running Tualatin on CuMine MB w/o Powerleap - Overclockers Forums

1 → AN3 → AJ3 → AK4 open with mainboard

2 → AK4 and AK26 short 3 → wire trick on Vcc (jumper setting Vcore to 1.5V even better) http://www1.oc.com.tw/forums/msg.asp...=74399&posit=5

if this is feaible Powerleap is making huge margin and Intel is also making immoral profits

Soil, I've been wondering about the same thing. I have a Soyo I-815EB motherboard(not a B step) and a 1.0A. I've been tempted to stick it in there just to see what happens, but I've also been a little afraid of doing an instant fry of the chip. I wasn't able to read the article on this system, but did they have to do any mods on that board to make it work? I'd love to get this to work here and get a little more milage out of that Soyo board, since it's less than a year old. I had planned on ordering a ST6 next week to replace it, but this surely would be a money saving solution.

Rick

this is another forum working on that, got a few images and faster / easier to access:

http://forum.hkgolden.com/view.asp?message=159158

the diagram is the pin position looking at the bottom (pin slug) of the CPU, NOT on the socket. the 2 corners w/o corner pins should then be on your left hand side it says that the guy can get a C1A be identified as a PIII but the Vcore is 2.2V(!) (for a Tualatin) and the CPU was not burn... it was not about modifying the MB, but the CPU, they tried not to permanently cut off AN3 → AJ3 → AK4 but having difficulties insulating such pins 1 → AN3 → AJ3 → AK4 open with mainboard 2 → AK4 and AK26 short it would be safer for CuMine MB using manual jumper setting the Vcore, if the CPU is to be permanently modified that would mean chopping AN3 and AJ3 off, use a fine wire to connect AK4 and AK26, chop off the excess part of AK4 and use a dot of non-conductive expoxy to seal the AK4 connection

they have done it, 10 hour stress test passed

http://forum.hkgolden.com/view.asp?message=159158

but only on win98

Well this is good news.

I am wondering if you mod chips like this will the be able to work on the fcpga boards also?

neat if that could be made easy on fcpga I can't read chinese though, I'm bad in electronics and I've a slot1 mobo...

it would be interesting to know what kind of voltage regulation they are using though..

Replying to doer.

Probably none. The extra voltage to the cache might make the chip more stable but it may also be more unstable also. Seems they are hitting 1.55GHz with the 1.0a but it runs too hot or something. i can't read chinese and babelfish sucks for translations.

I think that's their final approach: (have fun)

1.an3 → aj3 → ak4 open with mainboard using insulation lacquer (if you don't want to cut it). 2.vid25mv(ak36) connect to vss; get 1.45V Vcore. 3.upon boot up in BIOS increase Vcore to around 1.5 V translation: 1.76V initial attempt,@1.55g ok,stable. but a bit hot...41-51c. finally:1.65v...@1.45g,temp:33c-43c passed all stress tests(3dmark2001,super pi,cpumark99,hot cpu test,software sandra 2002) very stable.run for 10 hrs. everything fine!

Ok, but can someone PLEASE Explain this step by step, for the average idiot out there? (I am one of them!).

The pictures on the HK page were clear, but they don't make sense...I don't know what to do or why those pins are circled (except they have to be modified or wired somehow??) I don't understand ANY of these terms, and as I said in another thread, I have no electronics experience. So can someone please explain VERY carefully exactly what needs to be done and what parts I need? Please remember I am a complete idiot when it comes to this ! I tried to read this thread but all it is doing is giving me a headache And has anyone here actually tested this? if you have, please explan step by step, very carefully....some of us here have no clue (I'm one of those guys who will go "huh" when asked to unlock an Athlon XP, for example...but i'm willing to learn IF the instructions are clear enough And...... " 1 → AN3 → AJ3 → AK4 open with mainboard 2 → AK4 and AK26 short " What is this ????? what the heck is " 1iban ......" ? I dont see it anywhere. and what is meant by "ibak4 and ak26 short "? @@ I know you said it better, with : "I think that's their final approach: (have fun) 1.an3 → aj3 → ak4 open with mainboard using insulation lacquer (if you don't want to cut it). 2.vid25mv(ak36) connect to vss; get 1.45V Vcore. 3.upon boot up in BIOS increase Vcore to around 1.5 V " but........ open with mainboard?? huh? what does that mean? is that something on the slotket? (I have a MSI 2.3 master, so I guess I can use that...). What is insulation lacquer? How do you connect vid25mv to vssia ?

Replying to Falkentyne.

if you can't understand these you'd better not risking your CPU and MB, what your need to do is just for (1) insulate these 3 pins from contacting the socket (electrically open) and (2) soldering a wire in the MB connecting 2 pins for those who like hacking hardware (doing more than using pencil) they'll know what to do I'm still happy with my CuMine so I'm just letting more people know and so the more daring can continue the experiment

the 2nd page of the thread (click the hyperlink of the upper or lower right hand side) now has some more photos so it may be easier to understand how the mod is done:

http://forum.hkgolden.com/view.asp?message=159158

Would this mod work on an Abit BX133?

Would putting silicon on the pins that should be open work?

Replying to mikyseee.

The author did not know whether the mod would work on an Abit BX133. For insulating the open pins, they believed the experimenters ultimately used thin plastic wire insulation fitted over the contacts.

I wonder if this could be applied to an abit slotket3. It has voltage jumpers that go low enough for Tualatin use.

I just ordered a 1.0a ($79 shipped, newegg) to try this out.

I'll nailpolish those 3 pins and start with 1.4v on my slotket (the jumpers go even lower than that). If I cook it, hey it's only $79. If it works (and I can get to at least 1.4Ghz with it), I've gotten 75-80% of the performance of an xp1700/NW 2.0a on the cheap. When the Geforce4 ti4200's come out, I'll be golden for a year.

good luck, check those 2 threads from time to time, and feel free to post there in English. Well supposingly you need to be proficient in using soldering iron and HW hacking in doing that...

and so far they have only been successful in win9x/Me, not 2K/XP, and even for 98se, a Windows re-installation is needed. BTW I just know that Powleap 370/T (cumine/tualatin) adaptor is available locally for USD35, so you may want to judge whether the risk is worth that amount... and apparently the chip in the Powerleap adaptor is a logic gate chip and has something to do with voltage selection. please keep us informed anyway and nice to know somebody is interested in that thread

I'll post there too, in english of course.

Powerleap still bundles the adaptor with a 1.2 for $170, or about the same as a 1.2 and a new ST-6.

Replying to Ruiner.

you mean the slot1/Tualatin adaptor or the 370/Tualatin adaptor?

1.2 Tualatin to Slot 1 http://www.powerleap.com/Products/iP3T.htm

I have a slot 1 mobo (see sig)

Ruiner, I'll be waiting for your news. If you can do it I'll want to try it also.

Replying to Ruiner.

75-80% of the NUMBER maybe, but I doubt 75-80% of the actual system performance. I could be wrong though.

godfodda,

http://www.xbitlabs.com/cpu/celeron1a-oc/

It is pretty surprising how well the Celeratin supports the Ti500 at real resolutions. Of course it gets clobbered at low res, but who plays at 6x4 anymore.

Replying to Ruiner.

I don't game, so that part of it makes no sense to me. But that was a very nice link! The bench numbers look much better than I had expected. Unless someone can point me to another site that shows differently, I may have to eat my words.

I tried this hack on my 1.1A tually last night. Instead of using nail polish or wire wrap insulation to isolate the pins AN3 (DYN_OE), AJ3 (Reset#2), AK4 (VttPWRGD), I just removed them with needle nose pliers (this 1.1A was not a good OC'er so I did not care too much if the mod did not work and it killed the chip). Then I put the CPU into an 815E board that does not support tuallys (AZZA 815EP). This board works fine with P!!! coppermines but I could never get it to post/boot with the tually celertins. Also had to use the wire trick on the socket to pull VID25mV low (wired it to VSS, defaults to 1.45V).

No luck getting the tually 1.1A to work in the AZZA board. No post/boot, think the problem is with the Bios. But get this: when I put the CPU back into my ST6 board the thing still works! Booted right and loaded Win2K just like it did before. I was not expecting that, thought it would be history after losing those 3 pins. I guess they are not too critical, at least not in the ST6. Actually I was kind of hoping it would be toast so I could make another keychain out of it and order a PIII-S 1.40. Guess I will hang onto this 1.1A a little longer and wait for the PIII-S 1.53 instead.

should I say congrat ?

MaxFSB, can you try that on a BX board?

Replying to MaxFSB.

You were supposed to short two other pins.

I just pick up a refurbished 1.0A celeron for $54 shipped from newegg. I'm gonna try this on my BX133 board when I get it.

Replying to ol' man.

Are you referring to the VID25mV-to-VSS wire trick that sets the default core voltage to 1.45 V?