FORCE ACORP 6A815EPD TO USE 133MHZ FSB

Here is my story. I bought the Acorp 6a815EPD before a couple of days and I was very pissed that I couldn't run my 2 Pentium 3 700(7x100)Mhz CPU's at 133Mhz FSB, as on my previous VP6 which has died since summer. There are 2 jumpers on the motherboard to select the FSB you want between 66,100 and 133Mhz but even if you specify 133Mhz the mobo will read automatically the FSB from the CPU's and set it. So the jumpers are useless. If you use 100Mhz Pentium's the memory will work at the same frequency so you waste another 33Mhz from your lovely memory. Even the Voltage of the CPU's will be set automatically from the mobo as there is no setting in the BIOS.

So after a lot of searching I found that I could change the FSB of the processor's and their Voltage by a Hardware Modification on the CPU or on the Socket. On each CPU there are 2 pins called BSEL0(AJ33)  and BSEL1(AJ31). These pins are responsible for setting the FSB of each CPU.

The exact combinations are :

BSEL1 BSEL0 Frequency
0 0 66MHz
0 1 100MHz
1 0 Reserved
1 1 133MHz

After some searching on Google I found a site where someone said that I had to brake off the BSEL1 pin in order to get 133Mhz FSB. In order not to brake the pin I decided to twist it under the CPU. So I did that and I plugged it in the Socket but it didn't work. And after I tried to bring it back the pin just broke. I went mad for a moment but after a couple of minutes a just received an email from someone who replied to me about a thread I opened before on the overclockers.com forum regarding the Acorp FSB problem. I was told by him that I had to Wrap BSEL0 and BSEL1 with a copper strand of speaker wire so the mobo would think that 133Mhz signal was coming in. I decided to do that but on the Socket because I thought it was safer and I was afraid that something might go wrong under the CPU. So I inserted a wire in the Socket connecting the BSEL0 and BSEL1 pins together. I used the second CPU because the first one had one pin less, but again nothing happened. So my last chance was to use the first CPU which had the BSEL1 pin taken out. I plugged it in the Socket and after the boot I went to Heaven:) The Bios was reporting 133Mhz FSB for the CPU and Memory. So I broke the same pin from the second CPU straight away and got both of them working at 133Mhz.

The Place of AJ31 and AJ33 pin's under the CPU:

 

 

 So the procedure that worked for me was first to set the mobo jumpers for the FSB to AUTO(that means both of them must be out), then to brake the BSEL1 pin from each Processor and finally  connect the BSLE0 and BSLE1 pins together from the Socket using a very short wire from a UTP cable.( I don't know if I had to brake the BSEL0 pin as well, but as long as it works that way I won't bother).

After I saw both the processors start at 933Mhz I had to increase their Voltage to 1.80v or 1.85v in order to get stability at this speed. So after reading an article on overclockers.com site I found that I could do that quite easy by connecting the VID0,VID1,VID2,VID3 pins to VSS pin depending what CPU stepping I had and what Voltage I wanted to get.

Here is the Voltage Definitation Table:

 

All the possible configurations:

Stepping
Default Voltage
Wanted Voltage
Pins Closed Internally
Pins to Connect to VSSL2
cA2/cB0
1.60V
1.65V
VID1, VID2
VID0
cA2/cB0
1.60V
2.00V
VID1, VID2
VID3
cA2/cB0
1.60V
2.05V
VID1, VID2
VID0, VID3
cA2/cB0
1.65V
2.05V
VID0, VID1, VID2
VID3
cC0
1.70V
1.75V
VID3
VID0
cC0
1.70V
1.80V
VID3
VID1
cC0
1.70V
1.85V
VID3
VID0, VID1
cC0
1.70V
1.90V
VID3
VID2
cC0
1.70V
1.95V
VID3
VID0, VID2
cC0
1.70V
2.00V
VID3
VID1, VID2
cC0
1.70V
2.05V
VID3
VID0, VID1, VID2
cD0
1.75V
1.85V
VID0, VID3
VID1
cD0
1.75V
1.95V
VID0, VID3
VID2
cD0
1.75V
2.05V
VID0, VID3
VID1, VID2

 

The Pins we need under the CPU:

 

 

And the Pins on the Socket:

 

My CPU's are cC0 stepping, so for 1.80v I had to connect VID1 to VSS. I did that by using a piece of wire again on the Socket. After the boot I got 1.80v but the PC was crashing on its way to boot XPMCE. Then I tried 1.85v and it booted successfully to windows and after torturing for 2 hours the system with 2 instances of Prime and had no crash I knew it was as stable as hell. After I tried to increase the FSB more thatn 133Mhz but at 140Mhz the system would crash even with 1.90v. I didnt try more Voltage because I was afraid that maybe something go wrong. I could boot using 137FSB but it was not big deal so I left it at 133Mhz. Maybe I try again in next couple of days using better fan. My temps now are 37-38 degrees idle and 42-43 load using 2 ALPHA PAL 6035  and a pair of crap and noisy YS Tech. Overall I am very pleased with the performance of my PC now as it is very stable as well , but after I get some Arctic Silver 3 and better Fans I will try for more:)

I hope that this article will help people to get their Acorp @133Mhz, but have in your mind that I don't take any  responsibility if you screw up your system, as this is a bit dangerous and there's always a chance to kill your components. Have Fun...

My Rig
Acorp 6A815EPD
2 X Pentium 3 FCPGA 700@933Mhz
2 X Kingston 256MB PC133 SDRAM(2-2-3-5/7)
WD 20GB Hard Drive for OS/Apps
Maxtor 80GB Hard Drive for Backup/Downloading
TNT2 Sparkle->Leadtek Ti4200(coming soon)
Lite-On CDRW 52X(coming soon)
Lite-On DVD 166S(coming soon)
Hitachi 8130 CDROM 16x(very old but still works perfectly)
Crap Case->Enermax 3051 Black Case(coming soon)
350W Unknown PSU

Here are some pictures for you...

CPU

CPU Socket 1 - CPU Socket 2

PC 1 - PC 2

My Messy Room1My Messy Room2

DEAD Smashed  VP6 1 - DEAD Smashed  VP6 2

                                                                                                 ©2006-Nikolaos Anastasiou