A7N8X 2.0 del & XP2500-M Barton

[Rob Hemmings]

Has anyone *not* been able to o/c a (1.45V) mobile XP2500 at
200FSB (ie as an XP3200) on an A7N8X deluxe 2.0 mobo?
I currently use a locked XP2500 Barton, but can't run it reliably at
above 185Mhz FSB and need just a little more speed.
Thanks,

 

[dino]

is it the processor or your ram?...I have 2 setups with Barton 2500's..one
locked and 1 unlocked..both can hit 200 FSB in my A7N8X-Deluxe rev2..I am
running OCZ ram. I currently have the unlocked one in this machine..but have
tested the locked one with no prob's. Did you up the voltage to the
processor at all?

 

[kmchow]

Do it at your own risk, but I hear increasing the voltage is sometimes
needed to keep an overclock stable.

But why and/hor how does increasing the voltage help stability? Perhaps
another newgroup member can elaborate?

 

[Gareth Jones]

Do it at your own risk, but I hear increasing the voltage is sometimes
needed to keep an overclock stable.

Very often the case, but you usually only have to bump it up slightly.
Remember the mobile processors are basically standard ones that have
been tested to work at stock speed but at a reduced voltage and
therefore power.
Should be no problem pushing the chip to the normal 1.65V and a bit more
(that's why they are so sought after!)

But why and/hor how does increasing the voltage help stability? Perhaps
another newgroup member can elaborate?

The only reason they keep decreasing the voltage is to try and reduce
the power consumption and heat. This makes the data signals in effect
weaker (as they are smaller) and more susceptible to electrical noise,
capacitance effects etc. Remember that standard TTL and the first
pentiums were 5v chips!
Bumping the voltage back up can help 'strengthen' these data signals. If
you have adequate cooling, there isn't usually a problem. (I say
adequate as most CPU/Heatsink designs these days don't have a huge
degree of design latitude, hence you can certainly fry them if they get
overvolted and run too hot)




Just a thought, as someone else has mentioned, you may be stressing your
RAM here, but also it could be the AGP or PCI buss. I'd try the voltage
at something like 1.7v but @200FSB - your AGP/PC should be 66/33 then.
If it works, start bumping the voltage down until you get probs.
You do know there is a utility to do this in real time from windows for
the mobile chips? I've done it on my machine and it works fine.


--
__________________________________________________
Personal email for Gareth Jones can be sent to:
'usenet4gareth' followed by an at symbol
followed by 'uk2' followed by a dot
followed by 'net'
__________________________________________________

 

[Ben Pope]

Do it at your own risk, but I hear increasing the voltage is sometimes
needed to keep an overclock stable.

But why and/hor how does increasing the voltage help stability? Perhaps
another newgroup member can elaborate?

The following assumes virtually no knowledge, for completeness

Your whole computer basically works on switches. The "switches" in the
processor are transistors. Transistors basically work by sticking a voltage
on one terminal, which allows current to flow across the other two
terminals. Since the transistor has a capacitance, it takes a finite amount
of time to charge the gate before the voltage will appear, switching the
"switches". This charge time is effectively the rate at which you can clock
the processor, and thus limits the maximum frequency. If you run it too
fast, some of the transistors won't switch in time and the wrong result will
occur. When you up the voltage, the gates charge quicker and thus the
switching frequency can be increased. Obviously more voltage means more
charge, more current and thus more heat dissipation.

Roughly.

Ben

 

[Guest]

is it the processor or your ram?...I have 2 setups with Barton 2500's..one
locked and 1 unlocked..both can hit 200 FSB in my A7N8X-Deluxe rev2..I am
running OCZ ram. I currently have the unlocked one in this machine..but have
tested the locked one with no prob's. Did you up the voltage to the
processor at all?

It's the CPU - I also have 2 systems with this mobo. The OEM
2500 barton runs fine at 200FSB on both PCs, while the retail
one fails above 185 on both - I think I was just unlucky with that
one. I *think* I tried upping vcore, but may not have as I initially
had high CPU temps until I bought a better HSF (Coolermaster
Aero7) than the retail AMD one, so thanks for mentioning that -
I'll give it a go (will save UK50 if it works!
Cheers,

 

[OZoNE]

The following assumes virtually no knowledge, for completeness

Your whole computer basically works on switches. The "switches" in the
processor are transistors. Transistors basically work by sticking a voltage
on one terminal, which allows current to flow across the other two
terminals. Since the transistor has a capacitance, it takes a finite amount
of time to charge the gate before the voltage will appear, switching the
"switches". This charge time is effectively the rate at which you can clock
the processor, and thus limits the maximum frequency. If you run it too
fast, some of the transistors won't switch in time and the wrong result will
occur. When you up the voltage, the gates charge quicker and thus the
switching frequency can be increased. Obviously more voltage means more
charge, more current and thus more heat dissipation.

Roughly.

Ben

Ok Ben you just scared the bejesus outa me ... Makes to much sense.
Anyways back to subject.
I have a Mobile 2500+ @ 1.65Vcore and running at 2300Mhz (11.5x200) with OCZ
ram (6-3-3-2) On a Asus A7N8X-Non Deluxe rev 2.0.
Using a Thermaltek Volcano 9+ at 3500RPM (jet motor at full RPM)
My former 2100+ took 1.70Vcor to acheive 2200Mhz with same setup.
So yes a Vcore increase might help with cpu you have now.
OZoNE