omg i unlockeed tbird and almost lost it please re

G

Guest

Guest
fist off this is a long post and i really need your help so please bare with me here. im still shaking.

my system specs quickly :
abit kt7a-raid/t-bird 800 @ 880 110 x 8 ( trying to get 150 x 5.5 or 6
256 mbs of mushkin h.p. cas 2 rev 3 memory @ 150 mhz( not yet though thats what its garaunteed to run at)
herc prophet 2gts ultra/creative labs sblive etc...

well after trying everything else to get my fsb up to 150 or higher now that i have the mushkin memory cas 2 rev 3 @ 150 mhz i tried to unlock the multiplier.
yes weve all heard of this being done , sounds simple, actually it was ,the hardest thing was replacing the heat sink.
but after using epoxy paste from the overclocking store and connecting the l1 and l3 bridges the system not only wouldnt boot it shut down almost immediately.
i tried it twice then pulled the chip out.

all ive read says that if you screw this up thats it for the chip. i cleaned it and cleaned it again then replaced it and tada, still nothing.
now im pissed, i know i didnt cross those lines or hit any other bridges, and besides now its cleaned so what gives?

i cleaned the chip again and replaced the old thermal paste and this time it boots and runs .!
so did i have it right and the lack of thermal paste wouldnt allow it to boot due to temp?
wouldnt it have fried the chip if my epoxy trick had been done incorrectly?
i swore that if it would reboot id never do it again but now that its on im thinking.... what should i do?
the fsb is stuck at 110 as the multiplier of 8 is what the system boots to no matter what i set it to to get 880 at best.
im not interested in how far beyond 800 i can go , heck i dont even care if i run over 800 at all as long as my fsb is 120 or 150 or higher. the board , cpu , and memory i have all support the higher bus settings so optimally i would set it to 5.5 or 6 x 150 for 825 to 900

please help?? should i try again to unlock the multiplier?

on the chip itself there are 4 rows of connectors or whatever labeled l1 , l2 , l3 , l4 respectively.
as i understand it i am supposed to connect the l1 and l3 bridges together , each one horizontal to the next making sure not to connect anything other than the two connectors.

can anyone verify this for me? i have the picture of the chip showing the location of these but its not very specific on wich to connect. im sure its as i asked it is above as there really isnt any other logical way to do this.
and is this the only way i can get another # besides the 8 for a multiplier?
i know this has been covered over and over before but when the chip is out and i have know way to get here for any more questions it really hurts.
ive built the system from scratch and it runs awesome but i built this specificaly to achieve a min of 150 mhz fsb or higher and i wont be happy till i do .
thanx 4 reading yet another long post by lumper i really hope to get this right.
 
G

Guest

Guest
Hi,

You only need to do the L1 bridges. Maybe doing the others will affect it like you say. I dont know. But to unlock, just to L1.

Tom...
 

Bombastik

Distinguished
Feb 25, 2001
132
0
18,680
Look, i don't have my TB1Ghz, but form what i've seen and heard from the net and friends unlocking the multiplier should be easy enough...

Al you have to do is connect all the L1 bridges (get a confirmation on this though as i'm not sure, and i don't want blow your TB800, look in this site [Tom's Hardware Guide] , there should be a article about unlocking the multiplier on a Duron [which is the same in the TBird]).

To do this you can use either the best way or the easy and painless way.

The best way is to use a conductive silver pen, to connect all the bridges.
The easy way is to use a good all 0.5 HB Pencil leed to connect the bridges (as the leed is made of grafite and grafite is a weak electric conductor).

I don't know about the other bridges, only the L1.!
 

Orange

Distinguished
Mar 4, 2001
21
0
18,510
I also tried to unlock my processor, just the l1 bridges and it shuts down after 10 secs.
Tbird 750,abit kt7a.
 

Bombastik

Distinguished
Feb 25, 2001
132
0
18,680
Does it shut down also after 10 secs even at 7.5x100Mhz?!??!

If it does, there-se a slight possibility that there is somekind between the L1 bridges or the other bridges (electricity going thru the HeatSink)
 
G

Guest

Guest
mine reacted the same way the first three times, then i let the epoxy cure for 2 hours before i tried it again and it booted fine. i used arctic silver epoxy kit from crazypc.com

i cant seem to get the multiplier ( the smaller number, i get confused) below 7
so i have succesfuly booted at many combos all with the 7 or 7.5 x 124/133/140.
i want to get to 150 but i need to lower the 7 to a 6 and it just wont take it , it hangs then i force it to boot to its last safe boot clock speed and try again.

ive tried alot of combinatios and benchmarked each one with 3d mark / sandra 2001 standard / and q3/ ut
all results were improvements over all, and im happy with it but , theres always a but, i want 150 fsb min and then try for 155 or higher, but i must first get that 7 to a lower setting.
any suggestions??
thanx again for all the help thus far!
with your help im almost where i wanted this system to run at months before i even built it.
insane gamer....................lumper
 

Bombastik

Distinguished
Feb 25, 2001
132
0
18,680
I still have my old P2-333+Asus P2B, so it's gonna only in theory.

First of all, i really REALLY don't know how the VIA KT133A (Abit KT-7RAID chipset) handles FSB's above 133Mhz, from what i've saw, if you have a PC150 DIMM (that's a DIMM for 150Mhz Bus) the KT133A SHOULD handle it, but there aren't any garanties since, for example, the KT133 (the one before) was for 100Mhz FSB and it didn't handle 133Mhz very well!

Now, if for 150Mhz it's shaky for the KT133A, 155Mhz is even more shaky AND what's worse, PC150 DIMMs are for 150Mhz FSBs, not 155Mhz FSBs, so don't get you're hopes up one that one.

Ok, back to the picture,.... as i recall it you have a TB800 and you've (at last! :) ) unlocked the multiplier... so the multiplier can be decreased since TB800 is 8x100 and you're talking 7x140...

I'm assuming you have tried for example, 6x100Mhz (dunno if it can be done, but i think the AMD TBird should underclock OK, altough for a pointless reason) and it didn't boot?

I'm trying to find out if you can lower the multiplier or not, 'cause if you're trying 6x150 rightaway without tring 6x100 you can't tell if it's because the system can't do 6x or the system can't do 150Mhz FSB well....

Oh yeah, if you unlock it using silver epoxy, you really should wait a bit
 

ejsmith2

Distinguished
Feb 9, 2001
3,228
0
20,780
Sounds like the pencil lead you are using might not be all that conductive. You want the softest lead you can find for the mechanical pencil, and the lowest number for the 'old' style pencils. The HB lead may/may not work all the time.

You might try using a magnifying lens to connect the L1 bridges together. Make sure you are connecting them horizontally, and not vertically. The mag lens will help you check.
 
G

Guest

Guest
My Abit 7t7A-raid is now running stable at 154Mhz 128MB Mushkin rev. 3. So the board WILL handle it no prob.

Some settings of the multiplier work better than others. For example... 9.5x 133/33 +12 FSB freezes my comp shartly after booting windows. (FASTER MORE STABLE? I HAVEN"T A CLUE, but it's true) At 10x, same settings, runs perfectly stable. You need to find the sweet spot, man.

I can also change my multiplier to 12.5x 100/33 + 11 (host+PCI)which is a 116FSB and 154Mhz Ram clock- at 1450Mhz. From another approach- 10x 133/33 +12FSB(host clock) is also 1450 but at 147Mhz RAM. So which is better?

I want to hear everybody's guesses.



Screw the risks!
-Street-AMD 1450Mhz-Abit KT7A-Raid-256CL2-Fop38
 
G

Guest

Guest
ok lets see,
i have tried to underclock it not only with the 6 multiplier but also with 5 ,5.5 , 6, 6.5, 7, 7.5.
x 100 , x 110 , 112 , x 124 , x 133 etc...

it wont underclock at any multiplier.

so because of my previous attempts at unlocking i believe that its hard to isolate wether this is an AMD spec or if i shorted out other bridges.

also when using the epoy i hadnt thought of this before , perhaps the epoxy should only sit on the connector itself and none of it should touch the cast or die or whatever its called ( the chip itself)

i believe a little is on the cast itself.

im thinking about cleaning it off and doing it once more carefully this time, but i wonder if the damage is done, and this may be in vain and another risk of breaking something. ( took 4 tries to get it before so whho knows what i might have damaged)

perhaps this will ensure me that the lower multipliers are not available.

but since even after doing this again if the lower numbers , ie: 6.5 , 6 , 5.5, and 5 are not available i will still not be sure if its cause of the way i did the unlocking or if there simply not available.

i have it stable now at 980 133/33 plus hst clk 7 x 7 for a 140 / 33 x 7.
benchmarking it and monitoring the temp has been very interesting and has confirmed all that i had prevously assumed about the performance factors of the fsb.

ist off my board thru sandra shows the double pumped fsb at 280 mhz ( 140 x 2)

here are a few interseting facts about the relationship between the fsb and system throughput.

clocked @7 x 124/34 = 868 showed cpu speed at 618 ( the numbers are meaningless here just an example of how cpu speed changed with the fsb)

at 125 x 7 cpu speed = 583
at 133/33 x 7 cpu speed = 617
at 145/ 41 x 7 cpu speed= not stable enough to test due to heat (yet)

at 133/33 plus 7 x 7 = cpu speed 660

and it goes on but the 133/33 plus 7 for a 140 fsb d.pumped to 280 mhz fsb is by far the most stable and gives the most throughput, not only with the speed test but with 3dmark,sandra,q3,unreal, superbike 2000,tripleplay 2001, babco sysmark, etc...

this final setting never crashed although it did run a little hotter than normal ( 55 degrees celcius/122 degrees fare..) up from 34 degrees at 110 x 8
so i think ill keep this one at the 140 x 7 via the 133 plus 7 method and try the 1200 mhz chip im getting in two weeks.

btw, how can i change the pci , agp and isa mulitipliers?

i would like to go from 1/2 to 1/3 or 1/3 to 1/4 for greater stability at the higher fsb settings.

with that one ill be able to be more carefull and hopefully be able to determine wether or not the new chips can unlock fully or not.

any advice on the model, type , or such that i should look for specifically for the best overclockable chip?

i mean its gonna be a t-bird 1200, but made from where?

wich metals ?
ive heard of a few types of the same chip being available.

ive heard of some with brass pins vs aluminum pins,

something about copper pins being available and the best ?
what do you think ?

i cant wait fto get one of these at 150 or higher, prob 133/33 plus 17 or 22 x 6 !

<P ID="edit"><FONT SIZE=-1><EM>Edited by lumper123 on 03/06/01 01:43 PM.</EM></FONT></P>
 
do you have a network card scsi card or any other pci or isa card installed?when you overclock you overclock everything not just the processor,according to your post at 145 fsb your pci is at 41 instead of 33 so maybe one of your other cards can't handle the overclock.
 

ejsmith2

Distinguished
Feb 9, 2001
3,228
0
20,780
Sounds like you're getting some good experience with all this. I'd give the pencil lead a try first, and if that cures your problem you only have to get a little isopropyl and a cotton washcloth to clean it up. I remember reading where someone had shorted all the l1 connections together. After a little alcohol cleaning and careful pencil marking, the cpu would overclock just fine.

If you're not getting results, something else is amiss. Possibly even a bad cpu/motherboard pcb.
 

kal326

Distinguished
Dec 31, 2007
1,229
108
20,120
Call me curious, but why are you trying to get such a high fsb. The Althons for all i have heard do not like any fsb other then 100 or 133. The idea when over clocking a Tbird is to change the multiplier to a higher number. Like right now you have a Tbird 800(8*100) you could just use a multiplier of 10 at a 100 fsb and have a 1 gig machine. or you could use a 7.5 multiplier at 133 fsb and have a gig machine. I would avoid putting a tbird up to 150 fsb. I dont think i would take it. The new tbirds now only have a 133 fsb. My p3 733 will only run a 145 fsb and its intended to run at 133 fsb. Anything other then a 66, 100, or 133 fsb screws with the PCI, Mem and AGP speeds and this can lead to instability. I would just raise the multiplier and keep the 100 fsb or just try a 133 fsb and see what ya get.

All Your Base Are Belong To Us!!