**M570TU/ETU (Sager 5796/5797) Owners Thread**

really good.

 

Sounds too good to be true. ES go for 300 USD on ebay. Strange how they got an OEM with 220 but if you can get it... go for it.

 

Kaltmond, how is the FX3700M working out for you? High temps? What are your temps on the QX9200? Do you have the M860TU or ETU?

 

Its a sad day for us all. Well, mostly just for me. Went to boot up my system today when I was returned with nothing but a blinking light. Sadly, no form of removing components, resetting the BIOS, etc would work. Will have to send it in. If I hear back about what the issue was, I'll keep ya'll POSTed (forgive the terrible pun).

 

Here are some pics for the curious. I just flashed the GPU with a mild overclock of 600/1500/900 @ 1v. This weekend after finals, I will overhaul the GTX 280M heatsink and flash the card with higher clocks. My goal is to hit 650/1625/1000 and stop there.

Installation was a routine walk in the park. I did a rush job on the thermal paste as I will be opening the heatsink again in less than 48 hours.

So far I have played a little bit of GTAIV, Mirror's Edge, and L4D just to test out the card before I went back to studying. During my short test run, the GPU never exceeded 58*C while sitting on top of my NC2000 (fans off) with an ambient temperature of ~20*C.

One of the discreet differences that I have noticed with the new card is the new temperature thresholds for the GPU fan. The 9800M GTX would trigger the GPU fan to turn on when the core temperature hits 42*C and will turn off when the temperature hits 38*C. However, on the GTX 280M, the GPU fan would turn on at 38*C and turn off at 34*C.

9800M GTX on the left. GTX 280M on the right.

 

It seems that I have fallen behind schedule a little bit. I have one more final on Wednesday and those low clearance heatsinks have not arrived yet. So for now, here is the progress that I have begun.

I spent the last few hours peeling off all the stickers from my GTX 280M, and cleaning and purifying the copper surface. I also removed the stock purple thermal pads with high quality thermal pads custom cut just for my memory. Although thermal pads have less thermal conductance than metal, the new pads still perform better than the stock ones in both durability, flexibility, and conductance.

The next part took the longest. I spent most of my time lapping the face of the heatplate that goes over the core. Being careful not to remove any excess copper, I started off with 2000 grit and made my way up to 2500, 3000, and then 5000. Lapping in this case takes any burrs or manufacturing defects in the copper rather than flatten the copper plate; the plate was already near-perfectly flat to begin with. The results are very pleasing to the eye.

When the rest of my heatsinks come in the mail, I will begin mounting them onto the card. I have already planned a blueprint of how I will layout the heatsinks, though I still need the last set of them to be absolutely sure.

 

Where did you get the higher quality thermal pads?

 

I got my RMA and sent my laptop for GPU replacement on Friday. This is my second time. Hope they get it right this time.
I miss my laptop.

 

( http://cgi.ebay.com/BerGquist-Therm..._trksidZp1638Q2em118Q2el1247#ebayphotohosting)

Bergquist makes the best thermal pads on the market bar none.

 

Thanx a lot. When I will upgrade my notebook to 280M I will buy some as well.

 

What really makes them versatile is that they are thicker than the stock pads. So you will need to flatten the pads after you cut the desired size piece. The more dense the pad is, the better conductivity it will have.

 

Hmmm.... call me stupid but would a hammer do ?

 

Yeah, a hammer works. Sledges will give the best results.

I use two glass panes to mush the pieces flat, then edge away any excess compound with a razor.

 

Alright,thank you. When I will have more time and some money I will ... well... hammer the hell out of those thermal pads ).

 

Well after having my 570tu for a couple weeks now I only have one regret. That regret is that I didn't get a quad core cpu. Gta4 just doesn't play that well on a duel core. Sigh why couldn't Rockstar of done a better job porting that.

 

Hmm... i want to know if Bergquist will bring big help to ram oc. If so i´ll order too.

 

Here's the moment of truth. I have loaded Dox's 185.85 drivers and I have mounted most of the larger blocks onto my heatsink and spent a good amount of time during my break flashing new clocks onto my GTX 280M. Starting in small steps from 600/1500/900 to 650/1625/1000, I ran Furmark and ATITool parallel to find the temperature at which the card will begin artifacting.

625/1563/950 was stable before I stopped Furmark at 90*C.
630/1575/970 stable.
640/1600/980 stable.
645/1613/990 stable.
650/1625/1000 stable.

At my goal of 650/1625/1000 and knowing that the GTX 280M's core is tolerant up to 108*C, I decided to keep Furmark running to find the maximum temperature at which ATITool will begin reporting artifacts. ATITool began reporting artifacts at 92*C. Stepping down to 645/1613/990 ATITool also began reporting artifacts at ~92-93*C. At 640/1600/980 and 630/1575/970, artifacts did not appear until 95*C. 625/1563/950 proved to be stable at 95*C and I did not wish to push my card's temperature past that.

It seems that the core is operating just fine at those temperatures. However, it is the memory that is causing the artifacts. Modding the heatsink and replacing the thermal pads does help keep the memory running stable at higher temperatures, but not by much.

In regards to 3d06 and Vantage, those benchmarking applications can never take the card's temperature up to the levels that Furmark can accomplish because they are short running and induces variable loads less than 100% for the duration of each graphics test. Furmark and ATITool can keep pushing my GPU at 100% until I say stop. But with this said, it is very much plausible to run 3d06 and Vantage at 650/1625/1000 because they can never kick up the temperature to the point where the memory will begin artifacting, thus allowing us to record and publish our scores. As for me, I have run 3d06 and Vantage on my machine many times and have recorded each score and the conditions they were assessed in, including my current setup now. But I will remain unwilling to share my results for the sake of keeping the art of overclocking as a test of engineering limits.

In the end, we can say that in the M570TU, the GTX 280M is stable at 650/1625/1000 given that the card is never surpasses 90*C. However, this temperature can only be achieved in synthetic stress testing; I doubt that this temperature can ever be reached in heavy gaming. Reluctant to say though I am very happy with the results, the performance gain from 600/1500/900 to 650/1625/1000 is only a 8.3%/8.3%/11% difference and, in my opinion, is not worth it.

I have reverted to my original clocks of 600/1500/900 @ 1v and will forever stay at those clocks for the lifetime of this machine.

Finally, here are my GPU's throttling level's for anybody who wants to know guaranteed stable voltages.

Extra: 600/1500/900 @ 1v
3D: 400/800/300 @ 0.75v
Throttle: 300/600/300 @ 0.7v
2D: 160/320/80 @ 0.7v

Once the smaller copper heatsinks arrive, whenever they arrive, I will take a picture of the final design and share it with you all so that it may aid those seeking to plan a decent layout for the blocks.

 

625/1625/1000 at 1.05v on the Zalman and it never breaks 80c even after hours. So why wouldn't it be worth it?

 

Because I use my laptop without my NC2000 in hot areas quite frequently. I want sufficient leeway for my GPU to breathe when I am in those conditions. When I made this test yesterday during the Northern California heatwave, the computer lab was hitting 84*F, which were perfect temperatures for me to begin finding the maximum temperatures of my GPU. Along with my loud fans pissing off graduating seniors studying for their finals and luring Macheads and curious freshmen, I was able to get it done.

What I did not mention during the test was that I also choked the GPU fan to achieve a base of 88*C as quickly as possible, then allow the GPU to breathe again and slowly ramp up to it's plateau.

Also, to tell you the truth, I have never turned on my NC2000's fans during the entire time it has been under my ownership. Call me crazy, but I'm just lazy.

 

This is the last test that I will conduct for my GPU. Now that I have settled for 600/1500/900 @ 1v, I ran a stress test on Furmark to find out how hot my card will get after 20 minutes.

Here is my 9800M GTX test as a reference if you wish to compare. ( http://forum.notebookreview.com/showpost.php?p=4737862&postcount=231).

The conditions of the test are nearly identical to the one above. The thermal pads have been replaced, and most of the heatsink mod has been installed. The GTX 280M is clocked at 600/1500/900 @ 1v. The GPU is running on Dox 185.85 drivers. The laptop was sitting on top of an NC2000 (fans off). Ambient temperature was ~70*F. Furmark was configured for stability test in xtreme burning mode at full screen 1920x1200 with 4x MSAA and post-processing.

I began the test when the GPU core temperature rose to 36*C minutes after the GPU fan turned off at 34*C. For the duration of Furmark, the fans were in normal mode, not at maximum or silent mode. From here on I was playing the waiting game. During the test, the temperature rose to 60*C fairly quickly and began to flatten out at 66-67*C at the 15 minute mark and flatlined at those temperatures at the 20 minute mark.

15 minutes.


20 minutes.


I am surprised to say that the results of this test are identical to the results of my 9800M GTX test that I published over a month ago; an observed plateau of ~67*C. I believe this is because the heat generated from the last 16 shaders of the G92M core made up for the die shrink, and also because both the 9800M GTX and the GTX 280M have the same Samsung memory chips rated for 1000Mhz.

Overall, the GTX 280M runs rock stable and very cool after a grueling 20 minute stress test. In realistic applications such as games, the GPU would run well below the temperatures recorded from this test, as opposed to synthetic stress tests that push the card at 100%. However, not to contradict the statement I made from my "road to 650/1625/1000" test, the GTX 280M temperature arc inclines more sharply when operating at 1.05v at the same clock speeds, and even more so at higher clocks. I am very sure that there is a stable clock speed higher than 600/1500/900 at 1v, however I do not have the time or the energy to discover my card's highest stable clocks at 1v, nor is it worth the marginal performance gain.

 

600/1500/950 is stable as stock 1.00v. The card is designed to run 950 mhz memory stock, no reason not to.

 

I just got my 5797 with a q9000 2.0ghz and 280m... Might have been a bad choice. I wanted to get a desktop replacement which was better than my 320mb 8800gts and q6600 at 3.2. I realize the clock times are not comparable but I was shocked when I saw my 280m gettng beat out in 3dmark06 by 800 points in sm 2.0 and 700 in sm 3.0. My cpu score was awful the q9000 racked up a meager 2965 while my q6600 got 5001. Not only that 3dmark results said I had a 180m. Are these sub par scores because of a cpu bottleneck? Is there a way to overclock it setfsb didn't seem to have a setting for my q9000 help

 

It's a cpu bottleneck in the sense that 3dmark06 is too old to be a gpu benchmark these days. Compare actual gaming benchmarks before you make your final decision...

That said, the 280M is faster than the g80 320mb 8800 gts. The 3.2ghz desktop quad is definitely a bonus, but it will make less of a difference than you believe when it comes to actual gaming benchmarks.

You'll want a qx9300 if you're going to make a big deal about your 3dmark score... they can be had for $400 on ebay.

 

I'd have to disagree. Although the 'Theoretical Gflps' is slightly higher on the 280m vs the 320mb gts, the card performs better than the 640mb model in every scenerio, and in certain games and benches, even better than the 8800 Ultra. The G92 architecture is far more powerful clock for clock, shader for shader.

 

So my laptop just got shipped to my gf's, will be opening it up after I get back from work! Yippe!

 

Well I wouldn't have used 3d mark had I not felt choppy game play. Does it maybe have to do with windows 7 7100 build. ? I really think it is the q9000

 

I guess I'll be joining you Gents.

Just ordered my barebone from RJTech and HDD and Ram from the Egg. My QS qx9300 arrived this morning.

 

Hi guys!
After playing several hours of WoW on my np5796 my GPU (9800M GT) usually peaks out at 78C (no OC and I use Dox's 185.68) and after running a defragment program for a few hours my HDD temp peaks out at 63C (WDC WD3200BJKT-00F4T0 320GB 7200RPM)...are these temps normal? I'm not using notebook cooler and I cleaned out the vents a few days ago.

 

From my point of view they are alright. Although the HDD is a bit warm, but mine also gets around 60C when used intensively.

 

you mean agree?

choppy gameplay playing what? The Q9000 is fine, cpu clocks are overrated.

 

I asked the manufacturer to see if they can replace the faulty 9800m GT with a 280M and let me know on a price. Why not, they have it there already.
Let's see what happens.

 

Finally, these black heatsinks came in. Here is the final design for my heatsink overhaul. The last addition involved lining the heatpipes and the core with the aluminium blocks. I have not thoroughly tested the GPU on full load yet, but at idle, with an ambient tmperature of 20*C, my initial observations are that, under browsing NBR and other sites, it now takes longer for the GPU temperature to rise to 38*C and turn on the fan. Also, it takes shorter time for the fan to cool the GPU back to 34*C and then turn off. Sometimes the GPU just sits around at 35-36*C on an idle desktop with the fan off.

I find the first results intriguing as I did not expect a noticeable improvement from adding aluminium heatsinks. It may be because the six blocks directly over the GPU core are right under the bottom vent on the laptop, thus improving cooling by a fair level.

I will try later on tonight to see if there is any improvement under playing games and in stress tests.

 

a little cooling help for you coming


http://www.technologyreview.com/computing/22668/

 

Didn't Asus have something like this in their laptops?

 

Nope, it's a new thing, but it will also eat through your batter life like mad.

 

Ordered my 5797 from Xotic, waiting on the Customization and QA. Very excited for it to arrive, anyone know how long it usually takes for the customization? It changed phases on Tuesday, so I'm hoping it will ship on Monday at the latest. Here are the specs:

- Display: 17" WUXGA "Glare Type" Super Cl (1920 x 1200)
- Processor: ~Intel® T9550 45nm "Montevina" C
- Video Card: nVidia GeForce GTX 280M 1,024MB
- Ram: ~ 4,096MB DDR3 1066MHz Dual Chan
- Exterior Finish: NP579x Orange Frame Trim
- Optical Drive: ~Combo 8x8x6x4x Dual Layer DVD +
- Primary Hard Drive: ~ 320GB 7200RPM (Serial-ATA II 3

Can't wait for it

 

Hi,
sorry guys i searched alot but i didnt get it. I wanna overclock my t9600 (2,8 ghz) to 3 ghz. i read about setfsb and was very lucky. but i couldnt find an working ppl clock generator file for my pm45 chipset. i read about it is not possible to overclock the t9600 in an m570tu is that correct?

no setfsb working.. no other genius tool which allows me to overclock my cpu to 3 ghz? makes med sad...

greetings flode

ps. @eric nice system congratz you are some lucky guy you will have an awesome time with this baby

 

Welcome to NBR, ericb531. Very nice setup you got there too.

It seems that many M570TU owners went for the orange trim. Is there anyone else other than myself that has bought the two-tone silver trim? I would have gone with the orange myself as I love the sleek look of all black and a competition orange stripe around the edges. But where I live, I went with sterling silver to fight back the Mac.

 

I got silver. Not into bling, nor making my laptop look more appealing to thieves.

 

I chose black because my past 2 laptops weren't black (silver/grey toshiba and like a blueish grey metallic HP) Anything besides black gets old after a while...

 

Ok I got the X9100 which I mentioned before (It is truly OEM)

I runned 1 3DMark06 and I got these temps (CPU Overclocked to 3458Mhz and GPU overclocket too)



Do you think these temps are ok? CPU idle temp is 42-48C on "Balanced" mode.

 

those temps are perfectly fine.

you can put to to max performance and the temps will still be fine.

 

Damn!

I wanna know the name of that shop as well.

 

I have noticed that my wd 320gb black line drive likes to idle around 50c and extended gaming can lead to temps of 60-66c which i know is well out of line for safe hdd temps, even with my targus laptop cooler. Anyone have any ideas on how to bring the temps down to a more reasonable level?

 

I have the same CPU and I get the same temps.

 

What program are you all using to set "max performance"?

 

I runned Orthos on 3.06Ghz for 3+ hours, no problem, but when I tried to run it at 3.47Ghz, I get this at "CPU #0":

FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 3 minutes 3 seconds - 1 errors, 0 warnings.
Execution halted.

This means the CPU is defective? Should I send it back?

 

It´s not defective, just unstable @ 3.46G @ that V-Core.

 

So any other x9100 would fail too at this step?

+What's the point in 3.46Ghz if it's unstable? Or I can give it more V-Core?

 

It depends on the chip itself. Some are good some bad, but @default all no problem. The only way to overvolt is pinmod, but it creats too much heat.