View Full Version : 1.33 only likes 10 multiplier
No matter what I do my new AXIA 1.33 chip will only run stable at a 10 multiplier. All the L1's are unlocked. I've tried every FSB, voltage, driving strength, cas setting, and every other setting I have in my KT7A bios and the darned thing just will not run with anything other than X10. Anybody have any ideas?
asmodean
04-26-01, 01:56 AM
Did use a pencil for unlocking? I wouldn't trust the pencil trick farther than I move objects with my mind.
I had stability problems when using a pencil. Silver paint cleared the problem. Defogger kit seems to be quite popular also.
When you say it only likes 10, do you mean it wont run at say, 8, or 7? Or, do you mean it will not run greater than 10? If it will run 10x133 but not 8x133, you do need to look at the bridge job, although an L! bridge problem usually manifests as posting at a speed different than what you called for in the bios. If it posts and loads your O/S at a speed other than 10, but fails during benchmark or stress testing, than I doubt it is an L1 issue.
Hoot
The L1 bridges are factory unlocked, no pencil needed. and it doesnt like any multiplier, 5 -9.5, and 10.5 - 12.5. Only 10. It boots up ok but it will lock up a short time after it boots into windows.. and not after any certian program either. Just a random lockup 30 seconds to 5 minutes after it boots up.
portorock
04-26-01, 12:53 PM
weird, I have the same mobo and chip, I don't have that prob. did you try clearing the cmos, or changing the multiplier with the bios default, also, try an upgraded bios. don't know what else to say, but good luck
Yup, cleared the cmos and it already has the latest bios. I wonder if I got a bum chip or a bum Mobo.
Do you have a strong PS? I noticed you have great cooling but those things pull tons of amps. I am guess you already know this and you probably have at least a 350watt PS or I believe that may be your problem. I am at a total loss. Were you able to do better at say 9x155 maybe even 9x160?
Another thing maybe set the default voltage on the L6 or L7 to 1.85 seems to get some people stable again.
el (Apr 26, 2001 02:00 p.m.):
Do you have a strong PS? I noticed you have great cooling but those things pull tons of amps. I am guess you already know this and you probably have at least a 350watt PS or I believe that may be your problem. I am at a total loss. Were you able to do better at say 9x155 maybe even 9x160?
Another thing maybe set the default voltage on the L6 or L7 to 1.85 seems to get some people stable again.I was thinkin the same thing. I only have a 300 watt supply, I'm running the pelt on it's own supply and the pump is run off AC but I have 4 extra fans on there. Maybe I'll try a bigger supply before I throw it in the trash. :):):)
TT120 (Apr 26, 2001 07:44 a.m.):
The L1 bridges are factory unlocked, no pencil needed. and it doesnt like any multiplier, 5 -9.5, and 10.5 - 12.5. Only 10. It boots up ok but it will lock up a short time after it boots into windows.. and not after any certian program either. Just a random lockup 30 seconds to 5 minutes after it boots up.
This is very weird. If it boots at a certain multiplier it should retain that speed unless it overheats.
Now this is a hunch, but could it possibly be that your factory unlock traces are somehow broken? I know it's unlikely, but given all the heat stress and pressure that these chips are exposed to I could see that a trace that may be was not quite perfect to begin with got a microtear, and as the chip heats up during boot or while in the OS this tear opens just enough to disconnect the trace and crash the system.
Again this just speculation, but redoing the traces with a pencil would eliminate that option.
Yo
portorock
04-26-01, 05:52 PM
I don't think that it's a psu prob, because my sys has a 250w psu. I know that's crappy, but it runs
As long as I stay at 10, it runs fine, I can go up to 140 FSB which is as much as my generic pc133 ram can handle. I've tried putting the ram at cas 3 to see if I could get a higher FSB but that dont work either. 140 x 10 is where I'm stuck.
jetjaguar
04-26-01, 07:30 PM
hmmm sounds like u need a bigger psu ..
i running a enermax 431 watt and no probs here
the 1.33 use alot of power ..
i would try a bigger psu
Have you tried raising the I/O voltage to 3.5 or 3.6 ?
I have to do this on my 1.33 or it freezes up.
Well, I tried going from 3.4 where it's at to 3.5 but it didnt help at all and I'm a little gun shy about going any higher. I dont want to burn anything out yet. I have some Corsair pc 150 ram coming but it's still a week away. Maybe after I get it I will run that voltage up and see just how much this generic stick can take before it burns up.
TT120 (Apr 26, 2001 07:44 a.m.):
The L1 bridges are factory unlocked, no pencil needed. and it doesnt like any multiplier, 5 -9.5, and 10.5 - 12.5. Only 10. It boots up ok but it will lock up a short time after it boots into windows.. and not after any certian program either. Just a random lockup 30 seconds to 5 minutes after it boots up.
try a pencil over the closed L1's and tell us if it changes.
Uhhh, ok, whats the pencil trick?
LOL just kidding! :):):)
I'll give it a shot tonight when I get home from werk.
U know the one where you shake the pencil from the middle and it bends it while unlocking the multiplier!!!
I hope it works. Hearing your details I don't think it is the PS after all but at first I thought you were running everything off the one PS and I would be surprised if your light didn't dim when you turned on your machine!
TT120 (Apr 27, 2001 09:44 a.m.):
Uhhh, ok, whats the pencil trick?
LOL just kidding! :):):)
I'll give it a shot tonight when I get home from werk.
WOW - that was close... LOL
I dont know if you noticed my past problems but i had the same problem with my amd 1.2 axia and it came from the factory unlocked. I know some of the locals at computer shop i deal with so they did me a personal favor and swapped me for another cpu. I could than go to whatever my multiply setting heart desired running 1.33 now and guess what my first one and it was an axia wouldnt go over 10 on the multiplyer also my second one again does.
This is too wierd. My trusty 1.0G/266, after a major session with a peltier taking it from -4 up to 32 and back down, for several hours, has developed a similar problem. Quite suddenly, I can not choose multipliers of even numbers and the .5 multiplier stopped working.
It works okay at 7, 9 and 11, but selecting 12 gets 11, 10 gets 9, 8 gets 7, 6 fails post entirely, but then 5x never worked for me anyway. No .5 with either even or odd mults. Here's where it gets wierder. 12.5 works, but I think, based upon other reports, that is actually 13.
So, I could no longer use my favorite 10x145 2-2-2 setting. It ran 9x150 okay, but I don't like 150 on my system. The AGP GF2MX is right on the edge there and besides, it's slower.
Now your first thought would be the L1 bridges, but mine are soldered, securely. My 400W PSU is no where near its limit. I took the chip to work and gave it 30 minutes in the Branson degreaser. No help.
In a bold and as I look back, reckless state of dismay, I took a big unknown step. I cut and re-wired my L3, L4 and L6 traces for 1000 Mhz. Being a 266 rated CPU, it used to auto-detect as a 750. (750x133=1000). This involved some very touchy work with an engraver, under the stereo microscope and some conductive paint jumpers. Now, it auto-detects as a 1000 (10x100) and I can change the FSB to my favorite 145. I still can not select 12 or 8, nor the .5 mult jumps.
It works fine otherwise. Same benchmark results as before the incident. 1450 Mhz is not too shabby for a 1000 CPU. Eventually, I will replace it and make it a spare, but for the moment, it runs okay and more importantly, it's paid for.
Hoot
The funny thing is, that it will post and boot into Windows just fine with any other multiplier but it will lock up sometime after that. It's not program specific. I can just let it sit there and it will lock up a few seconds, or minutes later or I can go straight into quake 3 or sandra and it will lock up in a few seconds or minutes. Run it at 10 and it's fine.
I have the same problem. I've got the same board, same chip, except mine came unlocked. Anything over x10 and I get lockups... I gave up trying to figure out why.. So maybe someone here has an idea.
I have the same prob as Hoot does minus the solid numbers going one lower. I cannot use any .5 mult. It POSTs fine, gets into Windows, sits for a min, then, deep freeze. I can't even run 7.5X100, just wont work. I used a defogger kit and have checked the bridges many times. I would like to raise the fsb but I can't because the 8.5 mult won't work( 8.5X141 ) Oh well, maybe my next one will work better.
Fiz
Kurant, that's not the same problem as TT120 has. His won't work over or under 10x. Yours sounds like that is the limit of the chip. It's not uncommon to hit a multiplier, above which, your cpu won't work. Ditto on the FSB, though it is more often limited by other factors. Upping the core voltage can get you up a few more multiplier steps. Before my misfortune, My cpu would not run above 10x at the default 1.75 Vcore, but runs solid at 1.85 Vcore, all the way up to 12.5, but I have to lower the FSB to keep the total speed at 1450 or less, which also seems to be a "wall" for my cpu.
Fiz, my system does not malfunction with the .5 steps, it ignores my call for them altogether. IE I set bios for 7.5, I get 7, I set bios for 9.5, I get 9.
I'm gonna go out on a limb here and say that it was a bad move I made, soldering my L1 bridges. You do dumb things in ignorance. A fellow, who knows IC structure, told me that despite my care in not using too much heat, that the heat of soldering weakens the connection beneath the L1 bridges and it can separate. Given the mechanical stresses of going from hot to minus 0 while experimenting with the peltier, I think I proved his point. I will no longer encourage soldering L1 bridges to anyone.
Hoot
I'm sorry to hear that, ouch. Mine will accept the call for the .5 mults, it just dies when I get to windows. Oh well, i'll chalk it up to experience, I don't want to take the proc out unless I have to.
Fiz
mine accepts the call for ANY multiplier and will post and boot windows but then it just freezes up sometime later. I'm startin to wonder about my ram and vid card and power supply. I'm gettin a 256 meg stick of corsair pc150 cas 2 next week and will probably get an Enermax 431 watt power supply as well. Might as well eliminate all the variables. Vid card the week after.
A little update:
I removed the Pelt from my system (condensation problems) and now it seems that I can run at least an 11 multiplier. I'm still testing for stability but right now I'm at 133 x 11 at 1466 Mhz. Seems to be a temp thing near as I can figure. It doesnt seem to like the lower temps. Right now with my watercooler alone I'm at 42c with a case temp of 24c. Those temps seem a little high to me. What do you think?
I think it is a nice OC. good work :) I think your temps are fine but some people think anything over 45 causes problems. I just started working on my 750duron and hopefully I can get 46% out of it. only at 800 now but I am just loading stuff on it so I didn't want to push my luck. you running it at default vcore?
No, I have to jack it all the way up to 1.85 which actually comes out to 1.91. I didnt think 1466 was all that good out of an AXIA 1333 chip. I have a AXIA 1200 that will do 1400.
yeah I always think about the 200fsb instead of the 133 so i guess you only got a 10% OC :(. Did you turn down all you memory settings?
Yeah, but it still wont go over 140 and be stable. I have some Corsair pc150 ram coming in this week so I'm gonna ditch this stick for that one and hopefully I'll get it a bit higher.
UnseenMenace
05-01-01, 08:07 AM
I dont know if this helps but i have a similar problem kind of problem.. When overclocking my Duron i was suggested elsewhere on this board to change from 10x100 to a lower multiplier and higher fsb which i did and eventually ended up running at 7.5 x 136. (1020 mhz) cpu temp @ 37 c, system temp 25 c and stable. Oviously you cant have to much of a good thing and looking for more performance went back to the Bios looking for more (dam it why am i so greedy). When rebooting the system it did not POST or locked up when it did so i went back to the bios and found that the DRAM clock had changed from the HCLK+PCICLK needed for 133 mhz ram to HOST, which i believe is used for 100 mhz ram ( I have 133 mhz ram ) and this continues to be the case if I now use any FSB other than 100 mhz (however when using 100mhz fsb HCLK+PCICLK displays and stays in the bios) if i change the fsb now upon reboot the DRAM clock changes. I have loaded up the defaults and then re-adjusted the bios and still no change in the situation, I have also reset the CMOS (dunno why but im desperate) and still no changes.. anyone got any ideas.
Duron 800 @ 1000 mhz (10x100 again) <-- Ohh please come back 133 mhz fsb
Abit KT7A-Raid
512 mb (133 mhz Crucial Stuff)
Elsa Gforce with Blue Orb
Sound Blaster Live
Full Tower With 5 Fans (case fans on a seperate PSU)
Bible required for 133 mhz fsb type prayers
Host + PCI can only be envoked at FSB speeds at or below 120 Mhz.
Hoot
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.