I recently jumped into watercooling and took my Duron 650 up to 952Mhz. The temps were low, but my Abit KT7AR was maxed out on the voltage at 1.85V. I could not go any higher and remain stable.
I began researching ways to go past the maximum 1.85V. I posted some messages in the forum at the Overclockers.com forum, and based on detailed answers from Colin, one of the Senior Board Members, I got pointed in the right direction.
I also found information at www.tweakhardware.com under the “Voltage Mod” links. I compiled this information and decided to try the modification.
Core voltage on Abit KT7, KT7A, and KT7AR motherboards is controlled via a programmable IC. It’s located just between the AGP slot and 4 capacitors.
This chip is a HIP6301. I searched the net and found a datasheet for this chip
The pin used for this modification is pin 7. Pin 7 will regulate how much voltage this chip will output. Placing a resistance between pin 7 and a ground point makes the chip think it’s not putting out enough voltage and it increases output to compensate.
This chip also has an over-voltage protection circuit. If it detects a voltage 15% over maximum it shuts itself down. A little math:
From this you see that about 2.1V is the maximum output before over-voltage protection kicks in. Above 2.1V your system will cease to function and you’ll have to unplug the power supply to reset it.
If you want to over-ride the over-voltage protection you’ll also need to wire a resistance to pin 10 on the HIP6301. This will again trick the chip into thinking all is normal and allow you to go higher than 2.1V. I did not implement this connection as I never plan on going over 2.1V.
Some hardware sites recommend using a fixed resistor from pin 7 to ground to accomplish a higher voltage output. Colin had a great idea of using a micro potentiometer instead, which would allow the voltage to be varied, so I took his advice and used a 47k-Ohm micro potentiometer, Radio Shack Part # 271-283.
A potentiometer is merely a variable resistor. These micro potentiometers are tiny so a project box or other mounting device is highly recommended. I chose a small 2.125″ L x 1.375″ W x .58″ H project box, Radio Shack Part # 270-288A. I mounted two micro potentiometers in the box in the event I decide to implement the over-voltage protection modification at a later date.
I marked the locations of the three legs on the micro potentiometers and used a small drill bit to drill holes for the legs to go through.
The project box is two halves with a single screw that holds them together. I drilled two more small holes in the side of the box for the 18 gauge solid core wire that would be used in this project. The center pin on the potentiometer is the ground connection and you can use either of the other two pins for the HIP6301 connection.
If you use the left side pin when facing the back of the potentiometer, then as you turn the potentiometer clockwise, voltage will increase. I wired it up to the right side pin, but no big deal. I just have to turn it counterclockwise to increase voltage. Wiring to the left pin keeps it consistent with what you’re accustomed to in consumer electronics.
I did not wire anything to the second potentiometer. 2.1V is pretty high, so I figured I didn’t need to implement the over-voltage protection mod at this time. It’s there just in case I need it down the road.
I used double sided tape to mount the project box to the southbridge chip. That’s about the only nice flat location I could find to mount this box on the motherboard. I wanted to keep everything mounted to the board to lessen the risk of snagging a wire and ripping up a trace down the road.
Keeping everything on the board also allows me to remove the motherboard from the case without having to worry about moving wires or the project box.
I looked carefully at the route the wires should take, and put red wire to the HIP6301 and white wire to ground, making sure to keep a low profile to the board and the shortest distance to the soldering locations. I used solid core wire as it holds a shape when bent, making it easier to route.
Some people recommend soldering straight to the pins on the HIP6301. I chose to use a soldering pad close to the IC. There is a place on the ABIT KT7, KT7A, and KT7AR that looks like it was meant for just such a resistor.
There are two pads. The one closest to the chip is the one you need. The location is marked with “R131”. This soldering pad connects straight to pin 7 on the chip.
The trick to making this solder connection successful is to “pre-tin” the wire with solder. Then touch it to the solder pad, and use the soldering iron to just melt the solder enough for it to connect.
The other connection goes to a ground point. You can use a motherboard mounting screw, solder to one of the mounting screw metal trim areas, or do as I did and solder to the ground on a transistor. I used a digital multi-meter to determine which transistors would work, then chose one with the shortest route.
Soldering to these transistors isn’t easy as the pins are coated with something that makes solder not want to stick.
I just looped the wire around the pin securely, and soldered it. It took a couple of attempts.
If you also opt for the Stage 2 modification to over-ride the over-voltage protection, then just connect it in the same fashion, only solder directly to pin 10 on the HIP6301 and put the other wire to ground.
This is how everything looks once in place.
Make sure the potentiometer is turned all the way down to MAX RESISTANCE. The system will not boot if it’s over 2.1V.
MAKE SURE YOU HAVE EXCELLENT COOLING ON YOUR CPU.
Fire up the PC. Go into your favorite voltage monitoring program, I use MBM5, and check out the voltage. With the potentiometer set to zero resistance, and the voltage set to 1.85V in the BIOS, my reading was 1.92V in MBM5 and VIA Hardware Monitor.
Even with the potentiometer set to zero there will be some resistance as the voltage flows through the wire.
I used a small flathead screwdriver to adjust the potentiometer to make sure it worked. The voltage will change real time with no need to reboot.
I watched as the voltage went from 1.92V up to 1.97V with about 1/8th turn.
If you did the Stage 2 mod to over-ride the over-voltage protection and want more than 2.1V then turn the resistance of your pin 10 potentiometer up slightly. Then turn the pin 7 potentiometer up some. Going over 2.3V is not advised.
I opted not to do the Stage 2 over-voltage protection mod. More than 2.1V is too risky for me. If the screen goes blank, or the system reboots, then you hit the over-voltage protection and you need to turn the pin 10 potentiometer resistance up a little more.
With this mod you’ll need steady hands, nerves of steel, and a good soldering iron. I recommend one of the grounded type irons made for soldering circuitry. I picked one up at Radio Shack for $6.99 and it came with a “micro” tip perfect for soldering small circuits.
This project, although tedious, was worth the effort. My Duron 650 (Blue Die) is now humming along at 1.02Ghz (7.5 x 136) @ 2.05V. My idle temps are around 30 C and my full load Prime95 Torture Test temp is 42 C. According to Radiate II this makes my effective C/W .12 with my custom-built watercooling setup.
I frequent the Overclockers.com forum regularly (nick AMDGuy) or you can e-mail me with any questions about my setup.