Modifying and practicing overclocking on a graphical card can damage it irreparably.
It does some time, I read in a forum as somebody had managed to put a ViperII to 166MHz, this had obtained it clearing two resistance and replacing them by others, this way profit to increase the voltage to which the card works, so I decided to also modify my card. Until then, the only modifications that had done on my ViperII, they were to put a ventilator in the dissipator (this is very important) and small radiators to him in the memory, these sticks them using a little (very little) glue in the corners and thermal silicone by everything, not if this measurement improves the endurance of the card in situation of overclocking, but as the cost is minimum, not this of more. It is necessary to try to have all chilliest possible one.
I found as the exit voltage is regulated:
Vout=Vref·(1+R2/R1)+I Adj ·r2
The Vref is typically 1.25V, and I ADJ 50uA or what is the same 50·10-6 To, which does that the second I finish is despicable, so I solely consider the first one I finish. The resistance that regulate it are noticeable in the grafica card like R1=R104 and R2=R105, are very small resistance, which can scare to but of one, personally I am scared, how I am going to enter has with my tin soldering iron, but those technical problems leave them for the end and would begin to make theoretical measurements and calculations.
With the help of multímetro I obtained the values of the resistance, R104=108.2 W and R105=58.7 W , with the expression that relates the values of the resistance I calculate that it works to a 1.93V. voltage Which I look for is a voltage of greater exit, in addition I want to be able to choose it to my taste and necessity, chooses to replace R105 by a potentiometer (a variable resistance), this way and using multímetro to measure the resistance I can obtain the voltage that wishes.
The gain control of the music apparatuses is a potentiometer, I took one broken radio, I disassembled it and I measured the possible values of resistance that could obtain with that potentiometer and I saw that they served to me, since he was quite stable and it could obtain from almost 0 W to very high values of the order of the K W . Takes off it, me cost a little, since there am to admit that the tin welder is not mine, and much less to use it to desoldar...
After clearing I welded it to him a pair of threads, to be able to connect it?cómodamente? the hollow which it must have when it loosen the resistance R105. approached the point without return, had to clear the resistance, I decided to try a little before entering task, disassembles an old disquetera of 3.5? that towards much that had let work and with joy it can see multitude of that type of resistance, I tried all the process several times, to clear the resistance and to weld threads, to even return to put the resistance, just in case something failed, and all the tests finished in success, was much but simple of which it had thought, simply it warmed up the edges of the resistance and doing pressure of displacement they did not take in loosen, once loose was left the tin that maintained them united to the board, so simply it had to place upon the?montoncitos? of tin the thread that had welded to the potentiometer, to touch with the soldering iron and remained united perfectly.
When already tapeworm confidence sufficient takes card and I repeated what it had done before, I had problems to loosen resistance, it gave to the sensation of that was of another type, that tapeworm small legs inserted in the board, but finally loosen and I could see that it was just as those of the tests, I took the potentiometer and I welded them, I had much taken care of of which the threads in their part without cover did not touch anything, finally I placed a plastic underneath the potentiometer to avoid that it caused some short circuit. In the end I am left this...
and sight more close by:
Quite miserly, but very efficient the plastic....vale... tapeworm haste to prove it and was first that I found, by the way, for the peculiar ones, is the sticker of monitor;D
The operation had been a success. Single it was to regulate the potentiometer and to prove the card. I began raising a little, I put 70 to him W , Vout of 2,05 V. Llegaba the moment from greater voltage, thorn the card, connect the ventilator, the cable of the monitor and ignite the computer, goes perfectly, slight lightening, I put it to 155 and I prove 3dmark2000, goes perfectly, before it had never obtained that went to 155MHz, I put it to 160MHz and also it goes, I put it to 166 and when beginning, the computer is hung. I extinguish and I raise a little but the resistance, I put it to 80, I take the computer, all good, return to put it to 166MHz, prove 3dmark2000 and works, I raise 170MHz becomes to hang. Again I modify the value of the potentiometer to 95 W , return to prove to 170MHz and works, I raise 175MHz and it is hung, finally I decide to leave it thus, a voltage superior to the 2.4V seems to me excessive. The final result is the totally stable card working Core to 170MHz and the memory to 200MHz , when before the maximum that could obtain to era 150MHz of Core and 186MHz in the memory. Not this nothing badly, considering that the cost to be 0 ptas and I can choose the voltage to which I want that the card works. Now benchmark only reduces to make a few to be able to evaluate the increase of yield. Soon I will include, now this forming or my computer, finish pillar celeron 600, at the moment goes by 927MHz, but or we will see..., because that when it has all or, them coat, at the moment in 3DMark2000, I obtain a Fill rate of 671 MTexel/s.