I S*ck At Overclocking! Whats Going On!

Im about to throw the towel and leave this at 333mhz (which I hate because I want 400mhz so bad)
I have an E6600 with an Striker extreme and Corsair TWIN2X1024C4 DDR800 and a evga geforce 7900gs ko.

Current system configuration

SUMMARY
START
CPU= 2.4ghz - 1066fsb - c2d6600
RAM= 800mhz 4-4-4-12-1t - manual timing and voltage - TWIN2X1024-6400C4 2x512mB
FSB-Memory Clock Mode= manual
SLI-Ready Memory= disabled

END
CPU= 3.6ghz - 1600 fsb - 1.5xv
RAM= 800mhz 4-4-4-12-2t - manual timings
FSB-Memory Clock Mode: manual
SLI-Ready Memory: disabled

DETAILS
BIOS= 0402
Board Revision= 1.00G
AI Tuning: Manual
SLI Broadcast Aperture: disabled
Nvidia GPU Ex: Disabled
LinkBoost: disabled
HDD Smart Monitoring: disabled
SLI-Ready Memory: disabled

FSB - Memory Clock Mode: unlinked
FSB - Memory Ratio: manual
FSB (QDR), Mhz: 1600
Actual FSB (QDR), Mhz: 1600
MEM (DDR), Mhz: 800
Actual MEM (DDR), Mhz: 800
CPU Multiplier: 9

Vcore Voltage: manual - 1.63125v (not a typo) (varies around 1.52v-1.54v in PC Probe II)
Memory Voltage: 2.1v (2.14 in PC Probe II)
1.2v HT Voltage: 1.50 (1.55v in PC Probe II)
NB Core Voltage: 1.50 (1.52v in PC Probe II)
SB Core Voltage: 1.50 (1.52v in PC Probe II)
CPU VTT Voltage: 1.45 (1.50v in PC Probe II)
DDRII Controller Ref Voltage: auto
DDRII Channel A Ref Voltage: auto
DDRII Channel B Ref Voltage: auto

tCL : 4 (manual)
tRCD: 4 (manual)
tRP : 4 (manual)
tRAS: 12 (manual)
CMD: 2t (manual)

Advance Memory Settings
tRRD: AUTO
tRC: 22
tWR: AUTO
tWTR: AUTO
tREF: AUTO
Async Latency: AUTO

All Spectrums Disabled

CPU Internal Thermal Control: Disabled
Limit CPUID MaxVal: Disabled
Enhanced C1 (C1E): Disabled
Execute Disable Bit: Enabled
Virtualization Technology: disabled
Enhanced Intel Speedstep Tech: Disabled
LDT Frequency: 5x

PCIEX16_1 Frequency (Mhz): 100
PCIEX16_2 Frequency (Mhz): 100
PCIEX16_3 Frequency (Mhz): 100
SPP<->MCP Ref Clock, Mhz: Auto

Not Stable no matter what! orthos will fail! dont know what to do! Im about to throw the towel! Whats going on! Why cannot I get 400mhz FSB! :( My E6600 is L628A if it is needed.

Try fixing your vdroop first...3.6ghz is pretty high anyways.

I'm glad you censored "suck". I would've been deeply offended. :)

1.5v on e6600 24/7 is pretty high and anyways its not because orthos fail in 2minutes that its not stable. you can fail orthos and still have a stable system, unless you always do 50things at the same time : ripping,gaming,burning,downloading...

Dont feel to bad, I suck at overclocking! lol
I'm sure someone around here will be able to help you figure out whats up.

Quote:

---

Originally Posted by qgshadow

unless you always do 50things at the same time : ripping,gaming,burning,downloading...

---

Wait? You mean not everyone does that? :confused: LoL

guess thats why the main thing I check out when I build a PC is it must be an awesome multi-tasking!

Well it looks to me like your voltages are very high. Like i've said many times before. Most E6600's I've seen only need about 1.43v to get 3.6ghz.

Also Knowing your load and ambient temps would be helpful. Use Coretemp or the newest speedfan. Maybe even give an SS of your probe.

You might also want to try to lower your NB voltage to 1.40-1.45. This is because the trick to these boards is having a COOL NB.

first of all what kind of cooling do you use and what are your temps?
secondly i'm not sure are you aware that nvidia 680i sli has been known to have holes in the frontside bus.
so higher fsb with multi 8 might give some positive results. it did in my case :D
or you could even try with multi 9 and see if that does anything for you.

Ok, ambient and room temps I dont really know.

Updated to BIOS 1004, 0402 has a huge vdroop.

Im using a Zalman 9700CNPS and my case is a Thermaltake Tsunami with a 120mmfan in the back so it really helps the zalman.

Tired of freaking failing the test I decided to change some stuff, changed multiplier to 8x, im able to boot at 450mhz now. Memory is at 900mhz stable with memtest passed. vcore raised a lot.

here it is

CPU= 3.6ghz - 1800 fsb - 1.55v
RAM= 900mhz 4-4-4-12-2t - manual timings
FSB-Memory Clock Mode: manual

DETAILS
BIOS= 1004
AI Tuning: Manual
SLI Broadcast Aperture: disabled
Nvidia GPU Ex: Disabled
LinkBoost: disabled
HDD Smart Monitoring: disabled

FSB - Memory Clock Mode: unlinked
FSB - Memory Ratio: manual
FSB (QDR), Mhz: 1800
Actual FSB (QDR), Mhz: 1800
MEM (DDR), Mhz: 900
Actual MEM (DDR), Mhz: 900
CPU Multiplier: 8

Vcore Voltage: manual - 1.63125v (not a typo) (1.58v in PC Probe II)
Memory Voltage: 2.2v (2.24 in PC Probe II)
1.2v HT Voltage: 1.55 (1.62v in PC Probe II)
NB Core Voltage: 1.50 (1.52v in PC Probe II)
SB Core Voltage: 1.50 (1.52v in PC Probe II)
CPU VTT Voltage: 1.55 (1.62v in PC Probe II)
DDRII Controller Ref Voltage: auto
DDRII Channel A Ref Voltage: auto
DDRII Channel B Ref Voltage: auto

tCL : 4 (manual)
tRCD: 4 (manual)
tRP : 4 (manual)
tRAS: 12 (manual)
CMD: 2t (manual)

Advance Memory Settings

All Auto except for TrC at 22

All spectrums Disabled

CPU Internal Thermal Control: Disabled
Limit CPUID MaxVal: Disabled
Enhanced C1 (C1E): Disabled
Execute Disable Bit: Enabled
Virtualization Technology: disabled
Enhanced Intel Speedstep Tech: Disabled
LDT Frequency: 5x

PCIEX16_1 Frequency (Mhz): 100
PCIEX16_2 Frequency (Mhz): 100
PCIEX16_3 Frequency (Mhz): 100
SPP<->MCP Ref Clock, Mhz: Auto

So far I can boot and use every-day applications. I can even boot at 1.55v but I thought adding more vcore would make me pass the test and unforunately it didnt. Im happy that I am able to boot at 450mhz 1:1. Unfortunately, Orthos will fail Blend test after 3 mins.
But I have to add something really really really important

My vCore is in 1.63125 in BIOS (1.58 in probe II) and my temps with a zalman 9700CNPS are around 39c in idle and 51 in load (while running orthos) :S

Isnt this weird? CoreTemp shows tjunction 81c, is that bad? If I put less than 1.55vcore my system wont boot. it also shows core#0 51c and core 1#50c and samething for Speedfan. While Im running orthos, speedfan goes up to 56c but varies around 52c and 56c

What can be happening? why are temps so weird? shouldnt they be at like 60c!?
Also, why is my OC so inestable? I can use the pc and everything but it wont run Orthos for more than 3 mins.

***WILL BRING A SS AS SOON AS I GET HOME***

PD: I just want to be able to have an stable system at either 400x9 or 450x8 with a low vcore. I think +1.45 is crazy. I dont care if my system cant go higehr than 475mhz or over 500mhz.

ok your temps for cpu are fine, core temp is the most accurate temp program far as i know. i have noticed that in some cases it might help if you raise pci-express Frequency to 103- 109 it might solve the stability issue.
allso it might help your systems stability if you set the SB,NB and HT voltage options to auto.

Get a multimeter and fix vdroop...it'll take like an hour.

Solarfall, I just turned my PC on and the temps of my CPU are in 38c in Probe II! What do I have? A Power-hungry but always-cool CPU?

After running orthos and failer in like 5 seconds, the temp of my CPU in Probe II is 43c and CoreTemp shows 55c :S, stills good but I think now its showing some hot temps. Gonna try to lower vcore.

Gonna do some more testing like reducint 1.2VHT and NB voltage to see if this improves even a little bit.

BTW, yes Kobalt, I will as soon as I get the multimeter.

Quote:

---

Originally Posted by slim142

Solarfall, I just turned my PC on and the temps of my CPU are in 38c! What do I have? A Power-hungry but always-cool CPU?

Gonna do some more testing like reducint 1.2VHT and NB voltage to see if this improves even a little bit.

BTW, yes Kobalt, I will as soon as I get the multimeter.

---

i'm gessing that 38c is idle temp. but yes that sure does seem like a very odd temp reading you have there. truelly wierd my chip goes as high as 70c(100% load) with 1.5ish and that is with watercooling.

btw what psu do you have?

Yes, it is idle temp. When I ran orthos it went up to 53c (Probe II) I think and CoreTemp showed 55c

I have an Ultra X-Finity 600 Watts 2G w/ AfterSpin Tech

BTW, Im back to 9x multiplier, I want 3.6Ghz so I thought "what if the board cannot handle 450mhz?" maybe is too much for it. So I came back to 400mhz x9 and here is my current configuration while I write this.

CPU= 3.6ghz - 1600 fsb - 1.57v
RAM= 800mhz 4-4-4-12-2t - manual timings
FSB-Memory Clock Mode: manual

FSB - Memory Clock Mode: linked
FSB - Memory Ratio: Syncronized
FSB (QDR), Mhz: 1600
Actual FSB (QDR), Mhz: 1600
MEM (DDR), Mhz: 800
Actual MEM (DDR), Mhz: 800
CPU Multiplier: 9

Vcore Voltage: 1.62500v (1.57v in PC Probe II)
Memory Voltage: 2.1v (2.14 in PC Probe II)
1.2v HT Voltage: 1.60 (1.65v in PC Probe II)
NB Core Voltage: 1.40 (1.42v in PC Probe II)
SB Core Voltage: 1.50 (1.52v in PC Probe II)
CPU VTT Voltage: 1.55 (1.62v in PC Probe II)

I noticed that having 1.2VHT at at least 1.5 helped a lot. Also, VTT wouldnt help at 1.50, either 1.55 or nothing.
NB Core, yes, I can boot at 1.42v which is good, less heat. I knew 1.52 was too much. Memory is back to 2.1 because is now at default speeds.

Gonna Try Orthos :)

Picture of what I have, here you can see everything you need to

http://img265.imageshack.us/img265/1...ystems2wv8.jpg

bump

First, what are your specs? Please add them to your signature so it's easier to help you.

Second, I'm confused about what you're actually trying to achieve here. Are you going for high CPU clocks or are you trying to get a higher FSB?

Also, your PSU may not be powerful enough depending upon what your graphics card setup is. When you get a multimeter, you may want to measure the 12v line vdroop that feeds directly to the vcore circuit aka the 4 pin standalone connector. I can give you more details for this when you get your DMM.

looks like the PSU is the culprit. IMO those PSUs arent very good and thats what is probably causing all that Vdroop you are getting. When you buy a PSU get one thats non-modular. Remember every extra molex on the line is going to effect the vdroop in some way, so be careful w/ that. If you search PSUs on the forums im sure you can find a list of good ones.

yep i suck too but im sure someone here can get you over the hump...

Quote:

---

Originally Posted by slim142

Picture of what I have, here you can see everything you need to

http://img265.imageshack.us/img265/1...ystems2wv8.jpg

---

I see 55 and 56 for temps in this ss. I see no cpu load. i thought u said you were getting those temps at full load.?.? I didnt think your figures were correct.

Like solarfall i get 68-70 with 1.55 vcore with goood water cooling. I've had 6600's, 3060, 3070, 2 6800's all about the same temps with that voltage under load. I really cant see any zalman keeping any c2d w/ 1.63vcore at 51 under load. I'm calling you out on that one. Get us a screen shot with orthos running small fft. you also said you were running blend test. I always start with small fft for cpu overclock to make sure the cpu is stable. Then i'll up mem and use blend. Right now your stressing both with blend. Could be your memory failing blend and not the cpu.

Quote:

---

Originally Posted by qgshadow

1.5v on e6600 24/7 is pretty high and anyways its not because orthos fail in 2minutes that its not stable. you can fail orthos and still have a stable system, unless you always do 50things at the same time : ripping,gaming,burning,downloading...

---

Really? So whats the point of Orthos then?
How about 5 seconds of orthos and it fails? :)

U kinda missed the part where he has super pi running.

SuperPi never really gets the proc 100&#37; heated up tho.

Try a small fft test to see what happens and what temps you get. The problem is you can't really tell what your REAL load temps are until orthos has been running for like 10 min at least.

Btw 2 small things i noticed.

Execute Disable Bit: Enabled
This should be disabled.

SPP<->MCP Ref Clock, Mhz: Auto
Set this to 200.

My idle temps are at 34 and my load temps go up to 60C so i doubt your temps will only hit 56 or so.

Oh yea one thing I noticed is that you have a L28A which might be the difference. Most ppl with nice overclocks all have B.

Quote:

---

Originally Posted by nugzo

I see 55 and 56 for temps in this ss. I see no cpu load. i thought u said you were getting those temps at full load.?.? I didnt think your figures were correct.

Like solarfall i get 68-70 with 1.55 vcore with goood water cooling. I've had 6600's, 3060, 3070, 2 6800's all about the same temps with that voltage under load. I really cant see any zalman keeping any c2d w/ 1.63vcore at 51 under load. I'm calling you out on that one. Get us a screen shot with orthos running small fft. you also said you were running blend test. I always start with small fft for cpu overclock to make sure the cpu is stable. Then i'll up mem and use blend. Right now your stressing both with blend. Could be your memory failing blend and not the cpu.

---

maybe his ambient is less than 20c?? i agree lower mem and see if it is mem fail vs cpu fail

Quote:

---

Originally Posted by IFMU

Wait? You mean not everyone does that? :confused: LoL

guess thats why the main thing I check out when I build a PC is it must be an awesome multi-tasking!

---

maybe he's one of those guys that let's the computer complete every task on its own hehehe before loading another one

also; why so high with cpu vtt and 1.2ht???

personally i think you are expecting a bit too much from your particular cpu.....

a lot of good overclocking chippies reach higher clocks at lesser voltage (and no i dont have one)....but that's not helping you.

if you are as cool as you say you are keep upping the cpu volts and see what happens :lol:

as far as i am concerned i am convinced that excessive heat in any component will pretty much limit your oc.

y dont you just settle for a tad under 3.6 and be happy (like me)
or else get some better cooling and run higher volts.


...but i am fully open to any suggestions; this is my first intel since my old pentium2 :hehe: and i never heard of oc back then :slapass:

mayby different bios whould do the trick but honestly 3.6ghz is great oc. from what i have heard not that meny people are not able to get past 3.6ghz, usually due to limiting mobo. are you sure that you need that much vcore (1.49v) ?

You're not the only one! in fact, I was so stupid, I bought a mobo known to not break 300mhz.

LOL that mobo sucks serial hehehe

Quote:

---

Originally Posted by Solarfall

i'm gessing that 38c is idle temp. but yes that sure does seem like a very odd temp reading you have there. truelly wierd my chip goes as high as 70c(100% load) with 1.5ish and that is with watercooling.

btw what psu do you have?

---

Same here, E6600 @ 3,6 Ghz 1:1, BIOS 1.55v, AsusProbe 1.50v, after Vdroop 1,48v

50 C idle, Orthos 64 - 68 C with Apogee and 120.3

You must live in a freezer... :D

Quote:

---

Originally Posted by Mango31

Same here, E6600 @ 3,6 Ghz 1:1, BIOS 1.55v, AsusProbe 1.50v, after Vdroop 1,48v

50 C idle, Orthos 64 - 68 C with Apogee and 120.3

You must live in a freezer... :D

---

LoL pretty close Mango31 i live in Finland :D

Hey guys!
Thanks for all the answers!

I just got the multimeter and Im gonna do the vdroop mod so I wont have constants vdroops and fail orthos making me put higher voltages.

Also, I will lower the 1.2VHT and VTT, yes I know they are high but they looked like they helped a little bit.

Also, My memory is running at 800mhz at default voltages so it cannot be causing the problem.

Ill add my specs to my signature right now... also, why would the PSU be causing the problem? it has been running fine and its 600watts. I dont think my system would be using that much power. I dont have SLI and my current VGA is a 7900 GS KO just to let you know.

What Im trying to achieve is 400mhz with 9x multiplier. I want 3.6Ghz. At 3.6Ghz Ill be happy and satisfied. I dont care if it is 450mhz x8 or 400mhz x9.

I will disable Execute disable and make SPP MCP 200mhz, if XD is enabled will that make your OC inestable? I want to have some kind of protection thats why i left it enabled.

This board has weak dividers. If I put 1600, memory has to be at 800. If I put something like 1790 and memory at 800, for some reason, memory will run at 760 and that causes the motherboard not to post. So I have to be very carefull when increasing or decreasing the FSB with the Memory.

Im 100% is the CPU, I have ran Gromacs Core and it failed in 5 seconds and the PC restarted.

---
Now I have some question regarding the Multimeter. Which option out of the 20 that it has should i use? I see that this "wheel" in the multimeter can be switched to 1k, 200, 20, 2 200m 20M and has 5 sections: 1 V (with this thing over it ñ), 1 V with a line over it, 1 A with this thing ñ, the A with a line over it and there is another section with a symbol.

Can somebody tell me what option should I use to do the vdroop mod? I really want to do this ASAP.

Here is a picture to explain better what I mean.
http://img512.imageshack.us/img512/4215/picture1al7.jpg

Thanks to everybody!

The PSU can be a problem if its drooping on the 12v rail itself. This will cause a vdroop as well because the circuit that converts the voltage into around 1.4volts is fed by the 12v line on your power supply.

You want to use the 200k range when checking the resistance because I believe the normal resistance is 129.5k ohms, but check first as it may be different on your board. If its ~130k ohms to start with, then you should decrease to around 120k ohms and go down in 10k ohms steps. I can't really predict a number because its different on every board. For me I had to go all the way down to 10k ohms whereas others needed only to go to 77k or 105k.

When measuring voltage, measure in the 2v range (for CPU voltage). If you're going to measure anything higher than 2V (like the 12v rail), then you should set it at 20V.

Also, be sure to you dont short anything out with the multimeter...you don't want to accidentally connect something with the conductive end of the multimeter thereby causing a short circuit.

Kobalt, Am I suppose to use the red or the black stick in the resistor?

It doesn't matter for resistance, but for voltage you ground the black end and use the red end. It doesn't really matter that much though...if you do it opposite you'll just get a negative voltage. I usually put the black end inside the black molex pin for convenience.

Also, I forgot...the system MUST be off and DISCONNECTED from the power before you measure resistance. (For voltage it has to be on).

OK THANKS MAN!

Gonna do this NOW!

BRB! wish me luck

Measured!!!! I hope you guys can help me more now!

Vdroop resistor shows 63.00
Vcore resistor shows 92.7

Now, I would like to know, I have more pencils now. I have 4B pencils and 6B pencils, I heard it would be better to use 6B pencils. My question is, which one do you recommend me to use?

procedures
mark the resistor 3 times then measure, write results in a paper and compare them to the original values. Turn on the system and check the vcore in BIOS and windows. If vdroop stills severe, then I should repeat procedure...

Is this right?

Thanks!

Check original vdroop first so you can establish a benchmark. You know where to measure the vcore from right?

Basically what you want to do is lower the resistance around 8k ohms each time and keep checking vdroop. I can't really tell you how many times to mark the resistor.

vdroop the max i can get it is 57
vcore max i can get it is 88 i believe.

dont know what im doing wrong :(

for some reason marking them more is making them get higher, and also, why is this getting complicated?¨:(

57 what? 88 what? mV?

What was your original idle and load volts. And what are they now and with what resistance.

57 and 88 using 200k, i suppose 57ohms and 88ohms, this would be the resistance. I havent turned on the system yet.

what do i do to know my irignal idle and load volts? also, what do i have to do to know if my 12v rail psu is causing the vdroop?

I thought you were giving me voltages not resistance. Don't worry about the vcore mod now...lets try to get the vdroop gone first. What you need to do is boot windows and when its idle (0&#37; cpu), measure the vcore voltage using your multimeter. Then run something that will make it 100% cpu usage and measure the voltage. The difference is vdroop and we want it to be close to zero. I don't know where the vcore measuring point is on the striker though.

I give up

I putted the black stick on the ground molex connector and when I putted the red stick in the vcore resistor, boom my computer shutted down. I wont do that again.
I think I wont be able to run 400mhz w/ 8x nor 9x :( Im using 7x and everything is going SUPER FINE at 1.42vcore (1.45000 in BIOS) booted and running small FFT orthos for like 7 minutes already.
If I could just bring the FSB to 430 to AT LEAST have 3.0Ghz it would be GREAT.

Why do I give up? First, DMM tells me vcore resistor is 96.6 and vdroop is 59.6, fine, so I start marking, vdroop goes down to 58 (after like 5 marks) and vcore wouldnt go down. Sometimes after giving more marks to vdroop, it would go up back to 59.6 so I got tired and I think Im just gonna run the 7x multiplier with the vdroop thing.
Im running 7x to see what is the dam problem. So far I hate both my CPU and my motherboard. MB because of vdroop and CPU because of power-hungry (which now has me thinking cuz it booted fine at 400mhz with 1.42v and is running orthos fine :S )

Thanks Kobalt anyways, I know you tried to help me, maybe in a couple of months with new BIOS releases I might be able to fix this vdroop thing.

Do you think I might have gotten a defective motherboard or something? something that I think is impossible cuz its been running fine, maybe one of those strikers with low OC capabilities.

Ummm...you dont measure vcore on the resistor. You measure it on one of the mosfets, but I don't know which one because I don't have your board and I cant google anything.

and youre probably not pressing hard enough or using a pencil with enough graphite. Yes sometimes the resistance goes up if you dont press hard enough because you take off some graphite.

And BIOS releases don't do anything to vdroop...its a hardware thing.

I think my CPU problem is a CPU strap, not a FSB strap.
Take a look at this, so far running nice orthos small FTT's

http://img142.imageshack.us/img142/1552/xsqs2.jpg

Take a look at the voltage, I know is running at lower frequency, but the FSB is the same, super lower VTT than before as well as 1.2VHT. NB core at default as well as SB core. This is better than any other multiplier.
Gonna see what can I do to try to overclock the highest I can :)

Ill try tomorrow. Gonna ask my friend if I can keep the multimeter 1 more day.

Illl try to do it harder. I noticed something...

It is very hard to get the resistance, for the vcore and vdroop resistors, I had to put the red stick in the border (where the silver metal part is) to get the measure, and even when doing that, the DMM sometimes went crazy with some numbers like 22, 40 then it would go up to 128 140 then come back to 30 and I had to find the right spot to see what my resistance was until I finally got it to 59 which was my latest number for vdroop and 96 for vcore.

the DMM wouldnt give me any number if I putted the red stick in the middle (black part) of the resistor. Maybe is normal or not, I never tried this so thats why im saying this.

Um you should probably take some alcohol and clean off you board completely then start over from the vdroop. But before you do any mods you probalby need to reinstall windows agian that looks pretty messed.

When you took the initial resistance measurements from the resistors it's already so low which means u didn't clean it off well enough.

Also make sure you try 401 fsb instead of 400 which might move up 2 the next strap.

hi Slim142

would you do me a favour and please try this:

!! put " all " voltages on auto except cpu(1,45-1,50) and ram(2,250-2,30) !!

sp-mcp 200-others 100-
" all " ram subtimings auto(trc too on auto)
go linked + sync
fsb 400x9
disable bit-on + thermal control-on(you can disable all too-doesn´t matter)
try to boot-afterwards please tell me if its stable or not -
(had similar experience with this mobo like you-my hair is grey now!)
good luck
greetings from austria

don´t do mods until you got stable-but this is only my opinion.

When putting all the voltages in AUTO except for vcore and memory, system wouldnt boot. I think lack of voltge for NB and VTT.

Sp-mcp is in 200 and pci-e in 100
linked + sync
455x7
disabled everything

Cant try the timings and subtimings BECAUSE..............................

Here is what happened

Something exploded on my PC. Yes, not a typo. All I know is I heard a clap and burning smell came from the case. I was doing the vdroop mod on my motherboard (this time I applied very hard marks of pencil 6B but after I finished, I blew some air to get rid of the graphite in the motherboard) and decided to turn on the pc, after 3 seconds, there was a sound like a clap, PAMMM! and then my room started smelling like something burned. Turned off the PC and disconnected the AC power cable ASAP. The burning smell was really (really) strong coming from the case, some from one of the heatsinks near the CPU and some from the vdroop mod. Then to make sure everything was fine, I decided to turn on the PC again after 3 mins. Everything was fine and I got to boot into windows. I went into internet, checked my email, chat a little bit for like 15 mins and everything. So everything was fine until I decided to run memtest 1.70 to test the memory timings at 4-4-4-12 and speed of 920mhz (because I wanted to make sure my timings and ram voltages were correct with no errors). When I came back after eating, the pc was off (like after 10 mins). Tried turning it on the pc again and it wouldnt turn on. Motherboard lights go on and fans as well, LCD screens turns on and the onboard leds go on too as well as the power on reset and clear cmos button, also the yellow light next to the text STRIKER EXREME. But I have noticed there is a strange sound coming from the Power Supply, like if there was a wire stuck on one of the two fans (but actually there arent any), Before the memtest test, everything was running fine so I think.....

The Power supply just went boom and I need to change it.

BTW, I just checked the 24pin connector and there are no black marks, will try to see if it is really defective with my pentium III.

Any comments?

BTW, no more pencil/v mods until I become an electrical enginner (hopefully someday)

Were you checking resistance after every stroke?

Yes, the minimum I got it was 55.4ohms in vdroop, vcore I didnt mark it that hard.

guess what! it was a BIOS Failure :S
still dont know whats gonna happen when I get to 400mhzx7 and Ill have to put 1.42v :S

Im really scare of that sound, but Im positive, Im thinking it was just some pencil graphite on the board that made the sound probably.

It anything related with vcore woulda been destroyed or burned, I wouldnt be using this computer right now.

My computer has something called Against Overclocking

It wouldnt let me boot at 400x9 not even at 1.50v

I really dont understand what Im doing wrong...

You're probably just at the edge of stability. Try running a RAM divider to see if its your RAM or CPU. 3.6GHz on air is a difficult overclock to achieve. An E6300 system I just built maxed at 3.2GHz so be happy with what you have now.

Be happy with 355mhz FSB? Are you kidding me? even the suckiest E6600 can achieve better speeds! I expected more out of my rig.

I noticed something, I can run Gromacs Core forever, but blend test P1 for only 4 hours, then it will say "windows has found an error blah blah" and orthos will close. Weird?

Yesterday before going to school, I left Memtestx86 1.70 running with RAM at default timings,speed and voltage to see if the RAM had any problems, came back from school and it passed 47 tests. So I think RAM is fine, I heard it would be worth to try to see if 5-5-5-15 performs better and gives more stability. Will try that tomorrow but seriously, I dont think RAM timings might be affecting that much.

BTW, Im trying what Kroc did, change the pipes and put some good cooling. I noticed that by lowering the NB voltage from 1.40 to 1.35, it gave me 2 extra hours of orthos so that might be a factor causing inestability too.

I thought you were closer to 400fsb because most people don't up their fsb in 50mhz jumps when overclocking :rolleyes:

There are three things that can hold you back:

1) CPU
2) FSB/motherboard
3) RAM

At this point, since you have the capability of running your RAM unlinked, I would lower your CPU multiplier to something low like 5x or 6x - use stock CPU voltage and keep the speed under 2.4GHz. Then slowly up the FSB until you find the limit. If you can get to say, 460x5 then you know it's your CPU that's limiting you.

Well, this OC problem is gettin on my nerves so Im gonna try what Koc did, he bought 2 Thermalright HR-05 heatsinks for the NB and SB and some Swiftech heatsinks for the VR and a pair of Ultra 60mm Fans for the heatsinks. I bought all this stuff 2 days ago and they are coming on monday so monday will be a busy day for me, I guess on monday I will get tired of seeing so many heatsinks around hehe.

If is the NB the problem, then Ill be really happy because that would mean 3.6Ghz! yay!

BTW, is 48celsious for the NB in idle a normal temp? cuz thats whats mine is at :S

Oh yeah, Thelectic, I cant put 5x, the board doesnt allow it. Also, at 6x, my system is very unestable. Ill try this heatsink project first then Ill come back with the results.