I have got the camera, will upload what I hope is some clear enough pictures in a while.
I took heatsinks off the MOSFETs that I had thermal padded on, as they were in the way of taking clear photos.
Decided to leave them off after re-installing the CPU HS. Will put them back after the new caps are fitted.
I thought I would have another go at an overclock, to see how hot that area got without HS's there.
At 3700nhz 1.44vcore, the (CPU MOSFETs?) reached 158c. I couldn't find any info regarding max operating temps for them, I thought I had read somewhere 155c was the limit, but could be very wrong about that.
I didn't use a DMM on the Vcore, but if you look at the Speedfan voltage you can see how it's jumped about quite a bit. (vdroop pencil mod in place)
Edit: After some more testing using Intel Burntest, it looks like I get the same results with voltage, when running anything around 3.6ghz. Max stable voltage I can use is 1.38volts, after that it just drops voltage down(even with pencil mod) and thus fails stability tests.
Edit2: More testing with a DMM for vcore readings- Intel Burn Test set to run x10
450x8 - Vcore set at 1.45v in bios- reads 1.445 in windows, running IBT, lowest voltage was 1.406v=Passed test
463x8 - Vcore set at 1.465v in bios reads 1.465v in windows, after running IBT, lowest recorded voltage was 1.405v=Failed test
Maximum emperatures stayed around the same(under 70c). So I assume it's the voltage drop that is causing the failures.
Also thinking my vdroop pencil mod has stopped working and needs reapplying.
Edit3: Reapplied vdroop, and tested again. When testing @ 3700mhz(bios set at 1.45v/ real vore 1.465v DMM), vcore drops to 1.416v under heavy load, and flashed 1.3volts just before pc rebooted itself.
One thing I have really noticed(which have read many times before though) is how useless it is using voltage monitors in windows. They really have no relation to what voltage the CPU is getting at any given time. At a lower stable overclock the CPU voltage stayed completely stable on the DMM, where speedfan showed massive fluctuations.
I monitored temperature of the MOSFETS next to the CPU and as they reach higher temps the voltage does seem to drop more and more. With the PC case open then reached 146c, with it closed 155c.
I wonder am I flogging a dead horse trying to mod this board in hopes of better voltage stability?
Would this cap mod help this vcore stability issue, or is it really just to save cap damage? I imagine its the latter, and it's the 3 phase power that's to blame.
I took heatsinks off the MOSFETs that I had thermal padded on, as they were in the way of taking clear photos.
Decided to leave them off after re-installing the CPU HS. Will put them back after the new caps are fitted.
I thought I would have another go at an overclock, to see how hot that area got without HS's there.
At 3700nhz 1.44vcore, the (CPU MOSFETs?) reached 158c. I couldn't find any info regarding max operating temps for them, I thought I had read somewhere 155c was the limit, but could be very wrong about that.
I didn't use a DMM on the Vcore, but if you look at the Speedfan voltage you can see how it's jumped about quite a bit. (vdroop pencil mod in place)
Edit: After some more testing using Intel Burntest, it looks like I get the same results with voltage, when running anything around 3.6ghz. Max stable voltage I can use is 1.38volts, after that it just drops voltage down(even with pencil mod) and thus fails stability tests.
Edit2: More testing with a DMM for vcore readings- Intel Burn Test set to run x10
450x8 - Vcore set at 1.45v in bios- reads 1.445 in windows, running IBT, lowest voltage was 1.406v=Passed test
463x8 - Vcore set at 1.465v in bios reads 1.465v in windows, after running IBT, lowest recorded voltage was 1.405v=Failed test
Maximum emperatures stayed around the same(under 70c). So I assume it's the voltage drop that is causing the failures.
Also thinking my vdroop pencil mod has stopped working and needs reapplying.
Edit3: Reapplied vdroop, and tested again. When testing @ 3700mhz(bios set at 1.45v/ real vore 1.465v DMM), vcore drops to 1.416v under heavy load, and flashed 1.3volts just before pc rebooted itself.
One thing I have really noticed(which have read many times before though) is how useless it is using voltage monitors in windows. They really have no relation to what voltage the CPU is getting at any given time. At a lower stable overclock the CPU voltage stayed completely stable on the DMM, where speedfan showed massive fluctuations.
I monitored temperature of the MOSFETS next to the CPU and as they reach higher temps the voltage does seem to drop more and more. With the PC case open then reached 146c, with it closed 155c.
I wonder am I flogging a dead horse trying to mod this board in hopes of better voltage stability?
Would this cap mod help this vcore stability issue, or is it really just to save cap damage? I imagine its the latter, and it's the 3 phase power that's to blame.
Last edited: