Help with XP1600+ unlocking

[FPSguy]

I tried to unlock my Athlon XP1600+ (Newegg, AGOIA stepping) this weekend. I used the "superglue, rear window defogger" method. When I was done, I put the chip in a box with an Asus A7A266 motherboard (yes, I know it's not a good overclocker ...). The motherboard seems to recognize the chip as an Athlon XP1900+ and runs it at 1600 MHz, but it won't let me set the processor to run at any speeds other than 1600 or 1200 (12 x 133, 12 x 100, or 9 x 133). Can anyone diagnose this? Is it a sign that I mistakenly connected some bridges or grounded the bridges or didn't fill all the pits or something? Or, do I need to adjust a jumper on the motherboard to get above a 12x multiplier? I am supposed to be able to adjust multipliers through the BIOS, but there is something in the manual about a dip switch for multipliers 12.5x and above. But then again if it was unlocked properly I would think at a minimum I would have a choice of any multiplier at 12 or below.

Thanks in advance for any help anyone can provide!

 

[johncar]

Originally posted by: FPSguy
I tried to unlock my Athlon XP1600+ (Newegg, AGOIA stepping) this weekend. The motherboard seems to recognize the chip as an Athlon XP1900+ and runs it at 1600 MHz, but it won't let me set the processor to run at any speeds other than 1600 or 1200 (12 x 133, 12 x 100, or 9 x 133). Can anyone diagnose this? Is it a sign that I mistakenly connected some bridges or grounded the bridges or didn't fill all the pits or something? Or, do I need to adjust a jumper on the motherboard to get above a 12x multiplier? I am supposed to be able to adjust multipliers through the BIOS, but there is something in the manual about a dip switch for multipliers 12.5x and above.

1600 default Multiplier = 10.5X for which AMD sets the Sum of Bit Values HI to 7.5X, which means the .5X + 1X + 2X + 4X Bit Values are HI and the 8X BV is LO.

12X is set by having only the 1X BV HI, so we could say that you might have accidently grounded the .5X, 2X, and 4X BVs, which sets permanent LOs which cannot be reset HI again.

However, the 9X Multiplier which is set by having the 2X + 4X BVs HI contradicts the 12X analysis above which concluded the 2X and 4X BVs might have been reset permanently LO.

So we can't make any solid conclusions. Suggest you research the mobo for possible jumper settings etc you may have missed, before redoing the L1 bridges.

http://www.beachlink.com/candjac/index.htm link to Understanding Multiplier Code and Palomino articles for more details and to understand what we've written.
Good Luck, John C.

 

[FPSguy]

Thanks! I certainly learned something from that link. The fact that my chip/MB go to 12x and 9x, but no other multipliers, is strange. I don't see anything about it in the manual, though I will keep looking. If anyone has experienced this (or something similar) before and has any other tips, I would love to save some time diagnosing this ... 🙂

 

[dmcowen674]

Hi FPS,

I have the A7A266 as well and ordered a XP1600+<FONT size=1>AGOIA stepping from VEgames last friday since it was $54. I'll post on here how it does with mine. I only have PC2100 Memory in the board right now but it is running solid at 142 FSB (284). I hope to hit 160 with the new chip, we'll see.</FONT>

 

[johncar]

Originally posted by: FPSguy
Thanks! I certainly learned something from that link. The fact that my chip/MB go to 12x and 9x, but no other multipliers, is strange. I don't see anything about it in the manual, though I will keep looking. If anyone has experienced this (or something similar) before and has any other tips, I would love to save some time diagnosing this ... 🙂

A suggestion...get more data...ie...try ALL settings in the bios and post back with a Set vs Get Table. Perhaps we'll be able to deduce more with add'l data. Also consider doing the higher Multiplier tests at 100Mhz FSB to prevent not posting/booting because of too high an oc'g for the 1600.

Also see if you have learned enough about Multipier Code to do your own "fault decoding".
John C.

 

[FPSguy]

Okay. I learned something interesting last night. First, while the BIOS showed a 9x option when I first switched the processor, that is no longer available -- so, the BIOS says 12x is the only option. Second, even though the BIOS thinks it is operating at 133MHz, 12x multiplier (i.e., 1600MHz, Xp 1900+ speed), Windows and 3dMark say the processor is operating at 1478 (which is not even a real speed as best I can tell, though it is pretty close to 133 x 11). I am not sure whether Windows measures the speed accurately, but if it does then this might suggest that all multiplier bridges are grounded (maybe the superglue didn't fill the pits?). Any thoughts? Maybe I should go back and redo it with crayon this time so I can see better whether the pits are filled. Also, if I have conductive paint in the pits, does anyone have any tips for gettng it out?

As always, thanks for any help you can provide!

 

[dmcowen674]

Hi FPS,

I expect to get my AGAIO 1600 chip today via Fedex. It has been bouncing around the Fedex system since Friday. Anyway download the Sisoft Sandra, it seems to be best in detailing exactly how a system is running. I like it best compared to those other programs that are just for pushing Benchmark figures.

 

[dmcowen674]

Hey FSB,

The chip arrived about 5 minutes after my last post and I'm on the new 1600+ now. I have not done anything to the Multiplier Bridges yet (Have to get a kit). I am mainly shooting for higher FSB than Multiplier anyway since SETI processes faster with faster Memory than higher CPU Horsepower. Mine came right up at 1.53 GHZ at 292 FSB solid. If I try to notch it up one more click I get Windows Protection error. I am only running 2 Generic PC2100 128M DDR sticks. Not shabby for a $50 upgrade. I suspect I could probably hit somewhere between 300-320 FSB if I get some PC2700 sticks but I will do that when I get a KT333 chipset designed for DDR at 333 FSB. They are currently testing and certifying the MB's and chipsets for 400 FSB 400DDR chips.

 

[johncar]

Originally posted by: FPSguy
Okay. I learned something interesting last night. First, while the BIOS showed a 9x option when I first switched the processor, that is no longer available -- so, the BIOS says 12x is the only option. Second, even though the BIOS thinks it is operating at 133MHz, 12x multiplier (i.e., 1600MHz, Xp 1900+ speed), Windows and 3dMark say the processor is operating at 1478 (which is not even a real speed as best I can tell, though it is pretty close to 133 x 11). I am not sure whether Windows measures the speed accurately, but if it does then this might suggest that all multiplier bridges are grounded (maybe the superglue didn't fill the pits?). Any thoughts? Maybe I should go back and redo it with crayon this time so I can see better whether the pits are filled. Also, if I have conductive paint in the pits, does anyone have any tips for gettng it out?
As always, thanks for any help you can provide!

Very strange that the mobo/bios does not give a compete range of multiplier options....but we can't help there, maybe others can..or research that issue elsewhere.

On the 11X issue...yes, sure "looks like" your analysis is correct re that all Bit Values are grounded/reset LO, ( glad to see you understand the code now). Might try "just cleaning"/opening the L1s first, then testing for default 10.5X/1400Mhz/XP1600. Here's why...10.5X has the .5X, 1X, 2X, and 4X Bit values all set HI by default. So if you clean out the L1s successfully and get 10.5X, then you'll know that there are "no L1 grounds"/LOs...'cause if there are any grounds you'll not get 10.5X...get something else ( that you could decode/isolate to specfic L1 and fix).

As to how to get the conductive material out of the pit...we'd use a 5X eye loupe at least to start with, then maybe a needle to pick out the superglue, or acetone to dissolve it... but lightly with a Q-tip to avoid possible damge to organic pcb. And in the final analysis, you don't have to "get it all out"...just "disconnected" from both ends of the bridges. That's where the loupe and needle might help the most...then thee's always the default 10.5X to prove your mod repair.
Good Luck, John C.

 

[FPSguy]

Well, I tried to unlock it again this weekend and this time I quite clearly have an 11x multiplier and 1466 speed. Grounded again, I guess. I think I found part of my problem the first time, though. According to http://www.asus.com/mb/socketa/a7a266/faq.htm:

"Question :

Can A7A266 support AMD 1.4GHz CPU?

Answer :

Yes, A7A266 is able to support AMD 1.4GHz CPU (in locked frequency multiple)."

The bios has since been updated to handle even faster processors, but maybe it still needs a locked multiple above 12x (which is all the dip switches can handle, though you supposedly can adjust multiplier in the bios, though I have had not luck with it). I think I am going to try one more time, and this time try to connect the 2x and 8x bridges (for 13x, fixed). If that works, I'll take it. If it doesn't, maybe I'll reground everything and settle for 11x times whatever meager bus speed the board will do (and it IS meager).

Thanks for your help! I'll let you know what happens.

-FPSguy