480 antec psu not enough !!!

[pelikan]

Originally posted by: Soulkeeper
i just tryed the 3.3v mod with no success
maybe i am doing it wrong
but those guys were using 10k and 20k variable resistors
i put up to 45k and my voltage still didn't shift from 3.23
this blows

i'm gonna go see if i can find the thread that has another method for doing this
bbl

Are you sure you don't need less resistance to raise the voltage?

 

[Soulkeeper]

i tryed between 0 and 45kohms
i even completely cut the wire
nothing i do with the wire has an effect on voltages
not even .001

right now my 3.3 rail is at 3.2 and i'm only at 114fsb 6:5 mem ratio
if i raise my vdimm higher it will prob drop more

i guess my antec true blue 480 is different than the true power 430 and 550
this blows

 

[Soulkeeper]

at 2.98 vdimm the 3.3 rail drops to 3.18 when i run sisoft mem bench

 

[SinfulWeeper]

Ahh... while people rave for Antec's. I personally like Vantecs or Raidmax. They do not put emphisis on that stupid 5 volt rail. But more so on the 12 volt rail .
Do not get me wrong though, Antec does put out a good PSU. Just not good enough for my own PC's.

 

[Soulkeeper]

yeah i can't get my system stable over 114fsb 1:1 with mem
just upping the vdimm to 2.8 or 2.9 makes my system unstable
i'm gonna get a job at the home depot and buy me a pc power and cooling psu

 

[pelikan]

I think its definately not your psu that's holding you back. It could be your mb. Not all of them clock really high.

 

[drewdogg808]

i've also read that running 2x512mb in dual channel will not always overclock that high because it stresses the memory controller in nf2 boards. i had 2x512 in dual channel and it would run prime all day but as soon as i run 3dmark or some memory intensive game, it would crash. running in single channel allowed much higher overclocking with prime/3dmark stablity.

 

[MDE]

Your motherboard, CPU, or RAM are holding you back. Back off the RAM timings and see what happens. You sound like you just have bad luck. Goes to show that overclocking is the luck of the draw. I'm running a 2500@ 446FSB with an NSpire (never heard of them either) 420W PSU and cheapo Samsung DDR400. Your PSU is the last thing I'd look at.

 

[Soulkeeper]

yeah i've backed off the ram timings to 3-4-4-11 and upped voltages as follows:
vdimm 3.15
vdd 2.49
vcore 1.825

also i have an slk-800 on my cpu
custom copper heatsinks on both the nb and sb with 40mm 5k rpm fans
and i have mounted an 80mm fan .5" above my memory shooting right at it

this did not let me hold 220 fully stable in 3d games even with the cpu clocked as low as i could get it (12v rail load) and the vid card underclocked as low as i could get it (5v rail load) as well

i've already purchased and returned two nf2 boards that wouldn't even do 200fsb stable
this one does do 214 stable in everything, but the 3.3v rail drops to 3.18 and the system seems to crash when it gets any lower
i've also noticed that the more i increase the vdimm the lower the 3.3v rail goes



i'm gonna buy me a 600watt pc power and cooling psu cause this is pissing me off, i need adjustable pots and the mods i've seen on the internet for my antec don't have any effect


do you guys have a better idea ??

 

[Technonut]

The Antec True Control 550 is a great PSU, and less expensive than the PCP&Cooling. I have a PCP&Cooling 425W TurboCool, and I prefer the TrueControl 550..

 

[render]

you need a multimeter to check actual voltages of your ralis.

 

[Soulkeeper]

i have one
yeah the voltage meter reads them as rigth where they should be
but damnit i would still like to crank them up a bit

ok maybe i'm just in denial that my memory/motherboard can't handle anything near 220 speeds.........

i read all my voltages from my motherboard/psu with a multimeter manually btw

so i guess i lyed to everyone......


i'm gonna go head butt the wall cause there is no way i can get 220 or higher fsb fully stable




thanks for the help everyone
i guess i'm gonna stop posting for awhile cause there's prob no point in reading my junk anymore for now

if i fix things i'll come back and let you all know


btw i've voided the warranty on every component i've ever owned....

 

[Soulkeeper]

ok just ignore my last post
here's some benchmarks for you all at various mostly unstable speeds which resulted in BSOD, lock ups, spontaneous reboots, and the such:

225x8.0 1.78v 1804mhz
2-2-2-6 2.98v 1:1
vdd 1.86v
sisoft
mem 3344/3017
cpu arith 6764/2711
cpu mult 9703/10656
26c 38.5c

221x9.5 1.68v 2099ghz
2-2-2-6 2.8v 1:1
vdd 1.87v
sisoft
mem 3275/3045
cpu arith 7871/3160
cpu mult 11311/12423
28c 39c

221x10 1.73v 2210mhz
2-2-2-6 2.9v 1:1
vdd 1.97v
sisoft
mem
N/A

221x10 1.81v 2210mhz
2-2-2-6 2.91v 6:5
vdd 1.97v
sisoft
mem 2607/2439
cpu arith 8265/3314
cpu mult 11895/13074
28c 40.5c


215x10.5 1.73v 2260mhz
2-2-2-6 2.65v 5:4
vdd 1.83v
sisoft
mem 2666/2495
cpu arith 8505/3419
cpu mult 12158/13372
28c 40.5c


214x10.5 1.74v 2255mhz
2-2-2-6 2.65v 6:5
vdd 1.80v
sisoft
mem 2675/2509
cpu arith 8434/3389
cpu mult 12141/13325
29c 41.5c
3dmark2k3 1789 300/621

cpu interface enabled
**2794/2609
**8412/3357
**12109/13320
fsb/agp spread spectrum
**2789/2611
**8403/3418
**12146/13321
2-2-2-5 timings
**2787/2611
**8441/3376
**12121/13311
2-2-2-7 timings
**2782/2603
**8438/3389
**12122/13318
enhance pci performance
**2794/2609
**8448/3388
**12147/13319

220x10 1.73v 2200mhz
2.5-3-3-8 3.03v 1:1
vdd 2.04v
sisoft
mem 3328/3097
cpu arith 8264/3314
cpu mult 11859/13022
27c 39.5c

200x11 1.71v 2200mhz
2-2-2-6 2.8v 1:1
vdd 1.71v
sisoft
mem 2900/2730
cpu arith 8263/3314
cpu mult 11831/13009
29c 41c


202x10 1.65v 2020mhz
2-2-2-6 2.97v 1:1
vdd 1.83v
sisoft
mem 2921/2730
cpu arith 7575/3047
cpu mult 10894/11974
26c 39.5c




none of these are in any particular order either i opened the file and just through them in there a few times and forgot the put them in order of attempt

and the cpu speeds can mostly be ignored cause i was only trying for hi fsb/mem speeds and cared very little about the cpu speed, although i did only use cpu speeds that were lower than my highest stable speeds of around 2.3ghz
if there is no listed 3dmark score that means the system was not stable enough for 3dmark or i just thought the system wasn't fast enough to justify one

 

[Soulkeeper]

i had many more benchmarks but windows corrupted when i attempted 230fsb at 1:1
so i lost the text file when i had to reinstall windows
i'm too lazy to plug in my cd burner, locate and install drivers, and burn a bootable mem test cd


btw 214 was the highest fully stable fsb i've reached but for some reason it isn't stable today

214x10.5 1.74v 2255mhz
2-2-2-6 2.65v 6:5
vdd 1.80v
sisoft
mem 2675/2509
cpu arith 8434/3389
cpu mult 12141/13325
29c 41.5c
3dmark2k3 1789 300/621

cpu interface enabled
**2794/2609
**8412/3357
**12109/13320
fsb/agp spread spectrum
**2789/2611
**8403/3418
**12146/13321
2-2-2-5 timings
**2787/2611
**8441/3376
**12121/13311
2-2-2-7 timings
**2782/2603
**8438/3389
**12122/13318
enhance pci performance
**2794/2609
**8448/3388
**12147/13319


This yeilded my Highest 3dmark2k3 score yet
ok i'm going to sleep i hope these benchs help you all somehow

 

[Soulkeeper]

ok i just tryed for like 8 hours straight to get 214fsb 1:1 stable with no luck here's what i tryed


214x10 1.76v 2147mhz
2-2-2-6 2.80v 1:1
vdd 1.74v
sisoft
mem 3278/3079
cpu arith 8064/3228
cpu mult 11547/12682
27c 39.5c
3dmark2k3 (300/621)
crash program
vdd 1.80v
crash reboot
vdimm 2.97v
crash program
vdd 1.94v
crash program
vdd 1.65v
crash reboot
vcore 1.8v vdd 1.72 vdimm 2.86v
crash reboot
vcore 1.8v vdd 1.92v vdimm 2.97v
1816 marks
then rebooted when i opened explorer
rebooted ran seti to just out of curiousity and system rebooted again
vcore 1.82v vdd 1.92v vdimm 2.97v
ran seti again BSOD
vcore 1.68v vdd 1.92v vdimm 2.97v
seti + sisoft tests almost stable but rebooted
vcore 1.68v vdd 1.92v vdimm 2.86v
rebooted
vcore 1.68v vdd 1.92v vdimm 2.97v vagp 1.7v
almost then rebooted
vcore 1.68v vdd 1.92v vdimm 2.97v vagp 1.8v
rebooted
vcore 1.68v vdd 1.92v vdimm 3.03v vagp 1.6v
rebooted
vcore 1.68v vdd 1.92v vdimm 3.10v vagp 1.6v
rebooted
vcore 1.68v vdd 1.92v vdimm 3.06v vagp 1.6v
rebooted
vcore 1.74v vdd 1.92v vdimm 3.06v vagp 1.6v
3dmark crashed
vcore 1.82v vdd 1.92v vdimm 3.06v vagp 1.8v
would not load windows, rebooted
vcore 1.65v vdd 1.92v vdimm 3.06v vagp 1.5v
rebooted
vcore 1.7v vdd 1.92v vdimm 3.06v vagp 1.5v
rebooted
vcore 1.7v vdd 1.92v vdimm 3.06v vagp 1.5v with 12v cord connected
rebooted

gave up for now
booted 202x10.5 2128mhz vcore 1.7v vdd 1.77v vdimm 1.8v vagp 1.5v
hmmm that rebooted too


that psu just can't supply enough juice I guess


i could find settings in there to run sisoft benches but seti and 3dmark would crash any 214 settings i had even with the highest voltages i felt were safe to offer

one weird thing tho is the 3.3v rail is read off the psu connectors as 3.35v
but the bios, the winbond hardware monitor chip/program, and MBM 5 all list my 3.3v rail as inbetween 3.24 and 3.15v depending on what i set the other component's voltages to (like dimms)
it seems the more load i put on the system the lower all three of these report the 3.3v rail
but the psu connectors never fluctuate

this is odd

i'm at 202fsb now (mostly stable now i guess) till i figure this junk out or order myself a pc power and cooling psu

 

[pelikan]

I found the same discrepancy between software monitoring and my digital multimeter. Don't believe the software. I think if you want higher FSB you should get a new NF7 rev. 2. I got mine at Excaliberpc.

 

[Mark R]

Have you used a multimeter to measure the voltage directly? That's the only way to tell if the PSU is up to it.

If the voltages on the connectors are fine and the mobo-monitor software says voltages drop under load then it is due to inadequate power circuits in the motherboard, or a bad connection on the PSU-motherboard connectors (are the connectors hot under load?)

It looks like your system is taking a lot of power - most off the +12V line. Antec, particularly the 'true-power' line have weak (by modern standards) 12V lines. A couple of years ago, the true-power design would be ideal, but motherboard design has moved on, and it is now the 12V line that is most important. My advice to anyone wanting a good, potentially future-proof PSU is to get one with the strongest 12V line you can get.

 

[Technonut]

I found the same discrepancy between software monitoring and my digital multimeter. Don't believe the software

If you are only testing by sticking MM probes in the molex connectors, you are only seeing the PSU output voltage... As Mark R pointed out, it is what the mobo voltage regulators are doing with the juice from the PSU that is important......

 

[Soulkeeper]

every day i get a different overclock due to the changes in temperature
when it gets hot the antec psu (like others) can't supply as much juice cause it is hot
like i was stable for a whole day and then the next day out of now where BSOD

heh

yeah i am getting a pc power and cooling psu
i have limited it down to two models
they just came out with a new 510 watt psu which is like 189 bucks
also they have the crazy 600 watt sucker with like 70 amps on the 12v rail for like 369 bucks

i just don't care about the fact that they cost much more than other psu's cause they are the best out there

i e-mailed them earlier today before i went to college asking if the 600 has adjustable pots like the 510 does

i'll let you know what they say and what i get

 

[drewdogg808]

are you running 2x512 in dual channel?

 

[yak8998]

Mebbe you got a bad board or PSU?

 

[Technonut]

Here is a thread that may be of interest....

 

[pelikan]

Originally posted by: Technonut

I found the same discrepancy between software monitoring and my digital multimeter. Don't believe the software

If you are only testing by sticking MM probes in the molex connectors, you are only seeing the PSU output voltage... As Mark R pointed out, it is what the mobo voltage regulators are doing with the juice from the PSU that is important......

Yes, when I measured on my board I found that my NF7 does not undervolt, as shown in software monitoring. In fact, I found out that software monitoring is not even very useful for my board.

 

[Soulkeeper]

i've read in other forums with people that have adjustable psu voltage pots and the same board
and they adjust the pots to where in the bios/mbm5/winbond hwd it will display the 3.3v as 3.3 or more
cause mine drops to as low as 3.15v

i think it's just a matter of where they read the voltage
if you read it from the wires directly out of the psu you are not getting the same as a if you read it off the board or where the psu might be being stressed

i'm inclined to believe the readings from the software


anyways pc power and cooling hasn't e-mailed me back yet about whether their 600 watt psu has adjustable voltage pots

 

[Mark R]

i'm inclined to believe the readings from the software

You're right is some ways - the voltage given by the software is the voltage that the motherboard is getting. However, the PSU is designed to provide 3.3V at its connector - and motherboards are meant to work on this basis. If the voltage is inadequate in the motherboard, then it is a faulty or badly designed motherboard that is the problem and not the PSU.

It would be easy to make a system where the PSU senses the voltage at the core motherboard circuits, so there would not be a problem, but ATX does not allow this - so it would need a new design of motherboard and PSU to support this. This system is already used for CPU core voltage regulators - there is a specific pin on the CPU which is used for monitoring the voltage on the actual silicon - this is fed back to the voltage regulator.