How to overclock a Duron 800

[nathan_m_2005]

I am currently running a AMD Duron 800mhz cpu on a Future Power KT133A board. I can only overclock it to 896mhz through the FSB thing or something in the BIOS. Do I have to change the core or anything to get it up higher than that? Or do I need to get some software to overclock it higher? I want to do this because some people got their 800's up to 957 and even higher. My temps are still good. Its running at 37 C as I'm typing this.

[Disk11]

Probably, you can just raise the fsb, which is the front side bus, or the speed at which the cpu interacts with the rest of the system. On most boards, but I am not sure about yours, raising the fsb also raises the pci/apg speeds, and if too high can cause system instability. You really need to, if you haven't already done so, but the system under stress for a while to see if it is stable or not by playing an intensive game, 3dmark test, or something similar. If your computer locks up or gives you blue screens of death, lower fsb speed.

To multiplier overclock, you need to unlock the processor. I don't know how to do this, but this is easy to find.

Make sure you have good system cooling and happy trails:t

[Bovon]

You really need to unlock the Duron in order to have the multipliers available as well as fsb. Does your board support 133 fsb?..or just 100 fsb. I don't know anything about the Future Power KT133A, and am not gonna go look it up.

http://www6.tomshardware.com/cpu/00q3/000711/index.html

http://www.pusateri.org/cruft/images/thunder.html
Shows using a lead pencil to unlock the L1 bridges. Use a conductive ink instead, the graphite from a pencil lead is not permanant.

http://www.sharkyextreme.com/hardware/articles/amd_oc_10-00/

[Terminator]

I lead pencil unlocked my old Duron 600 over 1 year ago and still runs fine. After unlocking the L1 bridges I stuck some cellotape over the bridges so the lead didn't vapourise after time.
It hit 1050 at 10.5 x 100Mhz but I backed it off to 950 MHz. You'll need to bump up the core voltage. Depending on how high you want to go you might have to set it to 1.85V but make sure you have good cooling.Although mine went 350MHz past it's standard that does not mean you will get this because the Duron 600 was probably the best Duron overclocker but bet someone knows better. So don't expect to reach 800+350 = 1150MHz but you might manage it. Best bet if you really want to clock it high is a mix of multiplier and FSB clocking but I don't clock my FSB because it stresses the PCI/AGP/USB/etc but it's up to you.....if you do don't go too high unless that motherboard you have has 1/4 or 1/5 dividers.
General rule I use is if I overclock is take it as far as I can until you get lock ups or strange things happening then back off by about 10-15%.
Happy clocking.

T

:t

[nathan_m_2005]

Yes my board supports 133. But right now the jumper is currently set at 100. That was the default and I never changed it. All my ram on there is 133 though. I suppose I could change it. Do you know where I can get 3dmark?

[BlackTruckRyder]

www.futuremark.com

:t

[Bovon]

You can't change the jumper or bios...whichever...until you unlock the Duron. It comes set to run at 100 mhz fsb, and will not run at 133 mhz fsb until it has been unlocked. The multipliers will not work either until it has been unlocked..or at least thats how most of the boards are setup to work.

[causticVapor]

For the record, the Duron 800 is one of the best overclockers out there.:t

[Terminator]

hmmmmm.. looks like I stand corrected then? When I said that the 600Duron was probably the best Duron overclocker I meant that it clocked MUCH higher than 600MHz. At 1050MHz it was running 75% above standard. Does that mean that he can expect the 800Duron to hit 1400MHz?????
I only reduced my clocking to 950MHz because I couldn't stand the noise my thermaltake heatsink and 7000+ RPM Delta fan made not because it did not run stable.

Checking out the survey on here :-

http://www.sysopt.com/overc/amdcpustats.html

seems to indicate that the non-Morgan core Duron's don't go much above 1GHz but it all depends on the chip you get.

T

:t

[Bovon]

In 2000 for the most part, there was a great influx into clocking.

The new socket A series processors offered so much more than their predecessors, the K6 II and K6 III cpus.

Just like the K6 series, the socket A processors had good clockers, and some not so good. There was also some that were fantastic clockers.

What many people never heard or learned was, there were/are two AMD fabricating plants...one in Austin Texas which fabs (or did fab) from aluminum only. The Dresden Germany plant fabbed with all copper. The Dresden plant made only Thunderbirds, and the Austin plant made only Durons.

When people really got into the hobby of overclocking...the Durons was the cpu of choice because of the cost primarially...this caused a shortage of Durons on the shelves, and the Dresden plant fabbed a bunch of Durons over there, and out of copper. These Durons were spectacular in overclocking because they didn't heat up like the aluminum fabbed ones did, and for some unknown reason, every cpu that came from Dresden was a better clocker than those from Austin.

Back then, the stepping code was very important...and as soon as a new Tbird or Duron came out, the results of that series was posted everywhere. Even the core color was important because Dresden only used one color (blue) polishing dye and Austin used several...just about anything with the Dresden "blue" was greately in demand.

Today, there is no info out (that I am aware of) that mentions what Dresden or Austin fabs with, and the polishing dye detection method is not an indicator of where a processor is made. I suspect that the Dresden Germany processors still excell above the Austin counterparts...but figuring out which is which is much more difficult today...and AMD ain't saying!!

[paul0660]

I remember getting my 600 Duron up to 900+, mostly on multiplier cuz more than 110 mhz didnt work, dunno if it was the peripherals, the mb, or the chip that limited it. It was a long time ago, and a lot of fun! Dont break anything while you are fooling around or you will wish you spent 30 bucks on a new 1.1 gig duron in a hurry!

[Bovon]

Originally posted by paul0660
Dont break anything while you are fooling around or you will wish you spent 30 bucks on a new 1.1 gig duron in a hurry! LOL...boy!! you got that right. But, really thats a lesson learned the hard way that sticks with you. You screw up one time, and crush a core, or by some other means...kill that bird...I think the average guy-gal will have learned far more than he/she will ever learn from reading here or at some other forum.

I read lotsa posts where some guy with crocodile tears in his eyes wrote to say he had killed his"beloved" Dresden blue Tbird.

Whats that they use'ta say?..."the school of hard knocks"?

[causticVapor]

Hey, same here. I got a perfect AGOGA XP 1800+ and it only lasted 8 months. When I put new thermal compound on, the thermistor fell out of alignment while I put the block on and caused it to come in uneven contact with the core. When I screwed it down, CRRRUNCH! $120 CPU down the drain. :eek: :(

Fortunately now, I have an XP 1800+ T-bred AUIGA that I got for 83 euros downtown. Mr. 8-KHA+, after a few close calls, is still going strong. :)

[Bovon]

Originally posted by causticVapor
When I screwed it down, CRRRUNCH! $120 CPU down the drain. I'll bet you would never believe that was why shims were invented in the first place.

[causticVapor]

Only bad thing... A) the shims didn't fit around the waterblock (large and an extruded base) and B) all the ones I had were made out of copper. They also increased the temps... and I never had any problems before.

[nathan_m_2005]

I tried unlocking the processor with a pencil, and it didn't work. I still can't go to the 133 on the bus or higher than 896. How hard do I have to push down with the pencil?

[causticVapor]

You need to make the pencil traces stronger; it should work then. Cover it with tape afterwards. :t

[Terminator]

What you need to do is get a 0.5 propelling pencil and a piece of paper. Scribble on the paper at an angle til you get a sharp edge on the lead. Then use this to trace in one of the the L1 bridges. Go over it a few times and tilt the CPU at an angle in the light and you'll see the shiny lead line. Make sure NOT to cross over to other bridges. Repeat for the the other L1 bridges then tape it.

T

:t

[nathan_m_2005]

I tried it with the pencil lead you told me to use, and it still didn't work. I think I may have gone over too far on one of the bridges too. I think it might be touching another L1 bridge. What should I do?

[Bovon]

If you have shorted out the adjacent bridges, obviously you need to remove the graphite. Try a pencil eraser and some alchohol afterwards on a lent free cloth...you will probably also need a good strong magnifying glass to view the area. Clean it good and look at it for any left over graphite...then start over.

Personally, I do not like the graphite from a pencil method...lotsa guys have had good results with the method and lotsa guys have had problems. I used the conductive ink method and never looked back.

[nathan_m_2005]

What kind of pen has conductive ink though? Can I use about any pen for this?

[Bovon]

No, conductive ink pens are specialized, and were designed for circuit engineers to make temporary circuit traces when designing some circuits. It will probably not be found locally...I am told that Radio Shack has them from their online store, but not at the local ones. They are not cheap..costing around $10 bucks or maybe more. There is another simular substance that can be found at any well stocked automotive supply store...and that is a rear window defogger repair kit. This is a liquid that is conductive that is used to repair the circuits on those autos that have a heated rear window defogger.

I also recently read where someone used liquid solder bought at a hardware store..I donno about this stuff..if it is truly conductive ... like solder would be...and it will harden...then it would work ok. As soon as I can get to some store that has this stuff..I will buy a tube, draw a few lines with it...let it dry and then test it with an ohm meter to see if it actually conducts.

To do any unlocking of any processor, we generally have a strong magnifying glass, and a good light.. preferably one that can be mounted over the work and allow a good view while trying to close each bridge without bleeding over to the next one. I used a hypodermic needle...very tiny one for diabetics..placed a drop of the ink out on the edge of the ceramic base and dipped the needle in that. Even as small as this needle was, it held too much ink, and it would flow out over the next bridge. It was necessary to clean the work off and start again several times to get it just right.

Be patient and work slow...you will get there much quicker than if you have to wind up redoing your work many times.

[ichorid]

if there is no option in ur bios to change multiplier then the pencil trick wont give it - all linking the bridges does is enables it ( if it is there but grayed out)

[Bovon]

Originally posted by ichorid
if there is no option in ur bios to change multiplier then the pencil trick wont give it - all linking the bridges does is enables it ( if it is there but grayed out) Thats true, but you also cannot run a base 100 mhz cpu at 133 fsb without unlocking it. Most (if not all) of the older 100 mhz processors performed better bench marks when unlocked and run at 133 fsb.

[nathan_m_2005]

My board doesn't have the option to change the multiplier. It only has the option to change the voltage and fsb. Is there anything I can do?