How the hell...need help

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Rodknock

Member
Nov 13, 2006
177
0
0
Wow, I had never even heard of realtemp. Thanks for the tip!!!

I just downloaded it and am running it now. Ok, this seems more in line with a little thing called reality. Ambient temp in here right now--it was hot as all hell today for the bay area--is probably pushing 70*F, if not a bit higher in my bedroom. Realtemp is showing Core Temperature as:

1: 33*C
2: 33*C
3: 29*C
4: 29*C

That's more like it!

So is realtemp more or less accurate than coretemp/speedfan/pcprobe/etc.? It certainly seems to be right in between coretemp's possible overestimate and speedfan's probable underestimate. I'm relieved now, as coretemp is the only one showing higher-than-normal temps.

Thanks again for telling me about that prog.
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
realtemp is pretty much the standard for Qxxx and all the 45nm right now. Was written by a guy over at XS I believe.
 

WoodButcher

Platinum Member
Mar 10, 2001
2,158
0
76
Originally posted by: Rodknock
I was in a traffic jam (2 fatalities) today ALL fracking day, literally 4-5 hours. Haven't done shite today. All for a couple of molex splitters from Fry's. No turnoffs on this highway, I literally couldn't turn around. Wonderful.
Think you had a bad day? what about the 2 idiots and those they left behind?

Originally posted by: DerwenArtos12
I just don't like bikes getting vilified when it's the idiot riders fault.

thank you.

I didn't know about realtemp, thanks again, I'll be using that from now on.
 

Rodknock

Member
Nov 13, 2006
177
0
0
Well the two idiots that crashed have no one to blame but themselves, so I don't feel bad for them. I feel bad for their families for being related to such retards, but mostly I feel bad for the thousands of people that were stuck twiddling their thumbs for the day because two people lacked all sense of responsibility.
 
T

Tim

Much thanks for the realtemp tip. If these temps are indeed more accurate than coretemp for my 8400, then I'm mucho mucho happy. At ambient of 20C, I'm floating inbetween 20 and 21C, and at load I haven't gone any higher than 27C. Wicked!
 

Rodknock

Member
Nov 13, 2006
177
0
0
I don't mean to rain on your parade, but if your CPU temp is equal to your ambient temp, and you're using conventional water or air cooling, you've either got the best setup evar, or it's reading low.
 

Rodknock

Member
Nov 13, 2006
177
0
0
On the other side of the coin now:

I'm seeing abnormally high temps for my setup. Realtemp is reporting 45*C idle temps for cores 1 and 2, and 41*C for cores 3 and 4. I've got the q6600 overclocked to 3.2ghz (400 x 8), running 1.4250v. Should idle temps be this high with that kind of (fairly) conservative overclock?

I've had trouble overclocking the q6600 before on my old air setup. The most common real OC configs are 450 x 8 or 400 x 9, but mine won't even boot with that, no matter how high I crank the voltage. I'm wondering if my CPU is just on the low end of the batch, and runs hotter/clocks slower than some. Maybe it's time for an E8400.
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
Originally posted by: theplaidfad
Much thanks for the realtemp tip. If these temps are indeed more accurate than coretemp for my 8400, then I'm mucho mucho happy. At ambient of 20C, I'm floating inbetween 20 and 21C, and at load I haven't gone any higher than 27C. Wicked!

I believe you have to set an option for the 45nm to change the value it's comparing against. Intels temperature sensors are K-type or differential sensors, to the software reading them has to be set with the proper maximum temperature to subtract the actual reading from to get the actual temperature. ridiculous isn't it?
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
Originally posted by: Rodknock
On the other side of the coin now:

I'm seeing abnormally high temps for my setup. Realtemp is reporting 45*C idle temps for cores 1 and 2, and 41*C for cores 3 and 4. I've got the q6600 overclocked to 3.2ghz (400 x 8), running 1.4250v. Should idle temps be this high with that kind of (fairly) conservative overclock?

I've had trouble overclocking the q6600 before on my old air setup. The most common real OC configs are 450 x 8 or 400 x 9, but mine won't even boot with that, no matter how high I crank the voltage. I'm wondering if my CPU is just on the low end of the batch, and runs hotter/clocks slower than some. Maybe it's time for an E8400.

Thats a fairly substantial voltage from my understanding but, shouldn't be running that warm. How are your loads? What stepping is that Q6600?
 

Rodknock

Member
Nov 13, 2006
177
0
0
I think I found out where I went wrong and why my temps are so high. I applied the thermal paste incorrectly, using a method that's worked well for me with the video card, but apparently DOESN'T work with the cpu:

Spreading AS5 very thin with a razor blade. It's worked phenomenally well with my 8800gtx. I'm getting 40*C idle temps with this card, compared to stock temps in the 70s. Full load never gets above ~50* now, it's pretty amazing.

I guess this does NOT work on the CPU, and I'm an idiot for forgetting you're supposed to do a line across the casing and lead the heatsink smush it on there evenly when you mount it. So, I fucked up the thermal application:(

Can this be the ultimate cause of my woes? I've read some reports claiming a 10-15*C difference in incorrect vs. correct paste application. I fucked up, obviously, but if I reseat it with the paste applied properly via the line method, am I really going to see much better temps?
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
I still use the spread method to this day without issue so far but, *if* there is in fact too much or not enough TIM in there then yes, it's quite possible to see a substantial drop in temps. 10*C I would believe, 15*C seems like a lot but, were it completely botched beyond all belief it could happen.
 
T

Tim

Originally posted by: DerwenArtos12
Originally posted by: theplaidfad
Much thanks for the realtemp tip. If these temps are indeed more accurate than coretemp for my 8400, then I'm mucho mucho happy. At ambient of 20C, I'm floating inbetween 20 and 21C, and at load I haven't gone any higher than 27C. Wicked!

I believe you have to set an option for the 45nm to change the value it's comparing against. Intels temperature sensors are K-type or differential sensors, to the software reading them has to be set with the proper maximum temperature to subtract the actual reading from to get the actual temperature. ridiculous isn't it?

I knew it was too good to be true :(
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
Originally posted by: theplaidfad
Originally posted by: DerwenArtos12
Originally posted by: theplaidfad
Much thanks for the realtemp tip. If these temps are indeed more accurate than coretemp for my 8400, then I'm mucho mucho happy. At ambient of 20C, I'm floating inbetween 20 and 21C, and at load I haven't gone any higher than 27C. Wicked!

I believe you have to set an option for the 45nm to change the value it's comparing against. Intels temperature sensors are K-type or differential sensors, to the software reading them has to be set with the proper maximum temperature to subtract the actual reading from to get the actual temperature. ridiculous isn't it?

I knew it was too good to be true :(

sorry, on the positive side, I think you're only going to be off 5C. I think realtemp comes set with a 110C(Q6x00) max and your chip uses 115. You'll have to read through the realtemp site though as I'm not 100% sure.
 
T

Tim

And you mean to tell me that licking my thumb and sticking it out the window isn't an accurate judge of ambient? Sheesh... what a day!
 

Rodknock

Member
Nov 13, 2006
177
0
0
Derwen:

Apparently how you apply the thermal paste is especially important when using the intel quad cores. I'm hoping once I reseat the block and apply the paste correctly, it'll account for the 5-10*C over-normal I'm seeing now.
 

DerwenArtos12

Diamond Member
Apr 7, 2003
4,278
0
0
Yeah, because quads are actually two cores spaced apart you have a much wider heat distribution area so TIM has to be even and wider than duals. There fore the vertical line instead of the dot.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,067
3,574
126
okey Rod i did a little thinkn and i need a bunch of more variables to find out whats wrong.

Did you do the simple thermometer test?

If you dont know what that means its simple. Grab a thermometer. The kind you put in your mouth when your sick. Glass mercury is the best, dont get the electrical ones.

Now place it in front of your radiator where its drawing air. Record the temperature. This is your intake ambient.

Now after about 15-20 sec, make sure you take that ambient and record it.

Now next step, while your cpu is loading drop the thermometer in there for about 30 sec. Make sure your system is on load for at least 20-30min b4 you drop the thermo in. [It can take about this long for the water temp to =]

Record the max water temp on load, Then on idle, after it comes off load, wait 10 min, let the coolant = again. Record temps here as well.

I am looking for 3 things right now.

1. Intake Temp = Ambient temp.
2. Efficency Temp. Meaning the difference between idle and load. The number should not be that great.
3. The delta between idle and load in contrast to ambient temp.


Need these to see if something is wrong.
 

WoodButcher

Platinum Member
Mar 10, 2001
2,158
0
76
Glad you dropped in Aigo, I'm stumped, there are 2 or 3 new builds around here w quads that all have high temps, 2 on water, "Rodknock" also has a gpu on the loop but two 220 rads, but an other is "Timbob" just the cpu on a Petras kit w/ a 320 here. "Dawgdoc" also has a prob here and he has a thermochill 320.
What gives w/ these quads? Three different systems and all custom loops, I just don't get it.



ACK I accidentally hit edit instead of reply. Sorry about this sticky black thing at the bottom!

Case and Cooling Moderator Aigomorla
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,067
3,574
126
Right now i have no clue.

I just finished that legos guy on that custom build.

I was giving his problem some thought, heres what i think, he might have a kink somewhere thats hidering his flow. Having bad flow + accelerators = very bad scenario.

Another theory i have is he's drawing ambients from a hot source. Ambients control coolant temps. Coolant temps control cpu temps.

Tell the other two people to do what i asked and grab me those variables. This way i can see if something is wrong with the loop or if its hardware. You'd be suprised, sometimes hardware can be bad.
 

Rodknock

Member
Nov 13, 2006
177
0
0
Woodbutcher, you're slightly mistaken about my WC setup. The CPU and GPU are on independent but identical loops.

MCP655 > MCR220 > CPU (dtek fuzion) > MicroRes is loop 1.

MCP655 > MCR220 > GPU (MCW60) > MicroRes is loop 2.


With an accepted 'normal' ambient temp of 68* F (average temperature of a house I would assume), ***what should my temp range be, idle and load?***, based on what you already know of my setup?

I know that answer depends heavily on voltages, but for the moment just give me an educated guess.

I just moved back from college, and haven't set up my main rig yet, but here's what I can tell you from memory. Current overclock was 3.0ghz at *I'm 99% sure* 1.325v, give or take a hundredth of a volt or so. The Maximus Formula bios also has a ton of other voltage options, CPU PLL was set at 1.56, NB termination voltage was 1.4. Definitely a lot more complicated than my old 680i SLI board.

At those voltages, I was getting 32-33*C idle temps on the first two cores, and 28-29*C temps on the second two. In prime95, load temps got into the mid 50s.
 

WoodButcher

Platinum Member
Mar 10, 2001
2,158
0
76
Dual loops, got it, yes, I don't know why I was thinking single,:confused: Mid 50s at load still seems high. That I may expect from high end air but not water, not w/ an ambient 70 or below. You did drop your voltages as low as possible right?
These are temps we were seeing w/ the B3 6600s, not the G0.
 

Rodknock

Member
Nov 13, 2006
177
0
0
I will admit I haven't really sat down for a couple of days and fine tuned everything. The last couple of weeks have really been hectic, graduating/moving etc., so I haven't been able to devote the time to it that it deserves. A few issues I need to address:

1) I need to remount the cpu block and apply some Ceramique (or however AS spells it) as per the directions, rather than the razor blade method. Also of note, when I first mounted the block, one of the four screws stopped on its thread rather quickly when I was bolting everything down, so I'm going to rotate the mount if I can, and see if everything bolts down better.

2) The wiring is a mess. I went out and bought a SATA dvd drive so I can get rid of that hideous IDE cable, and I'm planning on routing as much of the wiring as I can under the motherboard.

3) I'm unclear if the fans I picked up for the radiators are sufficient, they are the Scythe S-Flex E's (there's D, E, and F), so maybe there isn't enough airflow through them.

4) I'm having an odd problem where the above-mentioned radiator fans are not working via the 3 pin connectors on the mainboard, I've had to connect them to the molex connectors directly from the PSU. If I connect them to the 3 pin slots on the mobo, they simply don't work. Is this a known conflict with these fans, or is something wrong?

5) Once all of that is finished, I just need to dedicate a good few days to stress testing and fine tuning, getting clocks where I want them and getting voltages as loooow as possible for those voltages. I have some problems with this though: mainly, I'm not sure what CPU PLL voltage is, and how raising/lowering it affects CPU temps like vcore does. In addition, I'm unclear of how proportionately the northbridge voltage affects CPU temps.

6) Are my video card temps acceptable? Idling in the low to mid 40s, load temps in the low 50s? Compared to stock temps obviously it's good, as there have been common reports of 8800gtx's running in the 80s C.



-Nick
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,067
3,574
126
1. i think you have a bleeding problem. do you have a top mounted radiator? If so those are insanely hard to bleed. tilt your case from corner to corner while only your pump is on to see if any bubbles get shot out. <wait you have a TJ-07?>

2. Blah, i like ide, i just cut and fold.

3. No, those fans suck hardcore. Find either the zalman F3's or Scythe kaze ultras and put them on controlers.

4. This problem is becuase there isnt enough heat to trigger the AI on the fans. You need to goto bios and set them to whatever voltage you want them at or use speedfan.

5. This is where your going to get tempted the most to cheat. DONT. Patience makes everything more pleasing when done.

6. Yes thats very good for gpu temps. :D
 

Rodknock

Member
Nov 13, 2006
177
0
0
Aig, yes I think you're right about the bleeding. When I put my computer in the car, laid down on its side, there were massive air pockets that rose to the cpu block from either the reservoir or the radiator, I don't remember.

I don't get it, how do you evacuate all the air from the lines? Does bleeding mean getting the air to the top of the reservoir, out of the fluid....or does it mean getting it out of the system *completely*, so that when you move the system they don't move around? I've always been confused about that.

Wow, my fans are that bad huh? I spent like 50 bucks on those:(

I have the Zalman ZF2s on the two 92mm fan slots at the back of the case, and I kind of think they're a bit too loud. What's the absolute best decibel per cfm 120mm fan out there?

I did notice my radiators get warm to the touch, btw.
 

Rodknock

Member
Nov 13, 2006
177
0
0
I just looked up the specs on the Ultra Kaze fan, screw that man, lol. 45 decibel? I don't want an F-22 Raptor taking off in my room. Isn't there something that better-balances noise and flow?