Video Card Problems Guide - Page 4

Capt. Insane · 05-21-2007, 07:50 PM

Quote:

Originally Posted by {DvT}JonahHex

Well hell...

Since I posted a pic of my system

the specs are:

CPU: Intel Core 2 Duo E6600, 2.4GHz
Mobo: EVGA 680i SLI
Video: 2x EVGA GeForce 8800GTX KO (factory O/C)
Monitor: Samsung 226BW, 22" widescreen flatpanel, 2ms, 3000:1 contrast (the pic was taken before I got the Samsung)
RAM: 2x1GB OCZ Platinum PC6400, 800MHz
PSU: BFG 1000W
Sound: Soundblaster Audigy
HDD: 2x Sata/300; Western Digital 250GB, Maxtor 320GB
DVD-RW: MadDog 16x Single Layer, 8x Dual Layer
CD-ROM: Creative 52x
Floppy: 3.5" 1.44MB
K/B: Logitech Gaming G15
Mouse: Logitech Gaming G7, wireless
UPS: APC BX1500LCD, 1500 VA
Speakers: Klipsch 5.1 Ultra
Coffee Cup: Clan Deviants logo'd
Internet: AT&T DSL, 3000MB/s d/l
O/S: Windoze XP Home

Best 3DMark06 score: 15,653 (when CPU O/C'd to 3.53GHz and system is barely stable enough to boot into Windoze and run 3DMark06. Hopefully the new case will cool everything enough to be able to crack the 16,000 barrier)

*drools* Man, I wish I could afford that rig. What speed are the 8800s? About the monitor: good, bad, recommended, price? Good deal man

{DvT}JonahHex · 05-21-2007, 08:14 PM

That rig is the result of screwing around doing nothing on a nickel video poker machine, not winning, not losing (I live in Reno, NV) and getting pissed off and switching to playing dollars - "Take my $ and let me go home" - and, instead, popping a royal flush to the tune of $4000. Over 3/4 it went into building that system. The FDD, sound card, speakers, DVD burner, CD drive, and internet connection are not new - "salvaged" them from the old system and they not included in the cost, above. Without that stroke of good fortune I'd still be scrimping and saving, hoping to be able to do an acceptable upgrade for UT3. You can afford the DvT coffee cup: $10.99 at http://www.cafepress.com/clandvt. The DvT thong skivvies are only $7.99.

The stock (factory O/C) speeds on the 8800s are: Core, 626MHz; Memory, 1000MHz.

The monitor is 2ms response time, 3000:1 contrast, 300cd/m² brightness, no dead pixels, and (still) on sale at Best Buy for $329.99 ($20 off), which is where I got it. I probly could have found it a bit cheaper on-line, but at the time I wanted the instant gratification of hooking it up the same day I shelled out the $ for it. Out of the box, the contrast and brightness were set at 80%; I promptly turned them both down, using the monitor controls, to 50%. I'm quite happy with it.

Capt. Insane · 05-21-2007, 08:38 PM

Since you've been talking about heat issues in this thread, I'll post this here. My brother is an aeronautical engineer working for Avidyne in Boston, Mass. Graduated from Renselear Polytechnic Institute. Anywho, I asked him about dead spots in cases, and this is what he said:

Quote:

There's really no such thing as a "dead spot" in airflow, especially in
such a closed system. I understand what you're getting at - you're
thinking about localized points of circulation that don't mix well with
the surrounding airflow. Honestly, though, even if there are it's not
going to matter much in a computer case. The temperature of the entire
box won't be that much lower than the local temperature there. Just
make sure the size of the "in" fan is about the same size as the "out"
fan (or even make the "out" fan bigger, because you're really looking to
dump heat out of the box). The math involved in this is incredibly
complicated and would require computers and time to solve. That, and I
would need much more specific dimensions than what's supplied on that
website. Seriously, though, doing a thermo analysis on a computer case
is a waste of time; the people that designed it kept these things in mind.

If you're worried about airflow in your current case, just make sure all
the fans are clean. If any of them aren't running, replace them. If
your computer hasn't been shutting off randomly, you're fine. Even
then, it could be any number of other things.

If you're really, really worried, you should be able to check the cpu
temperature in the bios. Just run the computer for a little bit,
restart the machine, hit delete when it tells you to, and look through
the menus.

ChemicalWarrior · 05-21-2007, 08:55 PM

He got to the basic point, the goal is to dump air out, not to exactly bring tons of air in.

{DvT}JonahHex · 05-21-2007, 09:02 PM

What I know for an experimental fact is that, under load (and often when idling), my lower video card runs 5 or more degrees C hotter than the upper card. The experimental condition is shown in the pic of my current case: the 80mm side panel fan blows pretty much directly onto the upper card and misses the lower card, as far as direct flow from the fan is concerned. The 80mm fan is not a high volume fan, and when the side panel is in place, the center of the fan is slightly above the upper card. No way is the lower card getting the cool air the upper card receives.

With the new box, those huge fans will still be somewhat above the lower card, but not as much and they are much high volume output. Obviously the second part of the experiment, test the temps with the new box remains to be done. However, the purpose is to lower the overall temps to give more headroom between the low at-idle temps and the high under-load temps and between the highest under-load temp and the crash-the-system temps. The hypothesis is that the new box will accomplish this, ie, it duplicate previously published results from other user-reviewers.

However, the important part from your point of view with your faulty PSU fan is the effect of the failure to adequately cool electronic components, regardless of how it is done.

I have not previously mentioned that about a year ago my system was acting rather erratically...random system crashes at not really frequent intervals (more annoying than troublesome) that didn't get fixed until I realized one of my exhaust fans on the back of the Green case had failed. Replacing that fan fixed the prob as quickly as cleaning the heat sink fins on my 5950 fixed that prob. Get that PSU (or PSU fan) replaced.

ChemicalWarrior · 05-21-2007, 09:10 PM

You might look into creating a air tunnel of sorts with the bottom card. It is in a rather nasty position when you're talking airflow.

dudio · 05-21-2007, 09:16 PM

Quote:

Originally Posted by ChemicalWarrior

I’m probably going to delete the guides unless there are serious reasons given to keep them up. ..
...On a side note, I seriously hope you are not overclocking your processor with the stock Intel heatsink. Talk about heat issues, no amount of indirect cooling is going to stop you from ruining that processor far faster than it should…

Chem: I think you are very knowledgeable, and I appreciate these guides.
That said, I disagree with O/C'ing the CPU and keeping the stiock cooling.
Works fine for my P4 1.8 @ 2.4, and peaks at 60c (45 idle).
The fan is efficient enough, but may be of note that I need no core voltage increase

ChemicalWarrior · 05-21-2007, 09:23 PM

Quote:

Originally Posted by dudio

Chem: I think you are very knowledgeable, and I appreciate these guides.
That said, I disagree with O/C'ing the CPU and keeping the stiock cooling.
Works fine for my P4 1.8 @ 2.4, and peaks at 60c (45 idle).
The fan is efficient enough, but may be of note that I need no core voltage increase

Depends on the processor a bit, but the Core 2 Duo's are not exactly extremely cool running chips like they are so talked up as. Many actually think the chips are cooling than they actually are because there are only two or so programs that can properly read the sensor. I run my Core 2 Duo E6300 at 2.8Ghz (stock voltage) and at that it reaches 65C under load with a Arctic Freezer 7 Pro, with stock cooling I wouldn't have dared to do it, not only would the fan be insanely loud but the temperature another 10C higher most likely.

{DvT}JonahHex · 05-21-2007, 10:03 PM

I'm not a big fan of O/C'ing. I do it, but not to any extreme. Yeah, when I pushed it to get my highest possible 3DMark06 score, I went to an extreme but slowed it right back down once the testing was done. I knew I could have fried it while doing so, but the overtemp shutdown of the CPU seems to work quite effectively...so...I held my breath and went for it.

You are pushing your E6300 from 1.8GHz to 2.8GHz - 55% O/C. At the moment I'm running my E6600 at stock settings (2.4GHz), but will probably kick it up to 3.0 or 3.2 - 25% to 33% O/C when I see what the new case does for me. The E6600 will do 3.0MHz w/o tweaking the voltages from stock and I actually reduce some of the BIOS default voltages a notch when running at the stock 2.4GHz.

I have the nVidia Monitor, TAT (Intel Thermal Analysis Tool), CoreTemp, and the BIOS for watching my CPU temps. They all generally agree within 2 degrees. Right now (just posting here) TAT and nVidia Monitor say the CPU is at 35C, CoreTemp says 34C, and I'm not going to reboot to check the BIOS. TAT can push the CPU (both cores) to 100% load and boost the temps to over 75C; it's the only app I have right now that will or even can load both cores to 100% and hold them there (as far as I know).

While I am quite aware of the dangers of O/C'ing, my experience so far is that the stock HSF for my E6600 is up to dealing with what I have in mind.

{DvT}JonahHex · 05-21-2007, 10:13 PM

Quote:

Originally Posted by ChemicalWarrior

120C is a typical temperature set as to when a video card will start to throttle itself in order to lower temperatures. While not all cards are the exact same, 120C is a general estimate is actually probably highly likely. One person’s person experience really doesn’t mean much in statistical terms. For example, I’ve had personal experience with my own (and countless others) with temperatures in excess of 90C, I had a 6800 that ran at 85C load normally without any issues for over a year and a half.

When I "took issue" with the 120C value, it was based up two things: my individual experience with my 5950Ultra, not a statistically valid sample, and the reports that people are posting regarding the 8800 cards in the EVGA forums that are consistant w/ my own experience with my own, perhaps still not statistically valid in a strict sense, but indicative of expectations we can have in the real world - if we get better than that, great; if not, oh well, "we didn't expect any better."

After you spend some real time in a chemistry lab trying to get the black tar in the bottom of a flask to crystalize into the "pure white fluffy crystals" the literature says you are supposed to get from some chemical reaction or another you'll know what I mean.

When you succeed, you'll know what a real chemical warrior is. hehehe