Re: Problem booting 9NDA3J with AMD Athlon 64 4000+

Chris wrote:
> Hi,
>
> I have just tried to upgrade my 9NDA3J motherboard from an Athlon 64
> 3000+ to an Athlon 64 4000+ and can't get it to boot normally :-( .
>
> This is the processor:
>
> http://products.amd.com/en-us/Deskto...ail.aspx?id=93
>
> The system was working fine with the 3000+. I fitted the 4000+ with new
> heatsink compound, now when I press the front panel power button the
> fans spin for a second or so and the motherboard LEDs flash, then
> nothing else happens. If I press the power button a second time, nothing
> at all happens (not even fans / LEDs) until I switch the incoming supply
> off and on again.
>
> The strange thing is, if I switch off the incoming power, hold in the
> front panel power button then switch the incoming power on, it boots and
> runs perfectly?
>
> I tried taking out all the drives and cards, made no difference. I tried
> changing the 300W power supply to a Corsair HX520W power supply, made no
> difference. I did wonder if particular supply rails are not supplying
> enough power, but haven't been able to find any info on recommended
> minimum current on each supply rail to check?
>
> I have now changed back to the 3000+ and it is working fine again.
>
> Any suggestions / recommendations gratefully received.
>
> Cheers
>
> Chris

It kinda sounds like some power component is latching off, like
an overloaded Vcore. In other words, it could be a problem with
the motherboard itself, and the Vcore regulator next to the
CPU socket. If it was just a BIOS problem, then the board
should always respond to the pressing of the power button.

Paul

Chris wrote:
> Paul wrote:
>>
>> It kinda sounds like some power component is latching off, like
>> an overloaded Vcore. In other words, it could be a problem with
>> the motherboard itself, and the Vcore regulator next to the
>> CPU socket. If it was just a BIOS problem, then the board
>> should always respond to the pressing of the power button.
>>
>> Paul
>
> Thanks Paul
>
> After more searching, I finally hit on the 'right' search terms and
> found the answer (for my motherboard at least):
>
> http://www.aoaforums.com/forum/epox-...tml#post377135 (post number 30)
>
> Now a happy chappy
>
> Cheers
>
> Chris
>

http://www.aoaforums.com/forum/epox-...tml#post377135

"Dear Sir

May be your board have hardware support Venice and new code processor issue.

Please try to remove or cut the on board's one Diod (D1) for this issue check.
In english this means to remove a diode from the 9nda3 and then these cpus post.
I had to do the same thing for my opteron 165 to post. See the atached files.
This of course voids your epox warranty. If you dont feel adventurous enough
to perform this 2 second mod yourself and dont want to void your warranty,
Epox will do it for you. Hopefully the pics come through. Im a noob to this
forum.

http://www.aoaforums.com/forum/attac...ice-epox-1.bmp
http://www.aoaforums.com/forum/attac...ice-epox-2.bmp
http://www.aoaforums.com/forum/attac...ice-epox-3.bmp

That diode appears to be next to an Attansic ATTP1 eight pin chip. Now,
the only question I've got, is what is an overtemperature controller
doing on a S939 motherboard ?

http://web.archive.org/web/200305091.../attp1_1.0.pdf

The S939 has its own signal called THERMTRIP, to protect the processor.
A chip like the ATTP1 would have been something from the AthlonXP era.
And since ATTP1 behaved differently when the 4000+ was plugged in,
it implies the thermal diode in the CPU must be connected to it.
The diode being snipped, could be in parallel with it for some
reason. Really strange.

http://www.amd.com/us-en/assets/cont...docs/31411.pdf

The ATTP1 appears to have signals to route the PS_ON# signal, which is
how the ATTP1 can turn off the power.

Paul