• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Possible solution to the 2200+ and other Tbred in old mobos

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Lafueste

New Member
Joined
Jun 19, 2003
Location
Soria (SPAIN)
EXPLANATION OF THE PROBLEM:

In the option DEFAULT/AUTO of the BIOS (that of beginning for default), with the buffers had disabled and the buffers enabled B (the buffers C is always enabled) the mobo it should recognize the CPU inserted in the Socket based on the configuration of the bridges L3 automatically.
On the other hand, with the BIOS in USER DEFINED (buffers had enabled and buffers disabled B), it is indifferent the configuration of the bridges L3, since the BIOS should be able to change the default disposition fixed in L3, sew that it doesn't happen in some prepared old mobos for CPUs Athlon/Duron that are shown unable to change the 5º bridge L3 (logical Value 8X) because its circuits are prepared alone for 4 bits.
These mobos (with official limit until the 2100+ Palomino and 2000+ Thorougbred) they don't recognize the new CPUs XP appropriately starting from the 2100+ (13X), due to a circuit of existent buffer in the mobo that you feedback in an inadequate way with the values coming from the FID [3:0] that the Northbridge controls.
Due to the mentioned buffer circuit in the configuration DEFAULT/AUTO the value of the Multiplier is modified in its last 4 digits for that of the Northbridge and in the configuration USER DEFINED the value of the Northbridge is modified for the last 4 digits of the Multiplier.

When activating the option DEFAULT/AUTO in the BIOS, the buffers opens up B, for that that when installing a 2200+ with a configuration of bridges L3 for defect of:

O - C - O - C - O

it is sent to the Multiplier Module the logical value: 10101 (O = open bridge = 1, C = closed bridge = 0).

But this value is modified, via buffers C, in its last 4 digits for the value coming from the FID [3:0] that is always above 12,5X: 0011, giving as resultant a logical value: 10011 that it corresponds to the configuration of bridges L3:

O - O - C - C - O

that it corresponds at 4,5X, remapped at 20X, too high value for this CPU whose original configuration is: 13,5X, for what doesn't boot up.

POSSIBLE SOLUTIONS:

The solution to be able to boot up and to have access to the BIOS, when the System refuses to boot he/she consists on closing the 5º bridge L3, but to the right (value of Bit 8X), well in the surface of the CPU with ink or rubber conductive, or in the Socket with a fine wire in form of U invested between the position AJ27 and anyone of the Vss (ground) adjacent (AK28 or AH28) and reseat the CPU, with what should start up at 5,5X, because it corresponds to the configuration of L3:

O - C - O - C - C

and to allow the access from the BIOS in position USER DEFINED to the values between 5X and 12,5X. If some uncertainty of the System is appreciated, to go down the Bus to 100 MHz.
If the System works stable, to update the BIOS to the last version, and then 2 options exist:

Option 1

To leave it like it is, at 12,5X and to prove to increase the speed of the Bus (12,5 x 144 = 1800 MHz), although this can cause uncertainty of the memory, cards PCI, AGP and hard disks.

Option 2

Since the BIOS is unable to modify any bridge L3, if this is blocked in the Socket, uniting two pins by means of a fine wire in form of U invested (uniting Vss (ground) = Close, uniting Vcc (Vcore) = Open), we will be able to use this to get the wanted multiplier without it affects it to the Northbridge.

Therefore, we will put in the BIOS (in option USER DEFINED) the value 12,5X and fixing in the Socket the 5º bridge L3 (value of Bit 8X) in Open (AJ27 with Vcc (Vcore), that is to say with AK26 or AH26), the 3º bridge L3 (value of Bit 2X) in Open (AN25 with Vcc (Vcore), that is to say with AM24) and the 2º bridge L3 (value of Bit 1X) in Close (AL27 with Vss (ground), that is to say with AK28 or AM28). the BIOS sends This way that is to say the Northbridge the code of 12,5X,:

O - O - C - C - C

that it corresponds to a logical value 00011 whose last four digits (0011) they in fact correspond with those that the Northbridge should receive in superior configurations at 12,5X, however when being blocked in the Socket the bridges L3, 2º, 3º and 5º the sign that he/she receives the Multiplier is:

O - C - O - C - O

that is to say the corresponding to 13,5X (logical value 10101), that is to say the official for this CPU.

AS MAKING IT FOR OTHER CPUs OR OTHER MULTIPLIERS:

To write the configuration of L3 wanted for the Northbridge, that is to say 12,5X, and under the one that we want for the Multiplier of the CPU and to verify what values they are different, since these are those that it will be necessary to block in the Socket with the value that you/they have in the configuration wanted for the CPU (O = to connect with Vcc, C = to connect with Vss)


Example: To get 14,5X

Configuration of 12,5X O - O - C - C - C
Configuration of 14,5X O - O - O - C - O
* *

The differences are observed in the 3º and 5º bridges L3, then it will be necessary to block in the Socket the 3º in Open and the 5º in Open (AN25 with Vcc and AJ27 with Vcc).

For more information on the different combinations of L3 and on the topic to consult:
John Carcich Home Page: www.beachlink.com/candjac

For more information about the pin distribution in the Socket to consult:
AMD Athlon XP Processor Model 8 Data Sheet (25175.pdf)

EQUIVALENCE BETWEEN PINS AND BRIDGES:

Host Clock Multiplier (FID [4:0]) (L3) --> AN27 -- AL27 -- AN25 -- AL25 -- AJ27

Multiplier Id (FID [3:0]) --> W1 -- W3 -- Y1 -- Y3

VCore (VID [4:0]) (L11) --> L1 -- L3 -- L5 -- L7 -- J7

FSB (FSB_SENSE [1:0]) (L12) --> AG31 -- AH30



Gratefulness to John Carcich (www.beachlink.com/candjac) for the necessary technical information to make this article.
 
Back