getting AMD mobile to 2600++++
 

Hi all, need some help to tweak my setup

basic setup

Epox 8rda3+ stable until 221 fsb on 1.8V need to test it with 2.0V
Twinmos BH-5 running 221fsb on 2-2-2-5 with 2.9V can't test it higher due to mobo limit
AMD mobile 2600+
SP97 with enermax92
Enermax 465w power supply

the following happens:

it runs on daily base 2520mhz (12x210) on 1,725V with temp under load and low dba 45, nice to my oppinion

as max oc i can use 2616mhz (12*216) or (12,5*208) on 1,85V and load temp 48 with open case

the board is overvolting a little so i get like 1,87V.

any higher setup gives normal xp setup but instant reboot while i try to bench (super pi .... etc)
increasing to 1,95V doesn't help and same result, booting xp ok but instant reboot during bench

any ideas? first i tought my CH-5 was the problem because my max fsb was 216 with multiplier 10 but now with bh-5 it is 221 for sure
due to the mobo. but on multiplier 12,5 i can't increase to 2630 or higher and multiplier 13 and up does not work....

if i touch the sp97 it is warm but not extreme and reboot after few seconds of black screen gives me a temp below 40 so it is
cooling the cpu for sure

thx
duploxxx

shal I put it under my prommy ? ;-)

Try the L12 mod (google), u can gain a lot of FSB with it.

Reached the limit of the Chip.

My XP2600+ M; will reach a bit over 2.6 Ghz no more disregarding vcore increase.

you have a prommi now piotke?

wel Lazyman, it is possible that you have a point there, the limit for mine is indeed a bit higher then 2600 and then finished.

is it due to a produktion version? because i have seen 2600 mobiles doing 2700-2800+++

I dont think the L12 mod will help, the fsb limit is comming from my mobo and not the proc

thx for the help

duploxxx

I tought that the L12 just solver the problem with the mobo not getting high FSB'es

thx for the info, i'll check that

I think it was for proc who don't like high fsb

duploxxx

http://www.xtremesystems.org/forums/...threadid=15300

BTW, what is the highest FSB with a low MP, max VDD, max. memV (and maybe a mem div).

when the CPU reaches the point of diminishing return, it means there is much less output than input.

Example: raising 0.1 vcore to gain 5 mhz and 10C higher in temp.;-)

in theory it tricks the mobo into thinking it's a 200 mhz fsb proc instead of a 133 mhz fsb proc but it also helps in obtaining higher fsb's

thx Kristos

Did it help with your mobile to get a higher fsb?

duploxxx

actually, I havn't tried it yet untill just now:gay:

But since you asked I figured I might just as well do this now instead of postponing it again (I'm very good at that you know :king: )

I knew it worked though since a friend of mine has the same mobiles as I do (I ordered them through him, they're even from the same batch :)), and he gets about 22x fsb without l12 mod and about 24x with l12 mod.

He does it the old way btw, just connects the 3rd l12 bridge with a pencil, giving him 166 fsb setting by default. De doesn't even cover the laser cut with something none conductive, he just draws right through :D


I connected ah 28 with ah 30 as instructed on this page:
http://www.geocities.com/trats102002/nf7sFsbsense.html
this should of given me a default fsb of 200. It didn't work though.
I don't know why yet. I might have bad contact, I might not have the sort of mobile that is discribed on that page. Or the info may be wrong but that's something for tomorrow :)

thats a nice fsb. on what mobo? i am afraid my epox 8rda3+ won't cope with the 24+ fsb

but still nice to know that it could work, i'll check it out

duploxxx

he has always used abit nf7 (-s?) rev.2.0 in the past (killed 3 or so through various means of abuse :D)
The only other socket A boards capable of such speeds by my knowledge are the DFI lanparty/infinity rev.B's.

wel dit some test yesterday...

at vdimm 2.9 my bh-5 at 2-2-2-11 reaches de limit on 221fsb
i tought that was limit of mobo or proc

however i've relaxed the bh-5 a little to 2-2-3-11 and i get a nice
237fsb :) on multiplier 8.

using this on multiplier 11 with 230 reaches again the limit of the proc at 26++ but the pifast results are worse then 218*12

i think its the max for my proc.... upping Vcore to 1.9 doesn't help me

maybe different cooling might help...

duploxxx