Easiest Way to Unlock the XP2100

We're referring to the former mod which reverses the L10 bridges so that mobos can boot and recognize the 2100 as a 5X Multiplier, then get control of the 5X thru 12.5X Multiplier range with the bios.

The L10 bridges reversal mod, (and the wire and pins alternate mods), all reset the 8X Bit Value from default HI to LO. The wire and pins mod/s do that by "grounding" the 8X Signal Bus.

So just had the "AHA" syndrome...ie, easiest way to ground that bus is to NOT insulate the pit in the rightmost L1 bridge when closing all the L1 bridges. Then when closing all the L1s with conductive fluid, that last one should connect to the sub-surface ground plane, (that you're trying to prevent in the other 4 L1s).

Voila', that grounds the 8X Signal Bus, just the same as the wire and pins mod... the default HI is reset to LO.
John C.

thanks John,
now would you let the little gray cells figure out how to allow my Shuttle AK31 to operate with that 8X hi? :>)
wj

Will have to try in the near future. Thanks again for some great information! :)

Quote:

---

Originally posted by wymjym
thanks John,
now would you let the little gray cells figure out how to allow my Shuttle AK31 to operate with that 8X hi? :>)
wj

---

Don't we all wish someone knew.;-) Seriously, must be a mobo/bios issue...something like they never considered a 5th 8X bit since the early XPs had default Multipliers less than 13X...as did the Duron/Tbirds which only had 4 bit Multiplioer code. We've all too spoiled being able to pop cpu after cpu into the same Socket A, and somewhere along the way something became lacking.

But Speculation...maybe the problem is in the Northbridge where the FID Multiplier "ID" signals from the cpu are sent for setting up system timings etc. These are only 4 bit, even for 5 bit Palominos and Tbreds. And recall that Duron/Tbirds had to have the FID 4 bit Multiplier "ID" signals equal or greater than the Multiplier...else no boot. So with only 4 bits being sent to legacy Northbridge and so ignoring the 8X Bit Value setting, it might be making timing settings for a 5X Multiplier cpu...could that be the problem...the Northbridges???...which are not upgradable???

If so, then might be able to boot a 2100 by hard modding the FID signals to 12.5X, which needs to have the .5X and 1X Bit Values HI and the 2X and 4X BVs LO. Datasheets say the data Northbridge uses/sends is the same for 12.5X and greater.

But the problem is that there are no separate bridges for the FID signals, the pins are still there, but the signals seem to be auto-set/biased by the 1st 4 bits of the Multiplier signals internally within the cpu...see Palomino article. So to test this speculation, user would have to open the 4 traces from the 4 FID pins to the Northbridge, then connect Vcore to the .5X and 1X traces, and ground the 2X and 4X traces. That would tell the Northbridge to use data for a 12.5X Multiplier.

Quite a mod....Any volunteers???
John C.

PS:- http://www.beachlink.com/candjac/index.htm link to any articles referenced above...some article/s have pic/s which locate the FID pins.

And thanks for your comments Frank.

So if I am to understand correctly you fill in the first 4 L1 Bridges like you would normally do. Then the bridge on the far right dosent get the crazy glue in the middle like the others. You just use the the rear window defogger repair fluid and completly fill the bridge without the crazy glue therfore saturating the pit with the fluid and filling from side to side? hehe I'm new to overclocking and I just want to make sure I do this right. Anyone else try this?

Quote:

---

Originally posted by stylz
So if I am to understand correctly you fill in the first 4 L1 Bridges like you would normally do. Then the bridge on the far right dosent get the crazy glue in the middle like the others. You just use the the rear window defogger repair fluid and completly fill the bridge without the crazy glue therfore saturating the pit with the fluid and filling from side to side? hehe I'm new to overclocking and I just want to make sure I do this right. Anyone else try this?

---

You read it correctly stylz, and no one has tried it yet...but as we explained, analysis of the relevant "circuits" says it should work...and recall there have been several different mods to that 5th 8X Bit Value...like reversing the L10 bridges, and the pins mod using both wire or conductive ink...and they all do the same thing...reset that 8X Bit Value from HI to LO...so grounding the 5th L1 bridge to the sub-surface ground plane exposed in its pit is no different...we'd just be sure to look into the pit (with some magnification) to be absolutely sure that ?copper? ground plane "is" exposed...and if not scratch around a bit with a needle to do expose it...but also after all, if we're insulating the other 4 pits to "avoid" shorting to ground, then very good likelyhood those ground planes "are" exposed. So be a pioneer, do it, and post you result!!
John C.

So let me get this straight, (this is confusing me a little)

To unlock XP2100+ processor

1: Fill in all L1 pits with non-conductive substance (apart from last one)

2: Close all L1 bridges including last one making sure to fill the pit with conductive ink

3: Thats it??

Or do you need to do the L10 swap as well?

Thanks

Thief

sorry, was giving bad info.
wj

Quote:

---

Originally posted by CodeThief
So let me get this straight, (this is confusing me a little)
To unlock XP2100+ processor
1: Fill in all L1 pits with non-conductive substance (apart from last one)
2: Close all L1 bridges including last one making sure to fill the pit with conductive ink
3: Thats it??
Or do you need to do the L10 swap as well?
Thief

---

You read it right Code Thief, and no need to do the L10 swap...'cause the L10 swap sets a LO to the 8X Bit Value. The "successful/proven" pins mod resets a LO to that BV by grounding that signal bus..and no one does the L10 swap when they do the pins mod...right???
And closing the 5th L1 and grounding it at the same time by not filling/insulating the pit does exactly the same as the pins mod...grounds that 8X BV signal bus.

So should work...as long as ground plane is exposed in that pit. So someone...do it and verify...please.
John C.

Thanks for the reply Candjac. I may try this tonight, hell it's only £100 if it goes wron :eek: :D

Thief

Quote:

---

Originally posted by CodeThief
Thanks for the reply Candjac. I may try this tonight, hell it's only £100 if it goes wron :eek: :D
Thief

---

Thanks, and hope you try it....don't believe it can do any harm as it should only do what pins mod does...ground that signal bus. Only thing that may go wrong is that the grounding for some reason does not take effect...would be your luck to have the reported sub-surface ground plane not exposed.

But then....
1:- Why are we insulating all the other pits??
2:- Sloppy insulating of L1s has resulted in many cases of "unexpected" Multipliers...all of which have been explained by decoding which pointed to particular L1s being LO/grounded when the bios setting would have had them HI. Then repairing that particular L1 mod fixed.

Which finally, says the grounding is restorable, albeit might need a bit of work...acetone on a q-tip???

Has anyone sucessfully tried this unlocking mod? It is by far the easiest technique I have seen for the xp2100. I have been nervously staring at my CPU for several days now wanting to unlock it but I'd love some confirmation before I start.

I guess someone needs to be first....but then again if you are thefirst to report doing it...then everyone else will say, 'oh yeh, I did that a couple of months ago'
wj

DegreeC,

Hopefully someone will stop by that has done it...but in the meantime, welcome to Xtreme :)

:(

Just tried this this evening. Spent ages making sure it was PERFECT!!

Wouldnt let me post at any multiplier other than AUTO on a GigaByte GA7VRXP.

Black screen, no post no nothing :(

Thief

Quote:

---

Originally posted by CodeThief
:(Just tried this this evening. Spent ages making sure it was PERFECT!!

Wouldnt let me post at any multiplier other than AUTO on a GigaByte GA7VRXP.

Black screen, no post no nothing :( Thief

---

What a bummer...but did you inspect the pit to check that the sub-surface ground plane was visible?? Only other possibility is it may have been covered with film/debris.
Should work as it's no different from reversing the L10s, the pins/wire mod...they "all" just ground the 8X signal bus. A ground is a ground is a ground he mumbles.....(All the other 4 L1s closed w insulation OK??)

Final Q??....How can the system boot at default (8X Bit Value HI) if its signal bus is grounded??
John C.