It seems that no one has posted any Vmods of this motherboard. I am trying to vmod mine, and I would appreciate if the experts here can tell me if I am on the right track. (I probably won't mind killig the motherboard, but I certanly don't want to kill my last pair of Corsair BH-5s and the FX-53!)
Anyways, AV8 uses the Winbond W83303D for VDIMM, and soldering a 20ohm connecting the emplty soldering point to ground gives me 0.3V more, and a 10ohm 0.5V more.
Last edited by HKY; 09-29-2004 at 03:12 AM.
Vcore is regulated by a 6559, so a VR to the sixth leg should do the trick.
VDIMM ca be measured from the right leg of the center MOSFET near the DIMM socket and Vcore can be measured from one of the coils.
Please let me know if you have any thoughts about the VDIMM mod. Thanks.
For the Vcore, on the ISL6559CB the target pin(s) would be either #10 (FB) or #13 (Vsen). Do a connectivity test between the coil measuring Vcore and those 2 pins to see which one (if not both) is connected to the Vcore read point. A vast majority of the time the FB pin is the target on most mods, however sometimes (Vcore) the Vsen is the actual mod target as FB is tied to otehr controllers and not directly tied to Vcore.
On the Vdimm, I am at school and dont have that data sheet with me, but I am pretty certain its pin #19 or #21 that controls Vdimm for the W83303D IC. You have already found where you can add resistance to boost the Vdimm, so I would think that a 200-1000 ohm VR would work nicely where you have the fixed resistor in your pic.
Keep in mind that since you are adding a resistance where none exists (as far as you are not "directly" soldering your VR to another resistor), that you will increase the resistance on the VR to increase the Vdimm. Whereas you would typically decrease the resistance of a VR to increase the voltage. So, be sure to preset the VR to 0 ohms before soldering it into place.
GL with it and let us know what happens.
However, be sure you first check which pin is needed for the Vcore mod and then you will need to do a resistance check between that pin and ground to see what value VR is needed for this mod.
I like my women the way I like my processors...naked.
I am not the HellFire that writes bios files. I work strictly on voltage modifications.
Hi Fire, thanks for your help. I've been testing the VDIMM mod for a couple of hours (with a pair of cheap DDR400s, of course)and it seems to work fine. I am getting 3.25V (2.8V in BIOS)with the 20 ohm resistor, as shown in my photo. Nothing seems to be melting yet so I guess its probably OK.
As for the Vore it looks like its pin 10. I'll probably do this in the next couple of days when I have more time.
Cheers!
At 3.25v Vdimm, you will have zero problems with that BH5.
Let us know how the Vcore works out.
I like my women the way I like my processors...naked.
I am not the HellFire that writes bios files. I work strictly on voltage modifications.
HKY, About the Vmem-mod.
Have you soldered the resistor to left or right empty soldering point? It's hard to say from picture :-)
Ok, I did the Vmem-mod but for some reason it's not working very well.
I can get 2,86V without mods and the board can run around 218 MHz 1:1 with BH-5 memory.
When I raise Vmem to 2,88V I get BSOD very easily and the higher voltages are even worse.
Pics of my mods:
does the Vtt follows when u raise the Vdimm?
No I don't think so.Originally Posted by enduracell
I believe if the winbond controller is the same as on the other abit board then that chip also controls the Vtt. Whichever pin it is you should try and find it via the datasheets because Vtt will play a good role in your ram's overclockability.
Level 80 UD Death Knight on Murmur (Beta)
it doesn't matter what u think, it matters what it is. KV8 pro had Vdimm on 3.2 and reduced it to 2.8 on 1.5 BIOS because VTT didn't follow over 2.8 and caused major instability. So look in the Bios monitor screen for VTT sense or measure it on a fet around the neighbourhood.
heh heh, VTT doesn't follow.. with 2,8V from BIOS I measured 2,86V Vmem and 1,43V VTT from motherboard.
When I add the 20ohm resistor the VTT remains same.
Any ideas for VTT-mod?
try this one, if this R is VTTsense...
@Pc Ice ovp on AV8 is @1.95 and the mobo gives 1.85 out of the box...
Last edited by enduracell; 10-04-2004 at 12:45 AM.
I just received datasheet for W83303B from Winbond:
http://www.muropaketti.com/temp/w83303d.pdf
Thanks guys.
Sampsa is right. VTT does not follow and the increase in VDIMM is not very useful. I only got about 5mhz more from 2.8V to 3.15V.
Enduracell, I'll try that later to see how it works.
Endu, that doesn't seem to work. Nothing happened even when I turned the VR down to 5 ohms.try this one, if this R is VTTsense...
@Pc Ice ovp on AV8 is @1.95 and the mobo gives 1.85 out of the box...
That mod wouldnt work.
You need to remove the resistor, its a 0 ohm resistor just bridging the gap, and replace it with a 200-1000 ohm VR soldered to both solder pads in order to increase the Vtt. The VR has to be set to 0 ohms before soldering it in place, not max.
And just in case it still acts odd, you may have to solder the other vr leg to ground. Just make certain that the vr legs that are soldered to the the sodler pads are the legs reading 0 ohms before soldering the VR into place.
I like my women the way I like my processors...naked.
I am not the HellFire that writes bios files. I work strictly on voltage modifications.
HKY, check out the datasheet. The pin number 26. is the 1.2V_SEN and I think it needs some modification.
I've no idea what enduracell tried to tell with his instructions.
Ok my mistake I haven't done it, just speculating. Do what Hellfire says knows far better than me those thinks
I think is BT_SEN (pin 7), correct me if I m wrong. Also check this out.
Guys, remember that there are 2 instances of Vtt with A64 boards. One is driving the cpu and one drives the memory termination bus. If you try to boost the voltage by modding to pin #26, say good bye to your cpu.
It is the Vbt pin (#7) on the W83303D IC guys. Trust me.
I like my women the way I like my processors...naked.
I am not the HellFire that writes bios files. I work strictly on voltage modifications.
Ok, thanks. I'll try it later this week and let you know what happened, if HKY doesn't do it first
Yea I'm going to try out Fire's mod in the next few days.
Sampsa I have no idea which is the correct pin just by reading the data sheet but I am betting Fire is right. (I found the VDIMM mod by trial and error on the board only....)
Ok here is what I did:
I removed the 0 ohm resistor which was attached to pin 7.
I replaced it with 1 kilo ohm VR and set it to 0 ohm.
Unfortunately Vtt didn't change when I raised the VR to 1000 ohms.
What next? Solder the other leg of VR to ground?
I did the same thing with the same result, and 5 minutes ago I connected the other leg of the VR to ground and now the system won't boot up any more, even after I removed the ground. It stopped after "C1", (which means checking DRAM), and then return to "02" and then "C1" again and again. It can't even be turned off by the power on switch. The RAMs are OK because I checked them in another system.
I am afraid that either the motherboard (maybe the uGuru chip) or the memory controller of the CPU is dead.... I'll have to find another 939 board to check it out. Hopefully its the motherboard!
Last edited by HKY; 10-07-2004 at 07:36 AM.
Posting Permissions
Reply With Quote
Bookmarks