PIII 600 in a p2v

  • Thread starter Thread starter Luddite
  • Start date Start date
L

Luddite

I have a known good p3 600 at 1.65v that wont work in my p2v
it has 1004 beta 05,should it support this voltage?
I left the bus at the old settings of 83x3.5
any tips on this one or is time to upgrade?
 
I have a known good p3 600 at 1.65v that wont work in my p2v
it has 1004 beta 05,should it support this voltage?
I left the bus at the old settings of 83x3.5
any tips on this one or is time to upgrade?

The first question is, will a P3 work on that board ? I don't
know anything about that generation, and these are a few links
I dug up.

http://groups.google.com/[email protected]

This one says no support for Coppermine processors. Now, maybe this
is a BIOS thing, or something else.
http://www.asus.com/inside/Techref/images/tr19.jpg

OK, this page says:

http://www.asus.com.tw/support/cpusupport/cpusupport.aspx
P3-600 MHz(slot1) ALL BIOS 1003 or later

When you plugged in the processor and attempted to power up,
did the fans spin but there was no POST ? That means the 1.65V
the processor requested, cannot be provided by the voltage
regulator. The way some of these regulators work, is they
support 1.8V or more. Attempts to send the code for 1.65, 1.7,
1.75 etc., actually result in the regulator making 0 volts.
One way to fix this (which I've used), is to change the VID
code fed to the motherboard to 1.8V, giving the processor
a little more voltage. This is OK for Coppermine processors,
but not for something like a Tualatin S370 which uses 1.5V.

Here are two voltage regulator datasheets. The first one, for the
HIP6004A, only supports down to 1.8V. The second one, pin for pin
compatible, goes down to 1.3V. Compare the VID tables containing
the five bit VID codes and the resulting voltage.

http://www.intersil.com/data/FN/FN4/FN4417/FN4417.pdf
http://www.intersil.com/data/FN/FN4/FN4567/FN4567.pdf

Voltage VID4 VID3 VID2 VID1 VID0
1.65V 0 1 0 0 0
1.80V 0 0 1 0 1 <--- You want these values

Now, the way the VID signals work, is the voltage regulator has
pullup resistors built in. If nothing is connected to the VID pin
on the voltage regulator, a logic "1" results. If the VID pin is
grounded, a logic "0" results.

The VID driving circuit on processors has varied a bit over the
years, and the latest processors _may_ have a full driver in
them. Older processors drive to ground or are open circuit on
the VID pins. (Pumping more than 5 milliamps into one of these
grounds, will damage it, so that is why just connecting the pin
to VCC won't work without the potential for collateral damage.)

What this means is, if your vintage of processor makes a logic "1"
on a specific VID signal (an open circuit), you can make it into a
zero, safely, by grounding the signal. This can be done, for example,
with a U shaped bridge inserted in the holes next to the Slot 1 socket.
Now, the example on this web page, is modifying VID1, and you need to
modify VID3 from a "1" to a "0". This means finding a GND on the
connector, near to B119. (The 244452 datasheet mentioned below, shows
there are no GND signals on the B side of the connector, so a longer
wire to reach the A side would be needed.)

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

VID4 (Pin A121)
VID3 (Pin B119)
VID2 (Pin A119)
VID1 (Pin A120)
VID0 (Pin B120)

The hardest part, will be opening the circuit on VID2 and VID0. What
I did on my slocket, is cut the wire leading away from the gold pins
on the edge connector, as this breaks the GND signal coming from the
processor, to VID2 and VID0 on the motherboard. I use a hobby knife
(Xacto), and cut a 1mm section out of the track with two cuts. This
leaves a visible gap, suitable for verifying the connection is cleanly
cut. I make the cut far enough away from the edge card contact, that
I can solder a bridge wire in the future, if it is ever necessary
to return the module to "factory" condition.

Since you are going to that much trouble, a soldered GND connection from
VID3 to GND, might be marginally more stable than using the U shaped
wire. If you don't own any soldering equipment, then the U shaped wire
is a fine solution.

http://developer.intel.com/design/pentiumiii/documentation.htm
http://developer.intel.com/design/pentiumiii/datashts/244452.htm (pg.72)

If you look at page 24 of the datasheet, the max operating voltage
is listed as ("Operating Voltage plus 1.0 volts") and you are lifting
the voltage by a safe 0.15 volts. That is why this can work. (On a
Tualatin processor, 1.75V is max, and that is why this style of mod isn't
healthy for a Tualatin.)

If you are uncomfortable modifying the module, then perhaps someone
like "Homey" http://motherboardrepair.com/ could do it for you. You
would need to tell him the required mod, like including this post, so
he can get the details. I don't know how much exposed track there
is on the module to cut, as the slocket I modified, is a different
animal than a Slot1 processor module.

Another possibility, is to use tape or finger nail polish, to insulate
the pins you want the logic "1" on. I dislike modding that way, as
the chemicals you use, can foul the contact on the SC242 connector.
You cannot easily clean a socket like that, so if some gunk is left
in there, the motherboard may never work again. That means, whatever
insulating material you use, should be thin and very hard, so it
cannot be scraped off when the module is inserted. It also should not
interfere with adjacent pins (a tall order). That is why I like
cutting copper tracks instead.

In summary, insulate A119 and B120. Connect B119 to A118 (GND) with a
long wire, as shown on http://tipperlinne.com/p2b-ds.htm . Leave the
body of the wire insulated, leaving stripped ends to plug into the
holes on the SC242 connector. A nickel plated wire is best - bare
copper will oxidize over time.

HTH,
Paul
 
Luddite said:
I have a known good p3 600 at 1.65v that wont work in my p2v
it has 1004 beta 05,should it support this voltage?
I left the bus at the old settings of 83x3.5
any tips on this one or is time to upgrade?

I'd say your chances of getting this to work are good, because P2V BIOS
1004 beta 5 contains microcode updates for x0681 and x0683 processors -
which are both Coppermines (.18u process).

Your problem is almost certainly CPU core voltage, so if you do the VID
modification as detailed by Paul, it should work.

HTH

P2B
 
Paul said:
The first question is, will a P3 work on that board ? I don't
know anything about that generation, and these are a few links
I dug up.



The manual says so but the 600 was not on the list,just up to 550
tipperlinne.com was good I'll see where those pins are
now and cut or jump them above the contacts on the CPU so it asks for 1.8
what pin can I use as a ground reference?

This one says no support for Coppermine processors. Now, maybe this
is a BIOS thing, or something else.
http://www.asus.com/inside/Techref/images/tr19.jpg
OK, this page says:
When you plugged in the processor and attempted to power up,
did the fans spin but there was no POST ? That means the 1.65V
the processor requested, cannot be provided by the voltage
regulator. The way some of these regulators work, is they
support 1.8V or more. Attempts to send the code for 1.65, 1.7,
1.75 etc., actually result in the regulator making 0 volts.
One way to fix this (which I've used), is to change the VID
code fed to the motherboard to 1.8V, giving the processor
a little more voltage. This is OK for Coppermine processors,
but not for something like a Tualatin S370 which uses 1.5V.
Here are two voltage regulator datasheets. The first one, for the
HIP6004A, only supports down to 1.8V. The second one, pin for pin
compatible, goes down to 1.3V. Compare the VID tables containing
the five bit VID codes and the resulting voltage.

Voltage VID4 VID3 VID2 VID1 VID0
1.65V 0 1 0 0 0
1.80V 0 0 1 0 1 <--- You want these values
Now, the way the VID signals work, is the voltage regulator has
pullup resistors built in. If nothing is connected to the VID pin
on the voltage regulator, a logic "1" results. If the VID pin is
grounded, a logic "0" results.
The VID driving circuit on processors has varied a bit over the
years, and the latest processors _may_ have a full driver in
them. Older processors drive to ground or are open circuit on
the VID pins. (Pumping more than 5 milliamps into one of these
grounds, will damage it, so that is why just connecting the pin
to VCC won't work without the potential for collateral damage.)
What this means is, if your vintage of processor makes a logic "1"
on a specific VID signal (an open circuit), you can make it into a
zero, safely, by grounding the signal. This can be done, for example,
with a U shaped bridge inserted in the holes next to the Slot 1 socket.
Now, the example on this web page, is modifying VID1, and you need to
modify VID3 from a "1" to a "0". This means finding a GND on the
connector, near to B119. (The 244452 datasheet mentioned below, shows
there are no GND signals on the B side of the connector, so a longer
wire to reach the A side would be needed.)

VID4 (Pin A121)
VID3 (Pin B119)
VID2 (Pin A119)
VID1 (Pin A120)
VID0 (Pin B120)
The hardest part, will be opening the circuit on VID2 and VID0. What
I did on my slocket, is cut the wire leading away from the gold pins
on the edge connector, as this breaks the GND signal coming from the
processor, to VID2 and VID0 on the motherboard. I use a hobby knife
(Xacto), and cut a 1mm section out of the track with two cuts. This
leaves a visible gap, suitable for verifying the connection is cleanly
cut. I make the cut far enough away from the edge card contact, that
I can solder a bridge wire in the future, if it is ever necessary
to return the module to "factory" condition.
Since you are going to that much trouble, a soldered GND connection from
VID3 to GND, might be marginally more stable than using the U shaped
wire. If you don't own any soldering equipment, then the U shaped wire
is a fine solution.
If you look at page 24 of the datasheet, the max operating voltage
is listed as ("Operating Voltage plus 1.0 volts") and you are lifting
the voltage by a safe 0.15 volts. That is why this can work. (On a
Tualatin processor, 1.75V is max, and that is why this style of mod isn't
healthy for a Tualatin.)
If you are uncomfortable modifying the module, then perhaps someone
like "Homey" http://motherboardrepair.com/ could do it for you. You
would need to tell him the required mod, like including this post, so
he can get the details. I don't know how much exposed track there
is on the module to cut, as the slocket I modified, is a different
animal than a Slot1 processor module.
Another possibility, is to use tape or finger nail polish, to insulate
the pins you want the logic "1" on. I dislike modding that way, as
the chemicals you use, can foul the contact on the SC242 connector.
You cannot easily clean a socket like that, so if some gunk is left
in there, the motherboard may never work again. That means, whatever
insulating material you use, should be thin and very hard, so it
cannot be scraped off when the module is inserted. It also should not
interfere with adjacent pins (a tall order). That is why I like
cutting copper tracks instead.
 
Back
Top