Asus P5E-VM HDMI: best matx ever?

NichoTL said:
I think this is less than optimal: turn the heatsink 90 deg counterclockwise and invert the fan so that it pushes air IN.
Click to expand...

NichoTL/bing:

You might be right about inverting the fan so that it pushes fan into the HS rather than out, but I don't understand what difference does it make whether I rotate it 90 degrees one way or another. How does that make a difference?

Regardless, I'll try it and see if temperatures improve. I'd like to reduce them a little because I've noticed my system misbehaving after 24 hours of stress. So I had to increase the voltage last night to vCPU 1.46xx and we'll see what happens after another 24+ hours of stress. I also reduced the NB, FSB Term and PLL but left it at 400FSB (3.6 GHz) because I have a feeling they don't need to be that high...but if instability reoccurs I'll raise it again and leave the vCPU at 1.46xx. And the temps are actually higher than what I posted before since those temps were not under prime95. They're in the mid 70C's under prime95 so I'm not too happy about that.

Although I'd prefer to mount the fan on top of the HS rather than the way it is now, I don't really want to do that as I'd have to remove the HS and rotate it the other way around so that the fan rubbers can be pushed into the fins (otherwise it'll fall off). And since I don't want to deal with those pushpins again I doubt I'll try that.
 
yorkman said:
NichoTL/bing:

You might be right about inverting the fan so that it pushes fan into the HS rather than out, but I don't understand what difference does it make whether I rotate it 90 degrees one way or another. How does that make a difference?

Regardless, I'll try it and see if temperatures improve. I'd like to reduce them a little because I've noticed my system misbehaving after 24 hours of stress. So I had to increase the voltage last night to vCPU 1.46xx and we'll see what happens after another 24+ hours of stress. I also reduced the NB, FSB Term and PLL but left it at 400FSB (3.6 GHz) because I have a feeling they don't need to be that high...but if instability reoccurs I'll raise it again and leave the vCPU at 1.46xx. And the temps are actually higher than what I posted before since those temps were not under prime95. They're in the mid 70C's under prime95 so I'm not too happy about that.

Although I'd prefer to mount the fan on top of the HS rather than the way it is now, I don't really want to do that as I'd have to remove the HS and rotate it the other way around so that the fan rubbers can be pushed into the fins (otherwise it'll fall off). And since I don't want to deal with those pushpins again I doubt I'll try that.
Click to expand...

It's better to put it the way we suggested because the HS/Fan contributes to 'the natural flow of air' of your box. But you don't HAVE to do it.

Regarding your voltages, could you post NB,PLL,FSB?

If you're not going to rotate the heatsink 90degrees, you might as well leave everything as it is now. If you are going to do something, do as we described, IMHO.
 
NichoTL said:
It's better to put it the way we suggested because the HS/Fan contributes to 'the natural flow of air' of your box. But you don't HAVE to do it.

Regarding your voltages, could you post NB,PLL,FSB?

If you're not going to rotate the heatsink 90degrees, you might as well leave everything as it is now. If you are going to do something, do as we described, IMHO.
Click to expand...

Like I said, I'll still do it, I just don't understand how rotating helps since the fan itself already does the rotating. It still turns clockwise. Inverting I understand completely. So don't get me wrong, I'm not implying you're incorrect.

So I'll invert the fan and turn it 90 degrees counterclockwise when I get home tonight. I'll post back later with any difference that made, if any as well as the voltages.
 
Crap. I degraded my pencil mod when I was swapping out my hs/f the other day. A few questions I probably know the answers to but would like validation on.

1) I cleaned up the resistor and surrounding area with 90% isopropyl alcohol and let it dry overnight so I can start over. That's the safe thing to do and if I was careful applying the mod, nothing permanent was done. Right?

2) The only 2B pencils I can find in my area are leads for mechanical pencils. Is that ok? Actually seems more precise than wooden pencils.

3) Would carefully applying a small swatch of electrical tape over the mod have any ill effects on surround components (covering sensors, etc)?

Really bummed because I had it the mod nailed. vcore ould ride 1.344v solid load or not. It started drifting 0.01v on load which was acceptable now it's pretty much back to stock drooping all the way down below 1.2v on load and killing my 3.2Ghz OC.
 
yorkman said:
Like I said, I'll still do it, I just don't understand how rotating helps since the fan itself already does the rotating. It still turns clockwise. Inverting I understand completely. So don't get me wrong, I'm not implying you're incorrect.

So I'll invert the fan and turn it 90 degrees counterclockwise when I get home tonight. I'll post back later with any difference that made, if any as well as the voltages.
Click to expand...

When I look at your setup, the flow of air in your box is from the right (front of the case) to the left (back of the case). So the heatsink fan should be helping/using that flow. So it should move air from the right to the left. And because you already have a case fan pulling air out of the case (out of the heatsink too for that matter) then you are better off having the heatsink fan pushing air INto the heatsink.

Currently your heatsink fan is pulling air towards the bottom. Where does this air come from? Logically the top of the heatsink: but you don't have fresh air there: the closest source is the intake fan bottom right (front) of your case. The air path should/must be as simple as possible. You can't expect air to make several turns in your case. See what I mean?
 
yorkman said:
Well, here it is finally. You can see where the fan is mounted on the HS now. And it's pulling hot air away from the cooler (I put my hand right under it and it my cold hand was getting nice and warm so it's definitely working).

You can also see me temporarily using an Ultra PSU as backup for my dead OCZ 800EXS PSU. Btw, speaking with their tech support on their forums, it looks like I'll be getting my dead 800EXS replaced with a PC Power & Cooling (S75QB) Silencer 750 or PC Power & Cooling (PPCT860) Turbo Cool 860W. I told them I'd prefer the 2nd one (costs $100 more than what I paid for the 800EXS, has 7yr warranty and 200,000 MTBF, but at the price of more fan noise which I'm ok with.

Vendetta2.jpg
Click to expand...

def change it to follow the flow of your case, your sucking hot air off your vid card, or pulling air down onto it, either way your CPU fan should continue the flow of air IN from the front - across the CPU and out the back
 
Hi all,

I just got this board, and I need some help with OCing! :)

I built this in a sg02f case with e5200 on Zalman 8700 cooler, so the cooling is not all that great. Still working it out, wracking my brain for ideas (flip PSU, flip PSU fan, build air channel for PSU or HSF, etc)...

Meanwhile, I have started on OCing, and here's what I got so far:

- All voltage on auto
- Vcore = 1.4000V (1.384V in CPUZ... pencil modded)
- 288 x 12.5 = 3600MHz
- RAM at 864MHz (G.Skill 1066, 4gb)

Temps:
Idle/Load: 38/72 deg C
Mobo: 31 deg C

I can't get 3.8Ghz stable all the way up to 1.4625V, and since temps were going over the roof (75~79 deg C), I decided to wait on that for a bit.

I tried dropping the multiplier and raising the FSB, to see what I can get, with the following voltages:
- PLL: 1.64V
- FSB Term Volt: 1.46V
- NB Volt: 1.47V
- SB Volt: 1.20V
- everything else on auto

I can't post with 350MHz at all (even at ~3ghz), and 333x11.5 (3850MHz) fails prime95 right away, or won't post at all (1.4500V and 1.4625V)

Right now, I am running 333x11.0 (3663MHz), at 1.4125V (13.92V)... 1.4000V won't post at all either. Prime95 fails at 1.4125V after 19 min.

Any ideas as to why I can't seem to post with even 350MHz? Anything I can try to make the OC more stable without totally going overboard with the voltages?

If I can get cooling better, I really want to be able to hit 3.8~4.0Ghz range... but it's looking a bit grim right now :(
 
Battlestar said:
NichoTL - here's my screen cap bios. What could be wrong on here?



Click to expand...

Personally I leave CLock Over-Charging, CPU SPread Spectrum, PCIE Spread Spectrum, CPU Voltage Reference and NB Voltage Reference on Auto:
1. I don't know what they do
2. I haven't seen a difference when changing settings
3. They might come into play with the most extreme O/Cing (not our case)

Try this (it's very similar to my own O/C)
PLL 1.58
DRAM 2.1
FSB 1.32
NB 1.37

And if this doesn't work bump NB and FSB another 2 notches to see if it makes a difference.
 
NichoTL said:
Personally I leave CLock Over-Charging, CPU SPread Spectrum, PCIE Spread Spectrum, CPU Voltage Reference and NB Voltage Reference on Auto:
1. I don't know what they do
2. I haven't seen a difference when changing settings
3. They might come into play with the most extreme O/Cing (not our case)

Try this (it's very similar to my own O/C)
PLL 1.58
DRAM 2.1
FSB 1.32
NB 1.37

And if this doesn't work bump NB and FSB another 2 notches to see if it makes a difference.
Click to expand...

NichoTL - I have already tried mix and match:

PLL 1.58 - Failed to post
DRAM 2.1 - failed to post
FSB 1.32 - Failed to post
NB 1.37 - Failed to post

NB beyond 1.39v and up will cause some error that relates to my memory :(

Im now looking for the Team Xtreem PC6400 D9GMH. They have a good review and can be reach even 1066Mhz w/o any overvoltage.
 
Battlestar said:
NichoTL - I have already tried mix and match:

PLL 1.58 - Failed to post
DRAM 2.1 - failed to post
FSB 1.32 - Failed to post
NB 1.37 - Failed to post

NB beyond 1.39v and up will cause some error that relates to my memory :(

Im now looking for the Team Xtreem PC6400 D9GMH. They have a good review and can be reach even 1066Mhz w/o any overvoltage.
Click to expand...

Did you try all of these settings at the same time (PLL, DRAM, FSB, NB)? They work together. If they work...
And if that doesn't work, raise FSB and NB 2 notches ON TOP of this. And try DRAM 2.14 too.

(careful with the double-triple posting)
 
antaholics said:
I can't post with 350MHz at all (even at ~3ghz), and 333x11.5 (3850MHz) fails prime95 right away, or won't post at all (1.4500V and 1.4625V)

Right now, I am running 333x11.0 (3663MHz), at 1.4125V (13.92V)... 1.4000V won't post at all either. Prime95 fails at 1.4125V after 19 min.

Any ideas as to why I can't seem to post with even 350MHz? Anything I can try to make the OC more stable without totally going overboard with the voltages?

If I can get cooling better, I really want to be able to hit 3.8~4.0Ghz range... but it's looking a bit grim right now :(
Click to expand...

3.6 GHz is a decent OC for a E5200. Nothing to be ashamed of. I know that Tom's Hardware in their SBM had managed close to 4.0GHz a few months ago but they were using a better cooler, and processor samples vary quite a bit. There is nothing to say that you'll be able to go much hicher (if any) than that. On top of that, your temps are already quite high with your current O/C. If there is room for improvement for your CPU, I'm not sure your cooler will let you achieve it.
 
antaholics said:
Hi all,

I just got this board, and I need some help with OCing! :)

I built this in a sg02f case with e5200 on Zalman 8700 cooler, so the cooling is not all that great. Still working it out, wracking my brain for ideas (flip PSU, flip PSU fan, build air channel for PSU or HSF, etc)...

Meanwhile, I have started on OCing, and here's what I got so far:

- All voltage on auto
- Vcore = 1.4000V (1.384V in CPUZ... pencil modded)
- 288 x 12.5 = 3600MHz
- RAM at 864MHz (G.Skill 1066, 4gb)

Temps:
Idle/Load: 38/72 deg C
Mobo: 31 deg C

I can't get 3.8Ghz stable all the way up to 1.4625V, and since temps were going over the roof (75~79 deg C), I decided to wait on that for a bit.

I tried dropping the multiplier and raising the FSB, to see what I can get, with the following voltages:
- PLL: 1.64V
- FSB Term Volt: 1.46V
- NB Volt: 1.47V
- SB Volt: 1.20V
- everything else on auto

I can't post with 350MHz at all (even at ~3ghz), and 333x11.5 (3850MHz) fails prime95 right away, or won't post at all (1.4500V and 1.4625V)

Right now, I am running 333x11.0 (3663MHz), at 1.4125V (13.92V)... 1.4000V won't post at all either. Prime95 fails at 1.4125V after 19 min.

Any ideas as to why I can't seem to post with even 350MHz? Anything I can try to make the OC more stable without totally going overboard with the voltages?

If I can get cooling better, I really want to be able to hit 3.8~4.0Ghz range... but it's looking a bit grim right now :(
Click to expand...
I agree completly with NichoTL, but have you tried to lower your multiper and raise your FSB, so you don't is in a "FSB-hole".

Swed
P.S. Battlestar, both CPU & PCIE spread spectrum can safely be disebled, but leave the rest in auto.
 
Swed said:
I agree completly with NichoTL, but have you tried to lower your multiper and raise your FSB, so you don't is in a "FSB-hole".

Swed
P.S. Battlestar, both CPU & PCIE spread spectrum can safely be disebled, but leave the rest in auto.
Click to expand...

What do those 2 spectrum settings mean, by the way?
 
NichoTL said:
It's better to put it the way we suggested because the HS/Fan contributes to 'the natural flow of air' of your box. But you don't HAVE to do it.

If you're not going to rotate the heatsink 90degrees, you might as well leave everything as it is now. If you are going to do something, do as we described, IMHO.
Click to expand...

NichoTL/Mr. Guvernment:

Several curses later (thanks to the pushpins again), it turns out that blowing the air out of the HS is more efficient than in. This makes sense, at least in my case, because hot air is being pulled from the HS down and around the video card (the video card's temperature is only slightly affected which I don't care as I don't game much on this pc) and out the rear. The only adjustment I made which caused no change is the orientation of the fan. It turns out the fan is supposed to have been turned 180 degrees counterclockwise (as per the arrows hidden in the fan bracket). I also tried just 90 degrees as you suggested but again that made no difference. So I left it

I also learned something else. After installing your HS, if your thermal compound is not spread out to each corner of your cpu, I've noticed that SpeedFan reports each core's temperature by a wider spread. So for example, before my temps were: core0 60C core1 60C, core2 56C and core3 57C. Notice the last two cores are off by 3-4 degrees.

After inverting the HS back and forth a couple of times and observing this change, I now have the cores more closely matched as my temps are now: core0 62C core1 61C core2 60 core 3 59C..while I still didn't get it right, it is slightly better as the difference is now 2-3C. This is also debatable as it's possible that it's just a coincidence but that's what I noticed anyway. Of course these are under full load, just not with prime95.

With the side door open, the temps are obviously still much better:

speedfan_3.60GHz_fritz.jpg


NB,PLL,FSB?[/QUOTE]

CPU & PCIE Spread Spectrums: Disabled
Clock Overcharge: 0.80
CPU: 1.4625V
PLL: 1.58V
DRAM: 2.1V
FSB: 1.44V
NB: 1.39V
SB: 1.20V

I've had no problems with these settings yesterday so I decided to now lower them to:

CPU & PCIE Spread Spectrums: Auto
Clock Overcharge: Auto
CPU: no change
PLL: 1.54V
DRAM: 2.00 V
FSB: no change
NB: 1.37V
SB: Auto
 
NichoTL said:
What do those 2 spectrum settings mean, by the way?
Click to expand...
You could say that it's a metod to reduce electromagnetic radiation, to prevent interference with surrounding electronic.

http://en.wikipedia.org/wiki/Spread_spectrum

You ought to have it disabled because it's easyer to get the rig stabil then. And when your OC'ing you shall have it disabled.
Of course if your rig is stabil with it enebled, you can leave it.


Swed
 
Did you try all of these settings at the same time (PLL, DRAM, FSB, NB)? They work together. If they work...
And if that doesn't work, raise FSB and NB 2 notches ON TOP of this. And try DRAM 2.14 too.
Click to expand...
Both, I've tried one at a time then all settings at the same time.

P.S. Battlestar, both CPU & PCIE spread spectrum can safely be disebled, but leave the rest in auto.
Click to expand...
Ok, let me try that thing. Thanks :)
 
yorkman said:
NichoTL/Mr. Guvernment:

Several curses later (thanks to the pushpins again), it turns out that blowing the air out of the HS is more efficient than in. This makes sense, at least in my case, because hot air is being pulled from the HS down and around the video card (the video card's temperature is only slightly affected which I don't care as I don't game much on this pc) and out the rear. The only adjustment I made which caused no change is the orientation of the fan. It turns out the fan is supposed to have been turned 180 degrees counterclockwise (as per the arrows hidden in the fan bracket). I also tried just 90 degrees as you suggested but again that made no difference. So I left it
Click to expand...

Ouch. It feels like there has been misunderstanding all along: It's the heatsink that need to be rotated: rotating the fan on its axis has (obviously) no effect whatsoever. According to the picture of your setup, it should look like this:
< !
< !
(the shape of your heatsink: its back towards the rear of the case)
Then the fan has to be on the right-hand side of it (you can't attach it onto the other side) and then setup the fan (look at the blades) so that the air is pushed into.

But I'm curious: why did you have to do the pushpins again? And also what is it that you thought we were saying when we were talking about right to left air path (on your pic)?
 
NichoTL said:
Ouch. It feels like there has been misunderstanding all along: It's the heatsink that need to be rotated: rotating the fan on its axis has (obviously) no effect whatsoever. According to the picture of your setup, it should look like this:
< !
< !
(the shape of your heatsink: its back towards the rear of the case)
Then the fan has to be on the right-hand side of it (you can't attach it onto the other side) and then setup the fan (look at the blades) so that the air is pushed into.

But I'm curious: why did you have to do the pushpins again? And also what is it that you thought we were saying when we were talking about right to left air path (on your pic)?
Click to expand...

I'm not sure who's misunderstanding who but I'll try to explain again.

I had to deal with the pushpins again when I changed the orientation of the HS. If you want to rotate the HS you have to remove it first, hence the pushpins.

I'm not sure if you meant to post a picture above or not but if yes then it didn't work. So I don't understand how you want me to position the HS and fan.

If you meant that you wanted me to rotate the HS instead of the fan by 90 degrees (which you said to do) then no, I didn't try that. So, going by the last picture I posted...my tower is standing upright. So I think what you want me to try is have the HS's back (where it's not flat/even) face the back of the case so that the front of the HS is facing the front. The HS fan then goes in front of the HS, blowing air through it and out the back. The HS in my last picture would then have its long side going from top to bottom (vertical), with the fan on the right and the back of the HS on the left.

Is that how you're saying I should have it?
 
yorkman said:
I'm not sure who's misunderstanding who but I'll try to explain again.

I had to deal with the pushpins again when I changed the orientation of the HS. If you want to rotate the HS you have to remove it first, hence the pushpins.

I'm not sure if you meant to post a picture above or not but if yes then it didn't work. So I don't understand how you want me to position the HS and fan.

If you meant that you wanted me to rotate the HS instead of the fan by 90 degrees (which you said to do) then no, I didn't try that. So, going by the last picture I posted...my tower is standing upright. So I think what you want me to try is have the HS's back (where it's not flat/even) face the back of the case so that the front of the HS is facing the front. The HS fan then goes in front of the HS, blowing air through it and out the back. The HS in my last picture would then have its long side going from top to bottom (vertical), with the fan on the right and the back of the HS on the left.

Is that how you're saying I should have it?
Click to expand...


Spot on. And I was trying to mimick the shape of your HS with these simple characters: < and !. Hence the:
< !
< !
And I also meant misunderstanding, as in we misunderstood each other.
If you look at my 90degree post, I clearly say the heatsink. Not the fan. Then the fan had to be inverted (as in change from pull to push).
 
NichoTL said:
Spot on. And I was trying to mimick the shape of your HS with these simple characters: < and !. Hence the:
< !
< !
Click to expand...

Yikes. It never even occurred to me to try that, or that that's what you were trying to say. Ugh! Pushpins time again. I can hardly wait ;--[

So the voltages I posted are OK then?
 
yorkman said:
I didn't know you'd need a pic of my case to determine that. Or a pic of my temps?
Click to expand...

Nothing to do with the temps. Just wanted to check that we understood each other 100% orientation-wise.

Voltages: you have to understand that I'm only making (somewhat) educated guesses, I can't really say for sure. But 1.44 for FSB seems a little bit high and 1.39V for NB a bit low. Apart from that, it's hard to say.
 
NichoTL said:
Nothing to do with the temps. Just wanted to check that we understood each other 100% orientation-wise.

Voltages: you have to understand that I'm only making (somewhat) educated guesses, I can't really say for sure. But 1.44 for FSB seems a little bit high and 1.39V for NB a bit low. Apart from that, it's hard to say.
Click to expand...

Ok. Well, so far I'm using the pc as I normally do (24/7 full throttle) and no problems while I keep lowering the voltages. I'll continue lowering them until I have instability issues, then raise them back up a notch or two. I'd really like to lower the cpu voltage which would lower my temps but with the other voltages being higher before, I just didn't have the system stable 24/7. So I'm assuming I won't be able to lower that anymore, and that's the reason why I'm lowering the other voltages instead....though they don't seem to play a role in my temperatures :-(
 
NichoTL said:
3.6 GHz is a decent OC for a E5200. Nothing to be ashamed of. I know that Tom's Hardware in their SBM had managed close to 4.0GHz a few months ago but they were using a better cooler, and processor samples vary quite a bit. There is nothing to say that you'll be able to go much hicher (if any) than that. On top of that, your temps are already quite high with your current O/C. If there is room for improvement for your CPU, I'm not sure your cooler will let you achieve it.
Click to expand...

thanks for the tip... I did read that TH article, and that's where I got my idea from :p

either way, I'm going to do some testing later playing around with the PSU orientation to see if I can get temps down... 3.6ghz turned out to be not completely stable, so I have to bump to 1.4125V and try again.

meanwhile, how do I know if I'm in a FSB hole, and how do i determine if NB/SB/PLL voltages are high enough or too high for a given FSB?

how does FSB hole work? :confused:

thanks!
 
antaholics said:
thanks for the tip... I did read that TH article, and that's where I got my idea from :p

either way, I'm going to do some testing later playing around with the PSU orientation to see if I can get temps down... 3.6ghz turned out to be not completely stable, so I have to bump to 1.4125V and try again.

meanwhile, how do I know if I'm in a FSB hole, and how do i determine if NB/SB/PLL voltages are high enough or too high for a given FSB?

how does FSB hole work? :confused:

thanks!
Click to expand...

if the voltages are too low, then you will have instability problems one way or the other. I doubt you got into a FSB hole there. But anyway, I think for you now what's needed is to make better use of your RAM.
On strap 266, the highest divider is 2:3 (FSB1066/DDR2 800) which would require 333 to get to 1000MHz on your RAM and 355 or so to get 1065. Too high.
On strap 200, the highest divider is 1:2 (FSB800/DDR2 800) which needs 266 to get to 1066. Too low. You are already past that frequency.
Next one down, divider 3:5 (FSB800/DDR2 667) which needs 300 to get to 1000 and 320 for 1066. I think it's what you should be aiming at.

I think you should try to get 10*300 stable first, with a lower divider even (3:4) so your RAM is at 800. Once it's stable, change the divider to get your RAM at 1000 (nudge the VRAM voltage up until it works). Then finally push the CPU multiplier to 12 or even 12.5. If you get 12.5x300=3.75 working with your RAM at 1000, you should be pretty happy. You can also try with 320x11.5-12-12.5 and see what you get.
 
NichoTL said:
if the voltages are too low, then you will have instability problems one way or the other. I doubt you got into a FSB hole there. But anyway, I think for you now what's needed is to make better use of your RAM.
On strap 266, the highest divider is 2:3 (FSB1066/DDR2 800) which would require 333 to get to 1000MHz on your RAM and 355 or so to get 1065. Too high.
On strap 200, the highest divider is 1:2 (FSB800/DDR2 800) which needs 266 to get to 1066. Too low. You are already past that frequency.
Next one down, divider 3:5 (FSB800/DDR2 667) which needs 300 to get to 1000 and 320 for 1066. I think it's what you should be aiming at.

I think you should try to get 10*300 stable first, with a lower divider even (3:4) so your RAM is at 800. Once it's stable, change the divider to get your RAM at 1000 (nudge the VRAM voltage up until it works). Then finally push the CPU multiplier to 12 or even 12.5. If you get 12.5x300=3.75 working with your RAM at 1000, you should be pretty happy. You can also try with 320x11.5-12-12.5 and see what you get.
Click to expand...

What I meant about the voltages for mobo is... how do I know if I'm setting them too high?
Conversely, if the OC is unstable, how do I know if the auto mobo values are too low, or my Vcore isn't high enough? :confused:

I tried 300 x 12.5 = 3750MHz, but I couldn't get it stable up to 1.4625V (1.440V Vdroop)... 1.4500V failed in 23 minutes, and 1.4625V failed in 3 minutes :confused: Is that normal?? or is my chip dying already :(

Right now I'm settling for something a bit more modest to get a stable overclock first, as you recommended... 320 x 11.0 = 3520MHz, at 1.4000V (1.376 Vdroop), with RAM at 800MHz. If I can get this stable for 12 hours, I'll move up the ram or multiplier and go from there.

At the moment the CPU temp is at 62 deg, with the PSU sitting outside the case, so heat shouldn't be a problem (until I find a good permanent solution for airflow)
 
NichoTL said:
Nothing to do with the temps. Just wanted to check that we understood each other 100% orientation-wise.

Voltages: you have to understand that I'm only making (somewhat) educated guesses, I can't really say for sure. But 1.44 for FSB seems a little bit high and 1.39V for NB a bit low. Apart from that, it's hard to say.
Click to expand...

Well, I tried just about every combination possible now and the way I mounted the HS and fan originally is still the best. I mounted it the way you said and I don't know why but the it seems the fan air blowing through the HS doesn't go to the rear but down and to the rear, and missing my rear exhaust fan. You can verify this by looking at my gpu temp. which is much lower than with my original mounting. I tried rotating the fan 90 degrees a few times and it's always the same (blowing through hs and down).

Left is HS is mounted vertically and fan blowing air through it. Right is HS is horizontal and fan taking air out.
hs_vertical_fan2rear.jpg
speedfan_3.60GHz_fritz.jpg


So, for whatever reason, in my case at least, the most efficient cpu cooling is mounting the hs horizontally and the hs fan pulling air out from the hs. Of course the gpu takes the biggest hit but thats still only 41C and I'd rather have that higher than the cpu. Case is now closed.

Thanks for your suggestion. It made sense but it just didn't do as well.
 
It's all over for my Q6600!! *Sniff*

I've no idea what happen but you can guess. After I changed back the HS to the way it was, I couldn't POST anymore. Tried it using the Q6600 I took from the kids pc and it POSTS fine. Inspecting my cpu closely, I can see a tiny spec of "something", like a scratch, across one of the golden dots. It's so tiny and minor looking that I'm not sure this is the reason but it's the only thing I can see that's out of the ordinary.

I'm overclocking the kids q6600 to 3.45 as I couldn't get 3.6 with the same settings as my own q6600. Then again, I noticed the vdroop is back bigtime (I thought it may be that so I wiped it clean and reapplied the pencil mod again)...it goes from 1.400 to 1.344V when it hits 100% load. Not sure why the pencil mod isn't working anymore, using the same 2B pencil.

So I paid the "ultimate price" for not listening to my own advice of "if it ain't broken..."!!

The cpu is a year old as of January of this year. Think I've any chance of getting it RMA'd by Intel?
 
oh dear :eek:

i'm still always at least a little bit amazed when i plug everything in and it just works. :cool:

sorry dood. :(

i swear that's part of why i go in for the mounting bracket hardware kits. just less stressful all around. not saying it was the pushpins necessarily in your case, i just think it's a small price to pay to have a lot less hassle when pulling chips.

you didn't get thermal goop on the traces did you?

ps: don't fry the kids cpu too! :rolleyes:
 
bing said:
oh dear :eek:

i'm still always at least a little bit amazed when i plug everything in and it just works. :cool:

sorry dood. :(

i swear that's part of why i go in for the mounting bracket hardware kits. just less stressful all around. not saying it was the pushpins necessarily in your case, i just think it's a small price to pay to have a lot less hassle when pulling chips.

you didn't get thermal goop on the traces did you?

ps: don't fry the kids cpu too! :rolleyes:
Click to expand...

It wasn't the pushpins as I had no trouble this time. It even posted but then I shut it down to rotate only the fan on the HS and after I shut it down the 2nd time, that was it.

But I may be lucky afterall. I tried the cpu on the kids pc and it posted?!?! I tried it like 8 times on my board before that and it wouldn't post so I'll just leave it in my kids pc for now since I've already put the Vendetta2 on it. If I can't OC this one to 3.60 GHz then I'll try switching them back again. Gotta fix this vdroop though, holy smokes! It's major, and probably the reason why it won't OC to 3.60GHz with this cpu.

Update: Ok, vdroop fixed. Weird. Had to use a #2 pencil after trying the 2B pencil and it worked. When I tried it for the first time ever a month or so back I did it the other way around so it must be a combination of both that fixed it both times. So far stable at 3.5GHz which I couldn't do before so I should be able to get 3.6GHz on it now too.

Whew! Dodged a bullet on my cpu as it's working on my kids pc now. That wasn't fun, let me tell ya!
 
yorkman said:
Well, I tried just about every combination possible now and the way I mounted the HS and fan originally is still the best. I mounted it the way you said and I don't know why but the it seems the fan air blowing through the HS doesn't go to the rear but down and to the rear, and missing my rear exhaust fan. You can verify this by looking at my gpu temp. which is much lower than with my original mounting. I tried rotating the fan 90 degrees a few times and it's always the same (blowing through hs and down).

So, for whatever reason, in my case at least, the most efficient cpu cooling is mounting the hs horizontally and the hs fan pulling air out from the hs. Of course the gpu takes the biggest hit but thats still only 41C and I'd rather have that higher than the cpu. Case is now closed.

Thanks for your suggestion. It made sense but it just didn't do as well.
Click to expand...

I'm sorry but this just doesn't make sense. Ask anyone in the forums. And I mean really EVERYONE, and they'll probably say the same as me. There's something else that I/You/We have missed. Clearly. But it's your computer and you have decided now, so let's put it behind us.
 
NichoTL said:
I'm sorry but this just doesn't make sense. Ask anyone in the forums. And I mean really EVERYONE, and they'll probably say the same as me. There's something else that I/You/We have missed. Clearly. But it's your computer and you have decided now, so let's put it behind us.
Click to expand...

agreed 100%. here's a good shot that illustrates the concept.

3480inside.jpg
 
<< [exhaust fan ] << [hs ] << [hs fan] << [ front fan (optional)] << cool fresh air
 
antaholics said:
What I meant about the voltages for mobo is... how do I know if I'm setting them too high?
Conversely, if the OC is unstable, how do I know if the auto mobo values are too low, or my Vcore isn't high enough? :confused:

I tried 300 x 12.5 = 3750MHz, but I couldn't get it stable up to 1.4625V (1.440V Vdroop)... 1.4500V failed in 23 minutes, and 1.4625V failed in 3 minutes :confused: Is that normal?? or is my chip dying already :(

Right now I'm settling for something a bit more modest to get a stable overclock first, as you recommended... 320 x 11.0 = 3520MHz, at 1.4000V (1.376 Vdroop), with RAM at 800MHz. If I can get this stable for 12 hours, I'll move up the ram or multiplier and go from there.

At the moment the CPU temp is at 62 deg, with the PSU sitting outside the case, so heat shouldn't be a problem (until I find a good permanent solution for airflow)
Click to expand...

I got 8 hours P95 stable for 320 x 11.0 (3520MHz) at 1.4000V.... but when I up the RAM to 1067MHz, the system won't post at all (tried 3 times). It would post at 961MHz RAM, but I didn't test stability for that, as it is not ideal.

That confuses me, because the G.Skill ram should be very high quality and definitely should survive a 1MHz overclock. :confused:

Now testing 320 x 11.5 (3680MHz)!
 
The problem is that BIOS doesn't allow for looser than I think it's 42 or 43 TRFC timings, and our GSkill sticks require at least 45. I've emailed Asus to release a BIOS that allows for higher timings without success. I have to underclock my RAM because of this.
 
You must log in or register to reply here.
Back
Top