The 939SLI32-eSATA2 has a Vcore adjustment range up to only 1.45V and a Vdimm maximum of 2.8V when set to "Ultra High"....
Carefull inspection of the board has shown that the 939SLI32-eSATA2 has the exact same Vcore and almost the same Vdimm power supply as the 939Dual-SATA2 motherboard.....
Those voltage mods are here... Note that both the 3.15Vdimm and 3.3Vdimm mods are the same even though the circuit is slightly different. A variable Vdimm will be devoloped shortly...
I have performed the Vcore mod on the 939SLI32-eSATA2 using a 10K resistor and have measured:
1.46V when the BIOS is set to 1.3V 1.63V when the BIOS is set to 1.45V
If you do any mod, your doing them at your own risk. The Rebels Haven Computer Forum will not be responsible for damaged systems or components... If your unsure about the mods, I strongly suggest you do not implement them...
Vcore Mod For 1.70Vcore Soldering a 5.6K resistor as shown below, increases the Vcore by 0.25V. The BIOS will read the new Vcore fairly well and the max setting of 1.45V(Winchester) in the BIOS, yields 1.7V to the CPU. If the 0.25V increase is too much for your liking, a Variable 20K or 50K resistor can be used. Here's some resistors and the Vcore increase they will generate, for the CPU's available:
5.6K-->+0.25V Use this for "Clawhammer" type CPU's 7.5K-->+0.20V Use this for "Winchester" type CPU's 10K--->+0.15V Use this for "Venice/SanDiego" type CPU's
Note that the resistor can be any tolerence you can find, although I recommend 5% or better to get a fairly accurate output. Also, there is almost no wattage in the resistor, so as small as possible will make soldering easier....
You do not need any Vcore mod if your running "Opteron" type CPU's. But if you want a +0.1 increase, use a 12K resistor.
The resistor is connected to the left side of C607 and the right side of C600:
For those with a Digital Multimeter, here is where to measure Vcore. Any one of the 3 points is Vcore. Connect the ground or negative lead, to the I/O plate in the rear of the system, or to circuit ground if you know a ground point.
-------------------- Too Many Computers,... Too Little Time ..... Damn Phenoms! Posts: 25182 | From: Fire Island, NY | Registered: Feb 2003
To increase the Vdimm, use the following resistor:
1K increases the Vdimm by +0.11V 470 Ohm increases Vdimm by +0.24V
Add these increments, depending on which resistor you choose, to the following: Low = 2.58V Normal= 2.61V High = 2.72V Ultra = 2.81V
For a variable Vdimm, solder a 1K VR in series with a 470 Ohm. Install across Pin 9 and Pin 11...
This is where the mod resistor gets attached to:
If you do any mod, your doing them at your own risk. The Rebels Haven Computer Forum will not be responsible for damaged systems or components... If your unsure about the mods, I strongly suggest you do not implement them...
-------------------- Too Many Computers,... Too Little Time ..... Damn Phenoms! Posts: 25182 | From: Fire Island, NY | Registered: Feb 2003
posted
Both Mods have been implemented on my board and I have noticed a problem with the basic Vdimm circuit!
Under load, if Vdimm is set to more then 2.72V, running programs like Memtest, Prime, etc, Vdimm drops to 2.72V-2.77V and is jumping all around... This is very bad...
The "Ultra High" Vdimm setting does it also....
There appears to be fairly high voltage drop from the ATX power connector to the Vdimm Power Supply area that is causing the problem. I fed the 3.3V into the local Vdimm Power Supply area and the 2.8V setting was almost 100% stable...
I need to find a better spot to connect the 3.3V to...
What would also help is to increase the 3.3V rail to at least 3.45V....
Topic closed until I solve or give up on the issue....
-------------------- Too Many Computers,... Too Little Time ..... Damn Phenoms! Posts: 25182 | From: Fire Island, NY | Registered: Feb 2003
posted
This helps a little, but this Vdimm Power Supply circuit was never meant to go above 2.8V...
Here, I have run the 3.3V PSU rail right to the "Drain" pad of the Vdimm Regulator FET, the 2 pads are common. This increases the voltage available to the FET from 3.15V to 3.32V...
At 2.8 Vdimm, it's still bouncing around under load, but not nearly as bad. Also shown in the picture is the 470 Ohm Mod resistor soldered to Pin 9 and Pin 11 of the Vdimm Control chip, U13:
Speaking of U13, The 5V supply to the chip is measuring 4.84V, while the 5V rail in my system is measuring 5.11V. A jumper from the 5V rail to Pin 4 of U13, would probably help....
-------------------- Too Many Computers,... Too Little Time ..... Damn Phenoms! Posts: 25182 | From: Fire Island, NY | Registered: Feb 2003