This one took quite a bit of experimenting and to be honest, the yields were not worth the aggrivation from testing... There's been an ongoing debate as to whether 2200uF in place of the 1500uF caps were beneficial to the VP6 board for overclocking... Some say yes, and some say no... I was one that said no. The way I figured it was this: The engineers that designed this board are more knowledgeable then all of us put together on how these boards work... Remember, the reason for the capacitor failure was not the value of the cap, but rather a poor quality cap, which the engineers and designers have no control over.
Ok, here's the experiment:
Phase 1:
I acquired a VP6 that had bad caps off Ebay. I recapped this board with all stock values, not modifying anything.
Hardware used:
-Matched Pair of 933mhz P3 CPU's (S-Spec SL4ME)
-OCZ Copper HSF's
-512mb stick of Crucial/Micron PC133 SDRAM
-Nvidia GeForce2 Video card
-Adaptec 29160 SCSI Controller
-Quantum Atlas V 9.1GB U160 HDD
-Enermax 350w power supply
-Windows XP Pro SP1
Phase 2:
Using the above listed hardware, I pushed the FSB as high as I could before the board would fail to POST. I was able to get it to 1022mhz at a 146mhz FSB with a core voltage of 1.8.
The board would POST at this speed, however, it wouldn't boot into an OS without freezing or rebooting itself partially through boot.
I turned it down to 1000mhz at 143mhz FSB, and that seemed to be the magic number and the highest I could push it and it run reliably. At this speed, it would fully boot and run SiSoft Sandra 2004 Professional's burn-in cycle. I let this run for 48 hours, and it had no issues.
Phase 3:
I replaced the 8 1500uF caps around the VRM with 2200uF caps and ran the same test again. With these caps now @ 2200uF, the FSB would not run reliably at anything higher than 143, the same as with the 1500uF caps.
Phase 4:
I now replaced the 4 1500uF caps around the RAM slots with 2200uF caps and once again, tested again.... This time there was a bit of an overclocking improvement. I was able to run the FSB @ 146 reliably. SiSoft's burn-in ran for 48 hours without a hiccup. I then bumped the FSB speed up to 150. The board POSTed at 1050/150, and actually booted into Windows XP. I ran SiSoft again, but it would not run for over an hour without a BSoD or a freeze. I then backed off the FSB in 1mhz increments until I got back to a reliable setting, which ended up being 146MHz FSB.
Phase 5:
I removed the 8 2200uF caps I installed around the VRMs and reinstalled the 1500uF caps and ran the tests again.
There was no change, I was still able to run the board @ 146MHz FSB relaibly, but not able to take it any higher.
Phase 6:
I removed the 4 2200uF caps around the RAM slots and reinstalled 1500uF back.
This rendered the FSB speed of 146 unstable. I had to clock it back down to 143mhz FSB, and the board was once again stable.
Phase 7:
Once again, I removed the 4 1500uF caps around the RAM slots and reinstalled the 2200uF caps, and once again, 146FSB was stable.
The aftermath and conclusion:
Well, this was a 2 week ordeal that I knew wouldn't be anything earth-shattering... I've seen people get the VP6 to run higher than this, so maybe my pair of CPU's or RAM weren't very overclocking friendly... Whatever the reason, take note of the consistancies....
With the stock cap values, the highest I could push the board was 143 FSB. With all caps at 2200uF I could run it @ a FSB speed of 146MHz. With only the caps around the RAM slots @ 2200uF, the board would run @ 146FSB. Then all taken back to 1500uF, 146FSB would no longer run, and 143 was the highest I could go.
Was it worth the aggrivation?! SHIT NO! Not for a measly 3MHz of power.
However, if you would still like to mod your board, here's the caps I recommend replacing with 2200uF instead of 1500:
EC8, EC22, EC24, and EC28
If you try this and yield better or worse result, feel free to post them! I'd be curious to know!!
Ok, here's the experiment:
Phase 1:
I acquired a VP6 that had bad caps off Ebay. I recapped this board with all stock values, not modifying anything.
Hardware used:
-Matched Pair of 933mhz P3 CPU's (S-Spec SL4ME)
-OCZ Copper HSF's
-512mb stick of Crucial/Micron PC133 SDRAM
-Nvidia GeForce2 Video card
-Adaptec 29160 SCSI Controller
-Quantum Atlas V 9.1GB U160 HDD
-Enermax 350w power supply
-Windows XP Pro SP1
Phase 2:
Using the above listed hardware, I pushed the FSB as high as I could before the board would fail to POST. I was able to get it to 1022mhz at a 146mhz FSB with a core voltage of 1.8.
The board would POST at this speed, however, it wouldn't boot into an OS without freezing or rebooting itself partially through boot.
I turned it down to 1000mhz at 143mhz FSB, and that seemed to be the magic number and the highest I could push it and it run reliably. At this speed, it would fully boot and run SiSoft Sandra 2004 Professional's burn-in cycle. I let this run for 48 hours, and it had no issues.
Phase 3:
I replaced the 8 1500uF caps around the VRM with 2200uF caps and ran the same test again. With these caps now @ 2200uF, the FSB would not run reliably at anything higher than 143, the same as with the 1500uF caps.
Phase 4:
I now replaced the 4 1500uF caps around the RAM slots with 2200uF caps and once again, tested again.... This time there was a bit of an overclocking improvement. I was able to run the FSB @ 146 reliably. SiSoft's burn-in ran for 48 hours without a hiccup. I then bumped the FSB speed up to 150. The board POSTed at 1050/150, and actually booted into Windows XP. I ran SiSoft again, but it would not run for over an hour without a BSoD or a freeze. I then backed off the FSB in 1mhz increments until I got back to a reliable setting, which ended up being 146MHz FSB.
Phase 5:
I removed the 8 2200uF caps I installed around the VRMs and reinstalled the 1500uF caps and ran the tests again.
There was no change, I was still able to run the board @ 146MHz FSB relaibly, but not able to take it any higher.
Phase 6:
I removed the 4 2200uF caps around the RAM slots and reinstalled 1500uF back.
This rendered the FSB speed of 146 unstable. I had to clock it back down to 143mhz FSB, and the board was once again stable.
Phase 7:
Once again, I removed the 4 1500uF caps around the RAM slots and reinstalled the 2200uF caps, and once again, 146FSB was stable.
The aftermath and conclusion:
Well, this was a 2 week ordeal that I knew wouldn't be anything earth-shattering... I've seen people get the VP6 to run higher than this, so maybe my pair of CPU's or RAM weren't very overclocking friendly... Whatever the reason, take note of the consistancies....
With the stock cap values, the highest I could push the board was 143 FSB. With all caps at 2200uF I could run it @ a FSB speed of 146MHz. With only the caps around the RAM slots @ 2200uF, the board would run @ 146FSB. Then all taken back to 1500uF, 146FSB would no longer run, and 143 was the highest I could go.
Was it worth the aggrivation?! SHIT NO! Not for a measly 3MHz of power.
However, if you would still like to mod your board, here's the caps I recommend replacing with 2200uF instead of 1500:
EC8, EC22, EC24, and EC28
If you try this and yield better or worse result, feel free to post them! I'd be curious to know!!
Comment