GIGABYTE GA-990FXA-UD3 AM3+ Overclocking problem

[Buta-kun]

I would like to bring to your attention my problem with overclocking the 990FXA - UD3 rev4 (FX-8350 + Noctua NH- D14; GSkill Ares 1866 CL9 ).
I got a stable operation at 4.5 GHz with Vcore of about 1.38V and LLC medium. Given the low temperatures (slightly above 40°C ), I would like to increase the clock frequency but the CPU require voltages higher than 1.4V. When I set in the BIOS (rev. F2) a higher Vcore to 1.4V, the pc will not start. The mobo power off and restart it is necessary to remove the power cord. If I boot the PC with stable settings and then varied the Vcore by ET6, not just the beginning of the stability test the PC is turned off. I tried to enable or disable the power saving settings but nothing changes. The power supply is an old but efficient Thermaltake PP680W. I set the DDR3 clock 1333 MHz to prevent any instability caused by the memories.
Does anyone have any explanation for me? 
Thank you all.

PS: I'm sorry but I have to use google translate .....

[Vezina]

First impression is that you ve hit a board limitation.The board is going in some protection state.
FX needs a little more voltage to get stable over 4,5Ghz , around 1,45V.
You should also test with another power supply , just in case ,having in mind your PSU is old.

What happens if you disable one module or 2 ,are you able to push it higher ?!

[Buta-kun]

Thanks for the reply.
I have not tried to disable some cores but I can try, as soon as possible.
I do not have another PSU......
Ciao

[Buta-kun]

Enabling 6 cores, setting Vcore to 1.4V, the AMD Overdrive stability test fails at 4.6 GHz.
However, if you increase the Vcore (with power saving settings enabled by BIOS), AMD Overdrive limits the CPU multi to 22 (set in the BIOS).
If I change the parameters from the BIOS instead of AMD Overdrive, increasing Vcore, the PC is turned off.
With all the core ENABLED, the PC is stable at 4.4 GHz with approximately 1.32V Vcore. I can not stabilize at 4.5 kHz because if imposed 1.4V, the PC is switched off; with 1.3875V, the OCCT lynx test fails after 30 minutes and when I do video encoding, the PC is turned off again.
The temperatures do not exceed 44 °C.
At this point I decided to work at 4.4 GHz, but I remain curious to see what happens ....
Thanks!