Unlock your XP2100 without cutting Bridges!!

[Chong345]

Very Nice OPP. I wish I had known this before I sold my 2100 2 months ago cause I didn't think it could be unlocked.

[Kefka]

ok, im still a little clear on this. To overclock my 2100+ palimino, I close all the L1 bridges, and do the pin mod correct?

and when closing the bridges should the conductive material be in the pits too? or should I fill those with a non conductive material?

[OPPAINTER]

To fill the pits you use NON conductive material like super glue, and to connect the bridges you use Conductive Material.
And yes, close the L1 bridges and do the pin mod to unlock the the 2100.

OPP

[Kefka]

might i inquire as to what happens when u fill the pits with conduction material?

[majormav]

whooooooooooooo this 2100 is getting the treatment asap a new toy to play with might send the p4 bench rig to the for sale column in local paper

[majormav]

a big thanks to all done the mod and chip is unlocked .... thanks

chip is agoia week 16

[DaGooch]

Just an update about the pin modification with L1 modfication on a Palamino 2100+. The modification worked fine using an AT7 and KX7 but would refuse to boot on an AK35GTR with the latest bios. Due to differences in implentation of bioses and hardware pin signal routing on the motherboard, the modification will work on some motherboards but not all.

[Bravo]

A question, has anybody tried to identify the pins that use the l1 gates?

I've had a think about it, and the advantage would be the board would unlock the processor with no problems, and EVERY processor you plugged in would not require an unlock job

[candjac]

Originally posted by Sinnah
A question, has anybody tried to identify the pins that use the l1 gates?

I've had a think about it, and the advantage would be the board would unlock the processor with no problems, and EVERY processor you plugged in would not require an unlock job

Sinnah, Yes the pins have been identified, and they are documented at
http://www.beachlink.com/candjac/index.htm link to 1st Palomino article. They are identified by pin # like AJ27, also in a picture of pins, and also in diagram/s associated with particular L1 bridges, and Bit Value signal busses.

Then in the Tbred article there is a circuit diagram of a simple switched or jumpered small Radio Shack printed circuit board/pcb that a user could construct and then connect the wires to the 5 socketss connected to the L1 bridges marked on the same diagram...also 2 wires to Vcore and ground. That diagram has "all the necessary info in one place".

Edit:- If user just wants to be able to handle the 5th Bit L3 "workaround", then just one wire frm AJ27 plus Vcore and Ground wires are needed for a "switched/jumpered" mod...covered in Tbred article. Or, user could just solder one wire from AJ27 to nearby Ground pin = AK28 or AH28 for a "permanent unswitched" 5th L3 mod to the mobo.

However the L1 bridges must be closed for those circuits to be able to over-ride the default Multiplier settings..no problem on Tbred 2200, but L1s Xp 2100s and lower must be closed by user.
John C.

[DaGooch]

Found out why the pin mod did not work on the Shuttle AK35GTR. Recently a phone call to OP and his latest adventures in cooling experiments, the boot up problem on the AK35GTR came up. He asked me to check the bios on the AK35GTR and see if there was a 5X multiplier with the latest bios release and when I checked, there was none. The lowest multiplier available in the most recent bios was 5.5X.

[R3N3G@D3]

Heloooo

If i was to change the L10 Bridges over (closed/opened - opened/closed) this would then make it a 665 mhz processor (5x133). I was just wondering would the processor boot at that speed (Running on an Epox 8K3A+ 2508 bios).
Or would I have to change the L3 bridges to make it the same as a 1700+ XP (1463 Mhz). Or would I have to use another processor to set the FSB or Multi???
Also how exactly does the pin mod work ???
Could some one please post some illustrating pictures that make it a little more clear to understand ???

ThanX

[candjac]

Originally posted by R3N3G@D3
If i was to change the L10 Bridges over (closed/opened - opened/closed) this would then make it a 665 mhz processor (5x133). I was just wondering would the processor boot at that speed (Running on an Epox 8K3A+ 2508 bios).
Or would I have to change the L3 bridges to make it the same as a 1700+ XP (1463 Mhz). Or would I have to use another processor to set the FSB or Multi???
Also how exactly does the pin mod work ???
Could some one please post some illustrating pictures that make it a little more clear to understand ?

Most if not all mobos do post/boot at 5X.

So should be no need to manipulate other multiplier bridges...just close the L1s of course.

All variations of the pin mod simply ground the 8X Bit Value signal bus, which "permanently" resets it from default HI to LO...which is exactly what reversing the L10 bridges does. We'd do the single wire from AJ27 socket backside to case ground...in post above...easy to do and undo.

http://www.beachlink.com/candjac/index.htm various articles on AMD CPUs' Multipliers, codes etc, if interested...also circuit diagrams, you could use to see the grounding effect.
John C.

[salocin]

Im very new to unlocking And overclocking. But i think Ive unlocked my 2100 + THANKS, TO , YOU

My system specs

AMD 2100+
Epox 8k5a2+
512 mb Samsung pc2700
80 gig Maxtor

Just 1 question where did my 13* multiplier go, it now boots at 5 my 14* also at 5 and my 15 at 5.5

[candjac]

Salocin,
If you read the articles and threads you will realize that when you ground the 8X Bit Value by reversing the L10 bridges or the AJ27 pin mod, you lose ability to control any Multipliers above 12.5X.

What you're getting seems to be some erratic settings that your bios cobbles up because it can't reset the 8X Bit Value "hard ground" to HI again.

Overclocking the XP 2100 should be done with Multipliers at 12.5 and lower, combined with higher FSB frequency.
John C.

[salocin]

Thanks for the reply

Ok so if I could run at a 13 multiplier before at 145 mhz
and i can now get 11 multi at 175mhz

Thats an improvement althuogh slight ,Its a start

Any other suggestions on how to get more ?


Im a bit sorry I didnt get the 8k3a+ board there seems to be a lot more info on it .

I have read and tried to under stand the stuff I printed from tbredoc and it doesn't seem to apply to the 2100

Just one more thing if I go to ocinside and select the setting I would like run ,do i just copy that painting

thanks again

Exterme novice.com

[candjac]

Salocin wrote...
>Ok so if I could run at a 13 multiplier before at 145 mhz
>and i can now get 11 multi at 175mhz

If your memory is up to 175Mhz.

>Any other suggestions on how to get more ?

We specialize in Multiplier code, we'll let the more experienced overclockers answer this.

>I have read and tried to under stand the stuff I >printed from tbredoc and it doesn't seem to apply to >the 2100

Well the Multiplier bridges are different, but the Multiplier Code is exactly the same.

>Just one more thing if I go to ocinside and select the >setting I would like run ,do i just copy that painting

Sounds like your mobo has no Multiplier support for below 13X in addition to not being able to handle 13X and up???

If so, Yes copy Jens' layout, but be careful about insulating the pits, not cross-connecting, and don't cut too deep when opening. Best tool is finest pointed/narrowest angled diamond coated burr from your dentist. Epoxy into tube for handling like a pencil with finger strokes slowly making number "ones" across the bridge...with heel of hand fixed/steady. Use 5X eye loupe to inspect cutting progress.
John C.

[OC Detective]

I am a firm believer in that a moments stupidity is better than a lifetime of ignorance so I will ask this question!:p
I have an XP2000+ and where I live it is difficult to obtain conductive pens etc - so is there a way to wire mod the pins on the cpu to replicate closing the L1 bridges? I would assume that the pin mod workaround for the last bridge on the L3 to obtain multipliers lower than 12.5 would also have to be incorporated.

[N8]

OC Detective,

The default multiplier on the 2000+ is 12.5...so you only need to connect the L1's on it. You only connect the last L3 on the 2100+ because its mult is higher than 12.5.

Regarding whether there is a pin mod for unlocking yours, sorry to say, but I doubt it. Connecting the L1's is what you need to do to aquire all mults up to 12.5x. This has been done by soldering a small wire between the traces on the chip, but I'm sure it'd take a very well experienced solderer to perform such an act. Conductive paint will surely be your best bet, unless its impossible to find in your area.

Good Luck!

[OC Detective]

Sigh, thats what I feared - I guess I will just have to try and import some.

[Ajax]

Nice info guys but can you answer this question?

How can you unlock the XP2100+ but keep the highbit x8 enable?

The reason is that I use a 100FSB (KT133) and I can only play with the multiplier to increase the speed of my 2100. Otherwise I'm stuck at 1300Mhz (13 x 100). I need to add the x8 to my multiplier and get higher value i.e. (Multiplier at 7 = (7+8) = 15, then 15 x 100FSB = 1500Mhz etc...)

Thanks guys. I really appreciate your help !

Ajax

[candjac]

Originally posted by Ajax
Nice info guys but can you answer this question?

How can you unlock the XP2100+ but keep the highbit x8 enable?

The reason is that I use a 100FSB (KT133) and I can only play with the multiplier to increase the speed of my 2100. Otherwise I'm stuck at 1300Mhz (13 x 100). I need to add the x8 to my multiplier and get higher value i.e. (Multiplier at 7 = (7+8) = 15, then 15 x 100FSB = 1500Mhz etc...) Ajax

Ajax,
Your 8X Bit Value is already set HI/Enabled. But there is a chance to try for higher Multipliers, has worked on some mobo/s.

1: Close all the L1s except the rightmost, (saves some work, and prevents bios from changing the rightmost 8X Multiplier signal circuit, but others need to be closed for mobo to be able to change other 4 signal circuits as needed). Be sure to insulate the pits else all signal circuits go to ground/LO when you connect the L1s and mobo can't change those grounded LOs.

2: Set 8X less than what you want in the bios Multiplier setting option. Example, you want 15X, set bios to 7X. Hopefully the bios will rest the default multiplier with signals to the 1st 4 bits, (currently set to 5X), to 7X. Normally, a new mobo would also "disable" the 8X bit as well, set it LO, so the system would boot at 7X. But.....
A: Your older mobo very possibly is not designed to manipulate the 5th 8X bit. If so, it won't get reset LO, stays HI and the system boots at 15X...hopefully. Or...

B: Because you did not close the right most 8X L1, even if your mobo should send an over-riding LO to the 8X signal circuit, it won't get there to change the existing default HI...which you need as in "A:".

http://www.beachlink.com/candjac/index.htm link to Multiplier Code article to "understand" why this might work....CPU doesn't care "how" its 5 signal circuits are set/reset with HIs and LOs. This procedure tries to get the mobo to "reset" the 4 lowest Bit Values and keep the 5th 8X Bit Value at its default HI.
Also read Palomino article, maybe also Circuits article to see how this might work. Should work "in theory", but sometimes particular mobo/bios/?Northbridge? might have system setup code that causes problem like no boot. But has worked, worth a try. Let us know if it works. 14.5X and 15.5X are "Reserved"...might not work???
John C.

[deks]

I followed all your instructions to a T on my AGOIA XP2100+. Did the pin mod and painted the l1's. Yes i filled the pits with heatsync compound. I did the L1's about 3 times before I gave up. First time the PC booted up. 1st. attempt. It said I had a 933 or something. So I went to the bios to bump up the mulitplyer. Set it to like 10. Got in to windows looked at WCPU and the multiplier was still 13, but my FSB was 198. 2nd attempt. Cleaned l1's filled the pits and now It booted as a 1500. Try to set the ratio. And it would lockup. Booted into windows and I was running 198fsb and 13.X 3rd attempt, I had got 1700+, same thing couldnt adjust the multiplyiers. (Well the bios lets me but it dont take.) I couldnt raise or lower the FSB or I wouldnt get post. SO I had to leave the bios at what ever it was detecting at the time. Any Idea's? I use conductive paint. Very steady hand and good eye sight. I know for sure the L1's had to be unlocked. But why different results every time I retried painting the L1's? I have an A7N8X, and all this time i had the chip modded, it would not let me run at 2100+ speed no matter what. I really want to unlock this. Any Idea's? I noticed that link http://www.beachlink.com/candjac/index.htm should the FSB HI LO where different pins than that chinnease guys wire trick. Does this matter? I'm really desperate. I want to see how far i can raise this FSB on this puppy.

[Ajax]

Thanks John,

Can I assume by reading your post that your proposal of unlocking this 2100+ for the KT133 doesn't require any bridges cutting?

If I understand properly I only need to:

1. Close all the L1s except the rightmost

2. Set the jumpers (on the mobo or in the Bios) at a lower multiplier since the x8 bit is still enable. Therefore If I select 7x I should get a frequency of (7+8) x 100fsb = 1500Mhz

Is this right?

Thanks John, really appreciated

Ajax

[candjac]

Deks,
Not sure if this is meant for us/ our previous post, but following applies when testing "Multiplier changes". You should try to do so at 100 MHz FSB to avoid any possibility of failures from too high a cpu clocking. Once you've "proven" you can get the system to change Multipliers at will, that's the time to "optimize". That said, comments....

Heatsink compound not a good choice for insulating material... never really hardens so conductive ink might weep around it and make connection to sub-sutface ground plane...or even worse if Arctic Silver is used, yes not supposed to be electrically conductive, but even a bit might be NG, and does not harden as above. Suggest a good cleanup and redo L1s with a more "solid" insulating material, as the "changing results" point to a possible variability resulting from the L1s closing mod using hs compound.

On the other hand, the L10s/AJ27 to Gd pins mod changes the 13X default Multiplier to 8X less = 5X. So somewhere along the line you should have seen 5X. Then you go to the bios and have control from 5X thru 12.5X...but not higher, cause doubt that even your new mobo can reset the L10/pins mod grounded/LO 8X Bit Value back to HI again.

In brief, if you want 13X, restore the chip as found. If you want 5X thru 12.5X close the L1s carefully as above and do the AJ27 to GD pins mod. Should boot at 5X then change to suit in bios. Always "test" first at 100 Mhz if you can.
John C

[candjac]

Originally posted by Ajax
Can I assume by reading your post that your proposal of unlocking this 2100+ for the KT133 doesn't require any bridges cutting?

If I understand properly I only need to:

1. Close all the L1s except the rightmost

2. Set the jumpers (on the mobo or in the Bios) at a lower multiplier since the x8 bit is still enable. Therefore If I select 7x I should get a frequency of (7+8) x 100fsb = 1500Mhz Is this right?

You read it all right Ajax, let's hope your mobo/system "co-operates".

You see, changing the Multipliers also involves code in the Northbridge, tables for system setup/"timing" during the boot process for example. And the Northbridge gets Multiplier info from "another" set of pins, not the ones connected to the bios for the mobo/bios when changing. So there's always the possibility of a "mismatch" (esp when "modding") which might cause the system to not boot. We only know this in general from reading about it in datasheets but exact logic/details are not available, and might not understand them even if available. Might be similar situation in bios?? But this has worked on some mobo/s in past, Good Luck.
John C.