Upgradeware Slot-T plus
a Tualatin. The processor will run at 1.5 volts because the voltage
regulator can go that low. Finding a Slot-T at around $25 or so gives
the basis for a cheap upgrade.

Since your P2B is less than rev 1.12, you need a Powerleap to run a
Tualatin. Your voltage regulator only goes down to 1.8V (never less
than that), so that is too high a voltage for a Tualatin which has
1.75V absolute max and 1.5V normal operating voltage.

http://plpadmin.tempdomainname.com/PLiP3T.html

There is a iP3-T with a Tualatin 1400/100/256 processor for $119.
If your motherboard supports setting the FSB to 133MHz (where PCI
stays at 33MHz, but AGP becomes 89MHz instead of 66MHz), then you
could take a 1000/100 or 1100/100 processor and overclock it to
get performance closer to the Server version of the Tualatin.

I changed the voltage regulator on my board, which enabled me to use
a Slot-T plus Tualatin. Considering the incidentals purchased to do
the surgery, a Powerleap would have been cheaper.

If you can find a Coppermine Celeron FCPGA, you can
use a slocket with VID jumpers, even with your motherboard, by setting
the VID jumpers to 1.8 volts. I tried this originally with a 1.75V nominal
Coppermine on a modified Startech slocket (added VID jumpers) and it
worked fine. You might actually have more luck finding a 1GHz Pentium
cartridge than a Coppermine FCPGA, in which case there would still be
a requirement to hack a 1.8V VID setting for your voltage regulator
( assuming the cartridge needs less than 1.8V, which you can check
by entering the SSPEC number at processorfinder.intel.com ).

To see one unique way of setting the VID on a slocket without VID
jumpers, see this page, where a U shaped wire is used as a shunt.
(I soldered mine). The example shown here is a one wire mod to run
a 1.7volt processor at 1.8volts. Other processor voltages are a
little more complicated, because more than one VID signal needs
to be modified (may require cutting a track on the cartridge or
fitting wires and switches on the motherboard) -

http://www.tipperlinne.com/p2b-ds.htm

The latest batch of Tualatin processors have a CPUID of 06B4. The
microcode to support this processor isn't in the latest beta BIOS,
so you can add support with low risk, by using CTMC and instructions
found here:

http://www.tipperlinne.com/bios6b4.htm

The main page on the site is here:

http://www.tipperlinne.com/p2bmod.htm

In terms of processors, here are a selected few entries:

SSPEC Core/Bus/Cache CPUID Pkg Vcore Pwr Typ Tmax Geom

Pentium Coppermine with 100 or 133MHz bus (cartridge needs VID
setting hack depending on regulator) -
SL4BR 1000/100/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both
SL4BS 1000/133/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both

Coppermine in FCPGA (Coppermine slocket with VID set to 1.8V)
SL5QV 1000/100/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u OEM
SL5QW 1100/100/256 068A FC-PGA 1.75V 33W cD0 77C 0.18u OEM
SL5DV 1000/133/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u Boxed

Tualatin in FCPGA2 (Slot-T or Powerleap, depending on regulator)
SL6JQ 1000A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JR 1100A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JS 1200/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JT 1300/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JV 1400/100/256 06B4 FC-PGA2 1.5V 33.2W tB1 67C 0.13u Boxed

Tualatin server version (expensive) - same as Tualatin above
[ Benchmarks don't seem to get enough benefit from the cache... ]
SL6BW 1.13-S/133/512 06B4 FC-PGA2 1.45V 28.7W tB1 69C 0.13u OEM
SL6BX 1.26-S/133/512 06B4 FC-PGA2 1.45V 30.4W tB1 69C 0.13u Both
SL6BY 1.40-S/133/512 06B4 FC-PGA2 1.45V 32.2W tB1 69C 0.13u OEM

There are more types than that (some with 128K cache), but that
is all I can figure out from my list of processors. The Tualatins
can still be bought at retail, but have gone out of production.
The other processors will be Ebay material.

Have fun,
Paul

 

[P2B]

Hello.

Is it posible to get a Tualatin Celeron CPU to work in my old Asus P2b rev.
1.02 ?

I already have a slotkey konverter with volt adjustment and uses the latest
bios (1014.beta2 or something like that). Have been running a Celeron
600@900 so far, but that CPU has to go for an other system and now I need a
new for this one.

Thanks in advance for any answer.


Regards.
Sune
Denmark

According to Roland's web page, the cutoff for Tualatin compatible voltage
regulators is P2B rev 1.12 -

http://homepage.hispeed.ch/rscheidegger/p2b_procupgrade_faq.html

This means the owner of a P2B rev 1.12 can use an Upgradeware Slot-T plus
a Tualatin. The processor will run at 1.5 volts because the voltage
regulator can go that low. Finding a Slot-T at around $25 or so gives
the basis for a cheap upgrade.

Since your P2B is less than rev 1.12, you need a Powerleap to run a
Tualatin. Your voltage regulator only goes down to 1.8V (never less
than that), so that is too high a voltage for a Tualatin which has
1.75V absolute max and 1.5V normal operating voltage.

http://plpadmin.tempdomainname.com/PLiP3T.html

There is a iP3-T with a Tualatin 1400/100/256 processor for $119.
If your motherboard supports setting the FSB to 133MHz (where PCI
stays at 33MHz, but AGP becomes 89MHz instead of 66MHz), then you
could take a 1000/100 or 1100/100 processor and overclock it to
get performance closer to the Server version of the Tualatin.

I changed the voltage regulator on my board, which enabled me to use
a Slot-T plus Tualatin. Considering the incidentals purchased to do
the surgery, a Powerleap would have been cheaper.

If you can find a Coppermine Celeron FCPGA, you can
use a slocket with VID jumpers, even with your motherboard, by setting
the VID jumpers to 1.8 volts. I tried this originally with a 1.75V nominal
Coppermine on a modified Startech slocket (added VID jumpers) and it
worked fine. You might actually have more luck finding a 1GHz Pentium
cartridge than a Coppermine FCPGA, in which case there would still be
a requirement to hack a 1.8V VID setting for your voltage regulator
( assuming the cartridge needs less than 1.8V, which you can check
by entering the SSPEC number at processorfinder.intel.com ).

To see one unique way of setting the VID on a slocket without VID
jumpers, see this page, where a U shaped wire is used as a shunt.
(I soldered mine). The example shown here is a one wire mod to run
a 1.7volt processor at 1.8volts. Other processor voltages are a
little more complicated, because more than one VID signal needs
to be modified (may require cutting a track on the cartridge or
fitting wires and switches on the motherboard) -

http://www.tipperlinne.com/p2b-ds.htm

The latest batch of Tualatin processors have a CPUID of 06B4. The
microcode to support this processor isn't in the latest beta BIOS,
so you can add support with low risk, by using CTMC and instructions
found here:

http://www.tipperlinne.com/bios6b4.htm

The main page on the site is here:

http://www.tipperlinne.com/p2bmod.htm

In terms of processors, here are a selected few entries:

SSPEC Core/Bus/Cache CPUID Pkg Vcore Pwr Typ Tmax Geom

Pentium Coppermine with 100 or 133MHz bus (cartridge needs VID
setting hack depending on regulator) -
SL4BR 1000/100/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both
SL4BS 1000/133/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both

Coppermine in FCPGA (Coppermine slocket with VID set to 1.8V)
SL5QV 1000/100/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u OEM
SL5QW 1100/100/256 068A FC-PGA 1.75V 33W cD0 77C 0.18u OEM
SL5DV 1000/133/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u Boxed

Tualatin in FCPGA2 (Slot-T or Powerleap, depending on regulator)
SL6JQ 1000A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JR 1100A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JS 1200/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JT 1300/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JV 1400/100/256 06B4 FC-PGA2 1.5V 33.2W tB1 67C 0.13u Boxed

Tualatin server version (expensive) - same as Tualatin above
[ Benchmarks don't seem to get enough benefit from the cache... ]
SL6BW 1.13-S/133/512 06B4 FC-PGA2 1.45V 28.7W tB1 69C 0.13u OEM
SL6BX 1.26-S/133/512 06B4 FC-PGA2 1.45V 30.4W tB1 69C 0.13u Both
SL6BY 1.40-S/133/512 06B4 FC-PGA2 1.45V 32.2W tB1 69C 0.13u OEM

There are more types than that (some with 128K cache), but that
is all I can figure out from my list of processors. The Tualatins
can still be bought at retail, but have gone out of production.
The other processors will be Ebay material.

Have fun,
Paul

Paul, I don't suppose you could be tempted...

I have a few new retail P2B-S 1.04 boards (complete with cables etc.),
which have been taken out of the box once to install the 4th jumper and
verify stability at 150Mhz FSB, looking for good homes

 

[Paul]

Hello.

Is it posible to get a Tualatin Celeron CPU to work in my old Asus P2b rev.
1.02 ?

I already have a slotkey konverter with volt adjustment and uses the latest
bios (1014.beta2 or something like that). Have been running a Celeron
600@900 so far, but that CPU has to go for an other system and now I need a
new for this one.

Thanks in advance for any answer.


Regards.
Sune
Denmark

According to Roland's web page, the cutoff for Tualatin compatible voltage
regulators is P2B rev 1.12 -

http://homepage.hispeed.ch/rscheidegger/p2b_procupgrade_faq.html

This means the owner of a P2B rev 1.12 can use an Upgradeware Slot-T plus
a Tualatin. The processor will run at 1.5 volts because the voltage
regulator can go that low. Finding a Slot-T at around $25 or so gives
the basis for a cheap upgrade.

Since your P2B is less than rev 1.12, you need a Powerleap to run a
Tualatin. Your voltage regulator only goes down to 1.8V (never less
than that), so that is too high a voltage for a Tualatin which has
1.75V absolute max and 1.5V normal operating voltage.

http://plpadmin.tempdomainname.com/PLiP3T.html

There is a iP3-T with a Tualatin 1400/100/256 processor for $119.
If your motherboard supports setting the FSB to 133MHz (where PCI
stays at 33MHz, but AGP becomes 89MHz instead of 66MHz), then you
could take a 1000/100 or 1100/100 processor and overclock it to
get performance closer to the Server version of the Tualatin.

I changed the voltage regulator on my board, which enabled me to use
a Slot-T plus Tualatin. Considering the incidentals purchased to do
the surgery, a Powerleap would have been cheaper.

If you can find a Coppermine Celeron FCPGA, you can
use a slocket with VID jumpers, even with your motherboard, by setting
the VID jumpers to 1.8 volts. I tried this originally with a 1.75V nominal
Coppermine on a modified Startech slocket (added VID jumpers) and it
worked fine. You might actually have more luck finding a 1GHz Pentium
cartridge than a Coppermine FCPGA, in which case there would still be
a requirement to hack a 1.8V VID setting for your voltage regulator
( assuming the cartridge needs less than 1.8V, which you can check
by entering the SSPEC number at processorfinder.intel.com ).

To see one unique way of setting the VID on a slocket without VID
jumpers, see this page, where a U shaped wire is used as a shunt.
(I soldered mine). The example shown here is a one wire mod to run
a 1.7volt processor at 1.8volts. Other processor voltages are a
little more complicated, because more than one VID signal needs
to be modified (may require cutting a track on the cartridge or
fitting wires and switches on the motherboard) -

http://www.tipperlinne.com/p2b-ds.htm

The latest batch of Tualatin processors have a CPUID of 06B4. The
microcode to support this processor isn't in the latest beta BIOS,
so you can add support with low risk, by using CTMC and instructions
found here:

http://www.tipperlinne.com/bios6b4.htm

The main page on the site is here:

http://www.tipperlinne.com/p2bmod.htm

In terms of processors, here are a selected few entries:

SSPEC Core/Bus/Cache CPUID Pkg Vcore Pwr Typ Tmax Geom

Pentium Coppermine with 100 or 133MHz bus (cartridge needs VID
setting hack depending on regulator) -
SL4BR 1000/100/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both
SL4BS 1000/133/256 0686 SECC2 1.7V 26.1W cC0 70C 0.18u Both

Coppermine in FCPGA (Coppermine slocket with VID set to 1.8V)
SL5QV 1000/100/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u OEM
SL5QW 1100/100/256 068A FC-PGA 1.75V 33W cD0 77C 0.18u OEM
SL5DV 1000/133/256 068A FC-PGA 1.75V 29W cD0 75C 0.18u Boxed

Tualatin in FCPGA2 (Slot-T or Powerleap, depending on regulator)
SL6JQ 1000A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JR 1100A/100/256 06B4 FC-PGA2 1.5V 29.5W tB1 69C 0.13u Boxed
SL6JS 1200/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JT 1300/100/256 06B4 FC-PGA2 1.5V 32.0W tB1 69C 0.13u Boxed
SL6JV 1400/100/256 06B4 FC-PGA2 1.5V 33.2W tB1 67C 0.13u Boxed

Tualatin server version (expensive) - same as Tualatin above
[ Benchmarks don't seem to get enough benefit from the cache... ]
SL6BW 1.13-S/133/512 06B4 FC-PGA2 1.45V 28.7W tB1 69C 0.13u OEM
SL6BX 1.26-S/133/512 06B4 FC-PGA2 1.45V 30.4W tB1 69C 0.13u Both
SL6BY 1.40-S/133/512 06B4 FC-PGA2 1.45V 32.2W tB1 69C 0.13u OEM

There are more types than that (some with 128K cache), but that
is all I can figure out from my list of processors. The Tualatins
can still be bought at retail, but have gone out of production.
The other processors will be Ebay material.

Have fun,
Paul

Paul, I don't suppose you could be tempted...

I have a few new retail P2B-S 1.04 boards (complete with cables etc.),
which have been taken out of the box once to install the 4th jumper and
verify stability at 150Mhz FSB, looking for good homes

I wish I could, but I would have to throw out something, to make
room for something new. I think I need a bigger house

For me, the problem with computers is they take up too much space.
I can justify a high performance machine taking up several cubic
feet of space, but for a surf/email machine, they should be a lot
smaller. Especially now that LCD displays are becoming almost
palatable. I have a spare motherboard just sitting here right now,
but because it is an ATX size, that means another big case if I want
to play with it. It is too bad the ATX motherboard wasn't designed
to allow the excess PCI slots to be sawed off, or disconnected.

To give you some idea how bad my addiction is, I still have a
Timex Sinclair ZX81 computer. About 6" x 6", with a membrane keyboard
(that still is in perfect condition). The thing cost $66 brand new.
Output was to a black and white TV set, via a mudulator. Storage was
via audio to cassette tape. That thing was probably the most fun
I ever had with a computer, because it was so crude, and so much
of a challenge to shoe horn stuff into it. It had a BASIC interpreter
that fit in an 8KB ROM. You could play chess on it - if you could
wait 1/2 an hour for the computer to come up with its next move.
The chess program took 12 to 15 minutes to read off a tape
(no HDTACH needed, because you could count the bytes going by).
I think the processor ran at around 1MHz or so. Standard ram was
2KB, and the expansion pack I bought for it had a whole 16KB of
space.

How times have changed...

Paul

 

[P2B]

Paul wrote:

[snip]

I wish I could, but I would have to throw out something, to make
room for something new. I think I need a bigger house

I went with the bigger house - but I've nearly filled it up again

For me, the problem with computers is they take up too much space.
I can justify a high performance machine taking up several cubic
feet of space, but for a surf/email machine, they should be a lot
smaller. Especially now that LCD displays are becoming almost
palatable. I have a spare motherboard just sitting here right now,
but because it is an ATX size, that means another big case if I want
to play with it. It is too bad the ATX motherboard wasn't designed
to allow the excess PCI slots to be sawed off, or disconnected.

You could always skip the case - my play machine is another dual P2B-DS
1.5Ghz which sits on a sheet of cardboard connected to a scrounged Sun
SCSI drive enclosure, so I can boot my choice of 6 operating systems by
changing the boot disk ID in the BIOS.

To give you some idea how bad my addiction is, I still have a
Timex Sinclair ZX81 computer. About 6" x 6", with a membrane keyboard
(that still is in perfect condition). The thing cost $66 brand new.

Mine cost about half that - it's a Sinclair ZX81 clone, and came in kit
form. It was my first computer, and the only one I've ever assembled
from discrete components. According to the sticker on the box it's
stored in, I last powered it up in 1996 and ran the Yahtzee game I wrote
for it - and have re-written in every programming language I've learned
since. IIRC, the occasion was a side by side comparison of the original
to the first Windows version - written in Smalltalk.

My second computer was also a BASIC machine, a Sharp PC1500 - back when
'PC' stood for Pocket Computer. I used it in college, and when Sharp
published a technical manual for it, with full schematics and complete
documentation of the instruction set used by the proprietary processor,
I decided to code my Labs on it instead of using the 6502 evaluation
boards with hex keypads supplied by the college. That led to my first
surface-mount modification - I needed the processor's non-maskable
interrupt to be available on the expansion connector, but it was tied to
ground (unused) inside. I remember just opening the case was a delicate
operation, and I'd have been in deep financial trouble if I broke the
computer, but the modification was successful thanks to due care and the
excellent manual.

The assignment was to design circuitry and write an assembly language
program which would track a sine wave, i.e. read it from an ADC and
output it via a DAC, and the best marks went to the student whoose
solution tracked the input at the highest frequency before degenerating
into a sawtooth. I was already ahead of the game because the Sharp's
instruction set did more in a given number of clock cycles than the
6502's, but the real breakthough came when I realised I could save a
dozen or so instructions by having the interrupt routine change an
instruction in the main output loop from an increment to a decrement and
back instead of having both refer to a register. I won the speed
competition by about 500Hz, but received a stern warning from the
instructor to the effect I was never to write self-modifying code again!

That tiny computer gave the instructors challenges - we were allowed to
use pocket computers in exams, but the proctors would power them on and
clear the BASIC program memory before we started. I discovered that all
'new 0' did was reset the end-of-BASIC pointer to the beginning of
memory, so I wrote a little machine code routine that found the original
pointer and restored it. I stored my routine in the area reserved for
function keys, so after the proctor was done turning my computer into a
glorified calculator, I simply powered on and hit F4 to get my BASIC
programs back

The instructors got wise to my hack, and started framing exam questions
that were partially worked and often designed to trigger edge conditions
in poorly written programs. This meant I needed to *really* test my code
before exams, then insert a BREAK in the appropriate place, set
variables to the values provided, and restart the program using the
partially worked values during exams - so if I'd cribbed the program
from someone else, it would be worse than useless in an exam. The
instuctors figured I must have learned the material fairly thoroughly if
my code got the correct answers under these conditions, and left it at
that

I still use the PC1500, mostly as a calculator, but I haven't been able
to buy the tiny ballpoint pens it uses in it's 4-colour plotter/printer
for years - which is a shame, because it produced seriously cool Yahtzee
scorecards

How times have changed...

Indeed... imagine the interesting enhancements one could dream up if
motherboard manufacturers published technical manuals with even half the
detail Sharp provided for the PC1500

I never occurred to me to google for the PC1500 until I wrote the above,
and was pleasantly surprised to find http://www.pc1500.com/ , but I'm
impressed the guy has scanned all 160 pages of the technical manual and
posted it as a huge PDF! Now there's a toy I could play with again that
doesn't consume space...

P2B

 

[Sune H.]

Just want to say thanks for all the answers it gave me a clear view over my
posibilities.

Im affraid its a bit to expensive to go for a powerleap and the mod-thing
isn't really my strong side. I think I'll try to find a coppermine
900-1100MHz somewhere cheap.

Thanks.

Regards

Sune / Denmark.

 

[pcguy65]

www.newwaveupgrades.com

Just want to say thanks for all the answers it gave me a clear view over my
posibilities.

Im affraid its a bit to expensive to go for a powerleap and the mod-thing
isn't really my strong side. I think I'll try to find a coppermine
900-1100MHz somewhere cheap.

Thanks.

Regards

Sune / Denmark.