Manually Setting Duron(morgan) multiplier with connecting/breaking bridges

[MarkB]

I'm a cheap college student so I jumped at the ECS K7S5A/ 1.2 Duron for $109 deal at Fry's since i want to ease into DDR ie what til it's reasonably priced again...

I know this board is a pos for oc'ing but goddamn was it cheap.

It's running nice and stable at 1320 (110x12) via cpufsb.

The problem is it gets pissy when i set the ram asynchronously so i can't run my ram at anything more than the cpu fsb.

So I want to set it at 1333 (133 x 10).. figure it should hit that if i can do 1320 at default Voltage

I've been looking around and saw on ocinside.de that i could cut and connect some bridges to hard set the multiplier since this board doesn't offer mult change in the bios or dip switch.

I'm just wondering if there's a difference in the bridge config between the spitfire and morgan cores since its still on the ceramic packaging.

basically i'm asking if the layout is right for the morgan core on this page for durons:

http://www.ocinside.de/index_e.html..._resistors.html

Oh and I've been reading about burning the bridges with the 5V line from a comp PSU... safe?

I've got an old AT supply lying around i could use... if i can muster the balls to try it.. anyone out there done this?

 

[MarkB]

.bump.

 

[johncar]

Mark,
Morgan multiplier bridges layout is different than Duron. The L3/L4 "bridge pairs", which act as 3 pin/jumper sets, had the individual bridges of each "pair" side by side. The Morgan/Palomino has "physically offset" these bridge pairs, so you can't use the Duron layout/s on Morgan.

http://www.beachlink.com/candjac/index.htm has links to articles which diagram "both" Multiplier bridges layouts...Final Decoding for Palomino/Morgan, and Oc'g Duron/Tbird for Duron.

Have a good read, lots of marked up pics and circuit diagrams will enable you to "understand" exactly what's going on.
John C.

 

[MarkB]

thanks a lot


Anyone try to the burning of the bridges yet?

I would think it wouldn't harm the chip as long as the ps being used has a decent ground so that the current won't flow into the chip... ah hell i might as well try it once these mid terms are over.

 

[SonicTron]

Try thishttp://www.oc-athlonxp.com/bridges/PaloDecode0.htm

 

[MarkB]

Alrighty.. finally had time to sit down and do this...

results... right now i'm sitting at 133x10

138x10 has some problems... get ie errors and can't run prime95

I first lowered the multiplier to 8x and tried to see how high i could get the bus to go

Highest I got was 150 weird thing was I was only able to do this once... got into windows ran 3dmark2k1se.. shut down changed some stuff... forgot what.. i think the voltage and multiplier to go a little higher since it was at 8x150 = 1200 (stock speed of chip).

After that i couldn't get into windows no matter what i set the multiplier too.. i have a sneaking suspicion it was the voltage that was f'ing me up.

I guess when i burnt the bridge i really fried a couple of the volage ones because i had the hardes time getting them to register as closed.. so for awhile i was getting votages like 1.405 and sh!t like that which i didn't notice for awhile.

So i think when i have some more time I'll try to hit 150 again.

Here's some 3dmark2k1se scores at various speeds i hit.. oh and i'm still running pc133 ram so my scores are probably a little low

1320 (110 x 12) - 5798
1062 (133 x 8 ) - 6087
1200 (150 x 8 ) - 6635
1100 (138 x 8 ) - 6169
1333 (133 x 10)- 6382
1400 (133 x 10.5)-6436
1374(138 x 10) - 6531

and another thing.. when i had the voltage set at 1.85V.. the chip would over heat like a mf'er but when i set it too default which is supposed to be 1.75 but ends up being 1.81V its fine.. i remember reading about ECS K7S5A's heating up chips when they run short on power.. which is weird but i guess thats whats happening.. since I'm still on a 3.5 yr old 250 W supply.

So i think I'll try 150 x 8.5 or 9 next...