Voltage Question

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Hi,

I have a question.
Motherboard is A7N8X-E Deluxe, RAM is 2x512PC3200 Samsung

My new CPU, unlocked XP3000+ (AQXCA 0410), runs many 6h prime95
sessions @2200Mhz (11x200) 1,675v without a problem, but if I try say
2250Mhz(11,5*195) even with 1,850v prime95 reports errors, or crash in
the minute.
2220Mhz or 2214Mhz crash in the 15 minutes.
And for what I know about 'oc, the limit is not as brutal as what I see
with this one

I started to look more closer of the voltages an log in MMB5.
It reports for my 5v a min value of 4,81v.

Okay, MMB say.... Where is my multimeter ?
Put a multimeter on the +5v direct on the motherboard connector (the
grey one)
I run OCCT @2200Mhz (a quick stability test for 'oc).
It monitors all sensors for 1mn at idle, then makes a 25mn burning,
and then monitors all sensors at idle again for 4mn. After that it
makes graphs.
I notice a great loss on the 5v during the burning time to 4,82 (It's
MMB which collects data for OCCT).
While the test, I had some looks at the multimeter. It starts at 4,93
and drop to 4,88 for all the burn time, the goes up to 4,92.
Here are the MMB5 results
(nb : shown 4,82v on the graph when was 4,88v on the multimeter)

http://perso.wanadoo.fr/danoba/OCCT/volt5.gif

the others are good
http://perso.wanadoo.fr/danoba/OCCT/volt12.gif
http://perso.wanadoo.fr/danoba/OCCT/volt3.gif
http://perso.wanadoo.fr/danoba/OCCT/voltcore.gif
http://perso.wanadoo.fr/danoba/OCCT/cpugraph.gif

For sure, my PSU (cheap 460W) is'nt stable enought.
But while in burning, I'd take a look at the others 5v (molex, others
on the main connector) with the multimeter they all stay at 5,03v. I
assume that these ones are not drawn from the same part of the PSU.
Could it be a good/bad idea to take a +5v on a free molex and "Y" it to
the gray pin of the mainboard connector in order to get some bonus Amps
?

Some other results with the multimeter

@2200 Mhz Idle = 4,91v
Running prime95 = 4,86v

Let's see if I lower Mhz. Going to 100x11

@1100 Mhz Idle = 4,95v
Running prime95 = 4,93v

--
Alni

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Fri, 11 Feb 2005 22:34:27 +0100, Alni wrote:

> Hi,
>
> I have a question.
> Motherboard is A7N8X-E Deluxe, RAM is 2x512PC3200 Samsung
>
> My new CPU, unlocked XP3000+ (AQXCA 0410), runs many 6h prime95 sessions
> @2200Mhz (11x200) 1,675v without a problem, but if I try say
> 2250Mhz(11,5*195) even with 1,850v prime95 reports errors, or crash in
> the minute.
> 2220Mhz or 2214Mhz crash in the 15 minutes. And for what I know about
> 'oc, the limit is not as brutal as what I see with this one
>
> I started to look more closer of the voltages an log in MMB5. It reports
> for my 5v a min value of 4,81v.
[snip]

> For sure, my PSU (cheap 460W) is'nt stable enought.

How can you know that? Before you waste your money on a ?better power
supply I would say you need to look at your memory first. My bet is that
your memory is flaking out - NOT your power supply. 4.81volts is within
specs. Try upping the Vdimm voltage to your memory modules first.

Larry Gagnon, A+ certified tech.

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Fri, 11 Feb 2005 22:34:27 +0100, Alni wrote:

> Motherboard is A7N8X-E Deluxe, RAM is 2x512PC3200 Samsung
>
> My new CPU, unlocked XP3000+ (AQXCA 0410), runs many 6h prime95 sessions
> @2200Mhz (11x200) 1,675v without a problem, but if I try say
> 2250Mhz(11,5*195) even with 1,850v prime95 reports errors, or crash in the
> minute.
> 2220Mhz or 2214Mhz crash in the 15 minutes. And for what I know about 'oc,
> the limit is not as brutal as what I see with this one
>
When you take the FSB off of a 33Mhz boundary it affects the ram bus. an
example, if you have both ram and FSB set to 200, bith run at 200. Set fsb
to 195 and leave the ram at 200, and now the ram is tryong to run 200+ the
difference between 166and 195, for a total ram bus speed 229. If it works
fine at 11*200, leave the FSB at 200 and raise multiplier to 11.5 for next
step and increase vcore to whatever it take s to get it stable. Changing
the multiplier doesn't affect any of the buses..

> I started to look more closer of the voltages an log in MMB5. It reports
> for my 5v a min value of 4,81v.
>
Does your PSU have a label on that list the different amps supplied for
each rail? A 460W PSU should have plenty on each rail. And 4.81v is well
within specs.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
My server http://wesnewell.no-ip.com/cpu.htm
Verizon server http://mysite.verizon.net/res0exft/cpu.htm

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Bonjour,

Larry Gagnon a présenté l'énoncé suivant :

>
>> For sure, my PSU (cheap 460W) is'nt stable enought.
>
> How can you know that? Before you waste your money on a ?better power
> supply I would say you need to look at your memory first. My bet is that
> your memory is flaking out

My sticks are good at 11*200 (prime95 6H), why should them be bad at
something like 172*13.(tried) when they are also good in prime95 at
170*13.
I always run the Asus at a ratio RAM/FSB=100%, so at FSB172 the RAM is
also 172.

That's why I've look at the PSU.

--
Alni

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Bonjour,

Wes Newell a émis l'idée suivante :
> On Fri, 11 Feb 2005 22:34:27 +0100, Alni wrote:
>
>> Motherboard is A7N8X-E Deluxe, RAM is 2x512PC3200 Samsung
>>
>> My new CPU, unlocked XP3000+ (AQXCA 0410), runs many 6h prime95 sessions
>> @2200Mhz (11x200) 1,675v without a problem, but if I try say
>> 2250Mhz(11,5*195) even with 1,850v prime95 reports errors, or crash in the
>> minute.
>> 2220Mhz or 2214Mhz crash in the 15 minutes. And for what I know about 'oc,
>> the limit is not as brutal as what I see with this one
>>
> When you take the FSB off of a 33Mhz boundary it affects the ram bus. an
> example, if you have both ram and FSB set to 200, bith run at 200. Set fsb
> to 195 and leave the ram at 200, and now the ram is tryong to run 200+ the
> difference between 166and 195, for a total ram bus speed 229. If it works
> fine at 11*200, leave the FSB at 200 and raise multiplier to 11.5 for next
> step and increase vcore to whatever it take s to get it stable. Changing
> the multiplier doesn't affect any of the buses..

I always run FSB/RAM sync. Tried 11,5*200 @1,750. Boot, but windows
crashed.
Ok @1,800v but Prime95 reports error in the minute T°=56° (+3° from
2200)
The multimeter measure was about 4,82v. MMB5 reports 4,76v

> Does your PSU have a label on that list the different amps supplied for
> each rail?

Yes.
DC 12v 17A
DC 5v 18A, not spécified on which rails.
DC 3.3 16A
DC 5VSB 2A
DC -5v 0,5A
DC -12v 0,5A

> A 460W PSU should have plenty on each rail. And 4.81v is well
> within specs.

OK

--
Alni

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Alni wrote:
> Bonjour,
>
> Wes Newell a émis l'idée suivante :
>
>> On Fri, 11 Feb 2005 22:34:27 +0100, Alni wrote:
>>
>>> Motherboard is A7N8X-E Deluxe, RAM is 2x512PC3200 Samsung
>>>
>>> My new CPU, unlocked XP3000+ (AQXCA 0410), runs many 6h prime95 sessions
>>> @2200Mhz (11x200) 1,675v without a problem, but if I try say
>>> 2250Mhz(11,5*195) even with 1,850v prime95 reports errors, or crash
>>> in the
>>> minute.
>>> 2220Mhz or 2214Mhz crash in the 15 minutes. And for what I know about
>>> 'oc,
>>> the limit is not as brutal as what I see with this one
>>>
>> When you take the FSB off of a 33Mhz boundary it affects the ram bus. an
>> example, if you have both ram and FSB set to 200, bith run at 200. Set
>> fsb
>> to 195 and leave the ram at 200, and now the ram is tryong to run 200+
>> the
>> difference between 166and 195, for a total ram bus speed 229. If it works
>> fine at 11*200, leave the FSB at 200 and raise multiplier to 11.5 for
>> next
>> step and increase vcore to whatever it take s to get it stable. Changing
>> the multiplier doesn't affect any of the buses..
>
>
> I always run FSB/RAM sync. Tried 11,5*200 @1,750. Boot, but windows
> crashed.
> Ok @1,800v but Prime95 reports error in the minute T°=56° (+3° from 2200)
> The multimeter measure was about 4,82v. MMB5 reports 4,76v
>
>> Does your PSU have a label on that list the different amps supplied for
>> each rail?
>
>
> Yes.
> DC 12v 17A
> DC 5v 18A, not spécified on which rails.
> DC 3.3 16A
> DC 5VSB 2A
> DC -5v 0,5A
> DC -12v 0,5A
>
>> A 460W PSU should have plenty on each rail. And 4.81v is well
>> within specs.
>
>
> OK
>

I believe you are correct in suspecting the psu. My no-name 300w psu has
30A on the 5v output (max or nominal? Who knows?) Also, 20A on 3.3v...

AFAIK, the 5v and 3.3v are VERY important to cpu and memory powering and
stability, 12v is for your periphs (p4 psu has separate 12v rail for
cpu). I would make a humble recommendation to change your psu. Those
values don't look good for an xp 3000+, hell, mine wasn't good enough
for an xp 2000+, nor the sempron 2400+ I run now, but money's just not
growing on my tree! lol

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Sat, 12 Feb 2005 02:32:56 +0100, Alni wrote:

> Wes Newell a émis l'idée suivante :
>> Does your PSU have a label on that list the different amps supplied for
>> each rail?
>
> Yes.
> DC 12v 17A
> DC 5v 18A, not spécified on which rails. DC 3.3 16A
> DC 5VSB 2A
> DC -5v 0,5A
> DC -12v 0,5A
>
>> A 460W PSU should have plenty on each rail. And 4.81v is well within
>> specs.
>
> OK

Well, not this 460W PSU. If the above are the ratings it looks more like
a 250W PSU. Just for reference a 460W PSu should have at least 50A between
the +3.3 and 5V rails and at least 20A on the +12v rail. Here's the specs
on a $9US 300W PSU and it's a lot higher than your so called 460W.

DC Output : +5V 30A +/-5%, +3.3V 24A +/-4%, +12V 15A +/-5%,-12V 0.5A
+/-10%, -5V 0.5A +/-10%, , +5Vsb 2.0A +/-5%. Input Voltage : 110V/220V AC.

And here's an $11US 450W PSU;

DC Output : +5V 30A +/-5%, +3.3V 28A +/-4%, +12V 25A +/-5%,-12V 0.8A
+/-10%, -5V 0.3A +/-10%, , +5Vsb 2.0A +/-5%. With exclusive DC connector
for the P4 CPU. Input Voltage : 110V/220V AC.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
My server http://wesnewell.no-ip.com/cpu.htm
Verizon server http://mysite.verizon.net/res0exft/cpu.htm

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Sat, 12 Feb 2005 07:29:42 GMT, Wes Newell
<w.newell@TAKEOUTverizon.net> wrote:

>On Sat, 12 Feb 2005 02:32:56 +0100, Alni wrote:

>> Wes Newell a émis l'idée suivante :
>>> Does your PSU have a label on that list the different amps supplied for
>>> each rail?

>> Yes.
>> DC 12v 17A
>> DC 5v 18A, not spécified on which rails. DC 3.3 16A
>> DC 5VSB 2A
>> DC -5v 0,5A
>> DC -12v 0,5A

>>> A 460W PSU should have plenty on each rail. And 4.81v is well within
>>> specs.

>> OK

>Well, not this 460W PSU. If the above are the ratings it looks more like
>a 250W PSU. Just for reference a 460W PSu should have at least 50A between
>the +3.3 and 5V rails and at least 20A on the +12v rail. Here's the specs
>on a $9US 300W PSU and it's a lot higher than your so called 460W.

>DC Output : +5V 30A +/-5%, +3.3V 24A +/-4%, +12V 15A +/-5%,-12V 0.5A
>+/-10%, -5V 0.5A +/-10%, , +5Vsb 2.0A +/-5%. Input Voltage : 110V/220V AC.

>And here's an $11US 450W PSU;
>
>DC Output : +5V 30A +/-5%, +3.3V 28A +/-4%, +12V 25A +/-5%,-12V 0.8A
>+/-10%, -5V 0.3A +/-10%, , +5Vsb 2.0A +/-5%. With exclusive DC connector
>for the P4 CPU. Input Voltage : 110V/220V AC.



If I recall correctly, this MOBO is well known for it's shite capacity
to overclock it's flaky FSB, even with the best memory.

Ask around for Ben Pope.

http://www.ben.pope.name/a7n8x_faq.html

¸ô¶ó

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Ben was on the ATi group this AM

¸ô¶ó

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Bonjour,

borolad@myowseintheboro.org a écrit :
> On Sat, 12 Feb 2005 07:29:42 GMT, Wes Newell

> If I recall correctly, this MOBO is well known for it's shite capacity
> to overclock it's flaky FSB, even with the best memory.

May be, but I dont want to overclock it's FSB.
I know it's run fine at FSB 200,
I just want to run something like 12x190 and 12x200 or even 13x180

--
Alni

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

Bonjour,

Wes Newell avait prétendu :
>>
>>> A 460W PSU should have plenty on each rail. And 4.81v is well within
>>> specs.

4,81 OK, but 4.81 is the value at 2200Mhz, at 2300Mhz MMB5 has measured
4,75

> Well, not this 460W PSU. If the above are the ratings it looks more like
> a 250W PSU.

Okay, should I understand that you agree this PSU should be changed ?

--
Alni

 

[Guest]

Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Sat, 12 Feb 2005 11:55:54 +0100, Alni wrote:

> Bonjour,
>
> Wes Newell avait prétendu :
>>>
>>>> A 460W PSU should have plenty on each rail. And 4.81v is well within
>>>> specs.
>
> 4,81 OK, but 4.81 is the value at 2200Mhz, at 2300Mhz MMB5 has measured
> 4,75
>
>> Well, not this 460W PSU. If the above are the ratings it looks more like
>> a 250W PSU.
>
> Okay, should I understand that you agree this PSU should be changed ?

I would, The unit isn't a 460W unit. And while it has a somewhat decent
+12v rail, it leaves the +3.3 and +5v rails weak. I estimate this PSU to
be about 350W minus any combined limit of the +3.3v and +5v rails.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
My server http://wesnewell.no-ip.com/cpu.htm
Verizon server http://mysite.verizon.net/res0exft/cpu.htm