• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Just made the vcore mod to my KT Ultra =]

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Michal[mt]

Member
Joined
Sep 5, 2001
Hi all.
First of all I want to thank Tiger who gave me a little hand in made this mod working.
I have applied the vcore mod using a 47 KOhm pot(trimmer) to the voltage regulator named as HIP6302.
It was not too easy in soldering the 2 wires on the chip...but finally I obteined a nice result.
Now I have the chance to reach a multipier factor of 10.5 and increase the FSB as I wish.
Using the pot. it is very easy to increase the vcore till to 2.20 volts and more ;)
 
Good stuff there and a nice oc. You are braver than me to solder the connection. Just be careful going over 2.05, that chip already produces too much heat and requires serious cooling solutions. Sadly I lost my 1.33 AXIA Y to condensation a few months back but the XP1600 is an awesome replacement.
 
Hi Tiger.
Thanks for everything.
Today I have sold my pretty good Nanya 7ns DDR RAM to buy a stick of Samsung pc2700.
Maybe this kind of ram is not good as my previous Nanya one but in my town is very difficult to find something better than common Taiwanese ram sticks. However it was impossible for me to go over 150(2-2-2-5-1) mhz with my previous ram stick so I was forced to buy somewhat faster.
So this evening I have tried to push my cpu to its limit...these are my results:

155*10.5 @2.20(~1628 mhz) The cpu was not pretty stable and my PSU seemed to be not good enough. Only One on tree times the system started properly, the other two times it did not.

160*10.0 @2.20(~1600 mhz) The cpu seems to be stable, have played UT2003 for sometime and am writing this test using these settings.

Is my intention later to reach ~1650 mhz but I surelly will have to increase the vcore a little more...maybe till to 2.25 volts.
Temps are becoming too high...the full load temp is about 50C using a H2O cooling system.
 
Tiger...is there any mod to make my 300watt PSU becoming more stable?
If I well remember the trick to make the system more stable is to increase the 5v isn't it?
 
Firstly I am very surprised the 300W can give the power you need. Don't expect it to last too long with that kind of output. You need to get a bigger power supply. I have a generic 350 that is suprisingly good but I would look for at least 400 and even 500W.
 
To make my system workin one on tree times(I am referring to the boot/start session) I had to unplug all my removable devices like cd/cd-r ones.
The matter is that I have no enough money to buy now another power supply.
Have found around the net a trick that allows to boost the 350w Enermax PSU but obiviously I have no Enermax PSU ;)
 
My personal advice is that if you don't have a good grounding in electronic engineering and design then don't play with a power supply. They are quite sophisticated devices and a mod can go very wrong and take everything with it. This apart from the fact that they can give you one hell of an electric shock if you don't hand le them with great care. I repair them very occassionally but don't like fiddling around with them.
 
Thanks for your advertise Tiger.
However today I have well pondered this question and decided to do not do any possible mod. Then I went around on some online shops to see if there was something I could buy...so I have found a nice Thermaltake 430watt PSU that costs 63 euros + vat.
I have found a 430watt Enermax psu too but it cost a little more, somewhat like 93 euros + vat.
Prices are not so cheap ;)
 
Having given your problem some thought why not go the dual psu route. You could probably pick up an old 250 or 300W atx psu very cheaply and then run them in tandem using one for the m/board and the other to drive the fans and hard drives. It realy only involves splicing the green and one black on each atx to get them to power up simultaneously.
 
That's a nice idea ;) I just have another 300 watt psu but the matter would be where to position the second psu inside the case.
However today have tried to start my system with no devices pluged to the psu(both h.d and cd drives) and the result was not good as I was expecting: the system started not all the times as before.
Tiger, is going some wrong here?
 
Today I have learnt something about my cpu.
So...this nice processor don t like to have its multipier factor changed, both under and upper 10.0x
When I try to downgrade the multipier to 9.0 or to 9.5 it boots about one on tree times and the same thing happens when I increase it till to 10.5x
Another important question is the vcore that this cpu can support.
The max vcore that it accepts is 2.15 volts.
Have tried and tried today to find the best settings for my system and my cpu and found that the best frequency that I can obtein is 159*10.0(~1590 mhz).
The matter now is that with a FSB of 159 mhz the pci freq. is going to be 31.8 mhz and personally I don t like a frequency like this one for my pci bus.
However for tommor there would be some good news...maybe I will sell my great cpu and could by a XP1600+ one.
Hope that everything will go to the right verse.
Tiger...you surelly could not believe me but I am so sad about the idea to see my cpu going away from me tomorrow.
 
Michal - everything you are saying is consistent with a lack of power to the cpu. I can empathise with you on the cpu - get kind of attached to them.
 
Here I am again ;)
Hope you will read this Tiger.
Today I have bought a nice XP1600+ AGOIA 0213 and have made some testing.
The result was that this cpu don t like a FSB greater than 160 mhz so now I am running it @ 1680 mhz only.
The question now is in going to unlock it so I wanted to ask you something about.
Especially what are the right materials that I have to use in unlocking this processor?
I have heard some people that are able to solder all the L1 vids but it would be surelly very difficult.
Thanks again Tiger.
 
I found the Xp very difficult to unlock but it turned that there is a coating on the dots. I will tell you how I did mine. Filled the pits with a non-conductive twin pack expoxy using a 1mm screw driver blade. Blob about the size of a pin head on the blade and this was allowed to drip into the each pit. Waited for them to dry and then trimmed the excess off with a hobby knife blade. Using a magnifying glass I carefully scraped the surface of each dot with the point of a sowing needle and then cleaned the area well with acetone. Then I used the same screw driver blade to connect each pair of dots with conductive paint used to repair the de-mist wires on the rear windshield of cars.
Soldering would be nearly impossible because the dots are recessed below the surface.
 
Tomorrow I will ponder it.
Want to try if soldering pits would be posible or not...if not I will you use your method.
Thanks again Tiger.
 
If you get any conductive material in the pits it will cause a short to ground and the proc will probably never work again. If you plan to solder the pits must be filled first with non-conductive material. I would be careful with heat on the proc because the body is a epoxy composite material not like the old Athlons which were a ceramic type.
 
About the matter about filling pits I think there won t be any problem but about your second statement I have some doubts...I mean you are right in saying to be carefull by using a solder cause the cpu material is not ceramic but somewhat like plastic material.
Hope I will do a good job otherwise we will listen us not early ;)
 
Have made the first step, have filled pits with a super glue.
You were right Tiger...it is near impossible to solder all the vids without making somewhat like a mess.
It took me sometime to make a good work by filling pits...have tried to make a better and cleaner work than some other pics I have found around the net.
Now I'm waiting for the conductive pencil.
 
I learned few more things in these days.
The first one was that unlocking an Athlon XP is pretty difficult and no all materials are easy to find in manner to make this procedure.
So I have wasted two days in going around and finding somewhat that could be a electrically conductive pencil.
The one I have found was not e.conductive at all.
So I had to leave the idea to have my cpu unlocked.
The second one thing I learned was somewhat marvellous, I realized that my cpu could go above 160 mhz.
Infact the problem was caused by the RAM timings that did not allow me to get more speed.
Increasing cas timings to the worst ones I finally was able to reach the awesome frequency of 1917 mhz(182.56*10.5).
Before reaching this limit the following frequencies were my step by step experience:

166*10.5 @ 1.75
170*10.5 @ 1.75
172*10.5 @ 1.80
174*10.5(1827) @ 1.80
176*10.5(1848) @ 1.80
178*10.5(1869) @ 1.825
180*10.5(1890) @ 1.925
182*10.5(~1917) @ 2.15
(As you can see breaking the ~1870 mhz wall frequency the vcore becomes somewhat like exponential...the cpu begins to say you: "Wait! You are going to ask something more than my capabilities!")
Running my machine at this frequency requires a good H2O cooling system. Sometime the full-load temperature goes up till to 47-50 C degrees so it would be impossible to reach the same overclock result with air systems.
Obivious the vcore is high too, it goes near 2.15 volts.
The latest thing I want to say is that UT2003 rocks with my PowerVR KyroII ;)
 
Back