The Official MSI GT73VR Owners and Discussions Lounge

That is Serbian I will try and report results here. Maybe is good paste to go with this laptop. We will see. Noctua is very very densely paste.

 

It's the same

 

Both MasterGel and ICD are listed as (nano)diamond mix, so both could be similar in performance and side effects. You can't only go by conductivity numbers, testing methods and measurements can vary.

You might be interested in the thread started by @Vasudev, wasn't this a sticky?, but it's currently not:

Which Thermal Paste to buy and apply (Traditional and Liquid Metal)
http://forum.notebookreview.com/thr...nd-apply-traditional-and-liquid-metal.806840/

Here is another thread:

Thermal Grizzly Conductonaut vs Grizzly Kyronaut vs Arctic Silver 5
http://forum.notebookreview.com/thr...ly-kyronaut-vs-arctic-silver-5.802527/page-18

 

Thanks @hmscott
unlike ICD, Mastergel has low viscosity. Composition wise, they are similar and yeah it doesn't scratch the die which they advertise it on their website.
The thread isn't made a sticky, so I bookmarked it.

 

I missed page 1000!

 

I'm having the same problem when doing under volt, anything above -25mv @4.1 either instantly reset the laptop or after few mins, depend on CPU load where in past I was able to do more under volt. my guess may be it's Dragon Center but not sure

 

Delete xtuacpi.inf using Driver store explorer or using Dism++. Make sure you force delete that inf(Make a copy of it before deletion just in case, if it causes more problems) from MSI Dragon SW install folder. Check program Files x64 and Program Data(Caches drivers and msi/exe sometimes)

 

I located the file, its in
system32\driverstore folder and
program files\intel extreme tuning utility folder

so which one to delete as this file xtuacpi.inf (xtuacpidriver.inf) belongs to XTU ?

 

Use this to delete inf or older drivers from Windows driver store https://github.com/lostindark/Drive...ownload/v0.9.5/DriverStoreExplorer.v0.9.5.zip
Use autoruns and uncheck xtuacpidriver so that will take of the driver from starting again.

 

Doesn't make any difference, still even not able to do -30mv @4.1 in XTU

 

I would say that is way too much thermal paste, that paste would be enough for 4 applications, its a laptop heatsink, with low clamping pressure, its not a sandwich with a ton of butter, its just a filler to correct imperfections in the surface of the heatsink and the die.

To properly lap the heatsink you need to remove the 4 metal clips from the screws, and use a flat surface(5mm+ piece of glass, marble countertop), and go to town, your surface finish is pretty rough.

 

Did you also set XTUService in services.msc to manual.
Do you see any WHEA error in HWINFO sensors window?

 

@Falkentyne I just tried running my GT73 SLI with one 230w PSU. This causes throttling while gaming.
The laptop is probably detecting the voltage drop on the input lines.

 

I got nothing sanding down my HS. There is no any difference. One day after repaste again huge core difference. I will try with CM paste if not I will sell it. This is big crap of laptop :S

 

Use throttlestop or hwinfo to see temps, since HWMON always showed huge temp difference than HWINFO or TS.
Revert to stock clock multipliers and use Auto setting for CPU Overclocking in BIOS and control everything using TS. Undervolt it and you will see a difference. I feel Win 10 background services are using single or dual cores giving huge differentials. Use Cinebench to evaluate temps.

 

for 4 GHz CPU do I need to touch in TS anything more or just to set multiplayer to 40x. I will delete dragon center then... It is 6820 HK

 

If you delete MSI Dragon you cannot set fan speeds on the fly.

 

This time I didn't install MSI Dragon and my full speed button still works, for the rest, asking for Svet for the EC Editor and creating a custom profile bumping the speeds a bit up did it for me, no more software running to set fan speeds.

 

I can control fans using silent option, it is the same as DC.

 

good. Then use TS.
Disable hibernation. I hope you have disabled it already.

 

I do not. where I can disable that?

 

Use powershell or cmd prompt as admin and type

Code:

powercfg -h off

 

I am always getting function failed during execution ...

 

You didn't open Powershell as Admin?

 

explains a lot of things. thank you
did you ever work out what that MSIHiddenFun thing was?

 

Still the same with TS. Here is temps and settings. Is everything set well? I managed to undervolt with TS -80mv

 

Yeah looks good, Ivan

I have a question.
Are you using 1mm thermal pads or 0.5mm?

I noticed you said you sanded the entire heatsink.
But if you notice, you are still having the exact same 2 cores getting hotter. So there is still weak pressure on them. Sending you a PM.

 

I am using 0.5 mm pads and I first did a little bit of down where there cooler cores are and then I did entire a little bit with 1500 sandpaper and then polish everything with 2000 sandpaper. Could Cm paste change this to stay under 10 C? I am waiting for PM

 

How much did you sand? 0.1mm is about enough. To sand that part, you can use fully dry high grit sandpaper, or 400-800 grit wet.
Because i'm looking at your heatsink. It looks just fully polished. Notice that the two > marks on opposite corners are still there and look almost pristine still?
You can see its still a pressure problem with contact.

Another thing you can try is sanding the CPU core. there may be marks on it because of the pressure problem. When I stopped using Kryonaut, i had a rather --big-- scuff mark (actually looked like a stretched out star!) on the BOTTOM area of the core that was caused by the imbalanced contact where the two cooler cores had more contact. It wasn't a crack but it 'looked' like a crack. I took sandpaper (polishing) across the entire CPU die until I removed that blemish. DO NOT ever use dry sandpaper or 400 grit on the slug.

If your paste were actually completely drying out, all 4 cores would get hotter. But in fact, the cooler cores are still the same temp as day 1.

on your cpu, do you see any 'marks' around the bottom half? Can you take a picture of it?

 

I just think that I sanded enough. Maybe I will try one more time but I am waiting for tomorrow, I wanna try that CM first. I did everything you said, just I used 1500 git for bottom of HS. I am afraid to sand CPU DIE, I could take picture of it tomorrow because I am going to put CM to try. I can see it is still pressure problem, but I can not believe that is not even small improve now and before sanding at all.

 

I sanded (polished, not truly sanded) my CPU die multiple times. I had to, because of hardened LM. I also had a 'burn' or scuff scratch mark because of imbalanced heatsink when I was still using kryonaut. That scratch was hard to clean. Took like 5 minutes of wiping and cleaning on the CPU.. The trick to the CPU is to keep the sandpaper wet and use FINE GRIT only (2000+). But do my pressure test I just told you in PM first. We need to see how that paste is spreading on the CPU. Use a VERY VERY small dot. like, half a rice grain or smaller.

 

ivan u see when u get a new paste put very thin layer and spread it just to cover entire cpu and also all around cpu a bit to cover all corners im still having 6c difference and temps good after 4th day of repastemaybe around 5c higher but between cores max temp difference is 6c. if it stays like that for a longer time i wont be getting new hs yet, i still dont believe in that. i went around 50 times with a coffe filter paper and rubbed alc 90% over hs so it was absolutely shiny. and i also put all screws same make sure there are not tightened at max just some low force.
however i once experienced shut down where power button was blinking very fast and laptop was tryin to turn on for 2 secs then reboot and constantly happening until i removed charger and plug it back in @hmscott any ideas why that could happened? my bios reseted after that as well ive got a error log can post a picture.. i wasnt able to turn off laptop completely cos it was goin on and off until i unpluged and then holded power button and turned back on... havent clue why that happened. it happened while i was gaming

 

See this picture of this heatsink?
This is GT72VR heatsink, but the main base is the same as GT73. MSI just uses the same heatsink base and then attaches pipes or radiator fins for different chassis.



When I first did my first repaste, I saw a 'dark' mark like that too, as well as what seemed to be a scuff mark around the bottom half of the CPU.

 

I know there can be issues with wifi on these laptops but I'm having an issue where the wifi adapter isn't showing up. I can see the following devices in the device manager: ethernet, bluetooth, and WAN miniport. Both bluetooth and ethernet show up in the network connections in the control panel but not a wifi adapter. I tried a network reset but that didn't do anything. I reseated the wifi card but no luck. Anyone have any ideas on why I can't get the wifi adapter to show up? I usually only game with ethernet (which works) so not that big a deal but still brand new laptop and wifi doesn't work?!? I did see some posts to replace the "killer" wifi card but wanted to exhaust possibilities before I spent more money. Thanks

 

I will try, thank you. Have you screw all screws the same or some tightened more some less?

 

yes, a very slightly a bit more not max.. i just ran cinebench and can see already 8c difference between cores!!!! after few tests, as i predicted. so i will get that hs.. and spread method thin layer helped me most as well

 

where you can buy new Hs and how much it cost?

 

you would need to contact msi customer services via email, i think they are based in Poland but they will need to order hs from china factory. ive contacted uk support msi dont know if there is anything else
i dont know price yet as i have to wait till 23rd then i will know price etc. i think around 50 euros, 60

 

The BIOS CPU voltage out of the box is set high, but only for stock settings.

If you OC the undervolt potential is less and less and the voltage requirement goes up and up - the BIOS set voltage is too high for stock but at some point OC can reach a crossover point where you need +overvolt to maintain stability.

The 7820HK is great in that you can still do a little undervolt at OC. Even at 4.5ghz I had -15mV stable, and that may not seem like much but on the 5950HQ I needed +100mV to get stable at 4.3ghz. That's a big difference from then till now.

Even if you only have -30mV stable at full OC it's still worth setting, as it will have some positive effect, it is still less voltage at the same usage.

And, as I've mentioned before I was able to get a Haswell CPU with -15mV at stock CPU to grow it's undervolt over 18 months to -50mV, which you can do as well, maybe not as much at your full OC, but a little more helps.

-30mV or -25mV is better than needing to add +30mV or +25mV at full OC which is a good way of thinking about it.

 

That's why I posted about ICD vs Mastergel, @Ivan994 already bought Mastergel and ICD is cheap so why not test both? ICD is very thick so it will cover imperfections better than thinner Mastergel, and in this use case ICD might give the same or better results for cheaper.

 

Finally, problem solved.

I find the C:\Program Files\Realtek\Audio\HDA\RtkNGUI64.exe
And right click the analog 1 port, change from "speaker" to "headphone", then click the speaker icon(and I do nothing), then it is fixed.


But it still cant auto switch to headphone when I plug in my head phone.

 

Are you disable startup realtek ?

Sent from my CPH1723 using Tapatalk

 

You mean realtek hd audio manager?
I didnt do anything to it.

 

To be honest, we haven't noticed an uptick in Bluetooth issues with 1709, especially not compared to 1703. There have been some, but I don't remember any users specifically mentioning that it happened after a Windows Update. That's not to say it didn't. Just that no one has mentioned as such.

From a troubleshooting perspective, I've just come to the conclusion that Microsoft is happy to break Bluetooth with every update of Windows 10. I don't know why they find this acceptable, but it's definitely every major update, and it's not down to any one specific brand of radio or anything. I even had a spreadsheet going for a while. There is no consistency, as far as I can tell. This time, Intel seems to be affected more than most, but I can't tell if that's because of anything to do with Intel, other than that they are in more things.

@Ivan994 , what I would suggest is reverting to before the update, uninstalling the Bluetooth driver and checking "Remove the Driver Software for this Device," then letting the update go through. Once it goes through, update the driver to the latest version. Sometimes that helps. Other than that, it's a matter of waiting for Microsoft to fix the problem, as usual.

-- Anthony with Killer Networking

*-*-*Check out this review on the Killer Wireless 1535!*-*-*

 

Next time... Could you try remove the pads and instead use a few lines with paste. Attach heatsink as normal, then disassemble HS and take a pict of what you find?

 

Please do this.
You don't have to even turn on the laptop. Just do what master papusan said and take a picture.

Also, do the test i told you in PM. its' really easy. Do my test first (before removing the pads).
Then do papusan's test.

You can always put the pads back on later.
I want to specifically see the spread pattern.

I delidded again and repasted, did a pressure test with Kryonaut. It was not easy, but I could 'faintly' see reduced pressure around core #3 (top right on CPU, although not very top right corner, it was marked in a picture earlier, you saw it

I actually tried the coolermaster maker Nano. Decent paste. Kryonaut gave better initial temps (since it is higher w/mk) but the Nano was better than my Phobya nanogrease extreme. Got about 5C core temp difference, but cores 0, 2 and 3 were all within 1-2c of each other. Core 1 was like 4C lower. Went back to LM and sanded a few more times. Actually made it worse at first, was getting 3-4C a few times, with core 2 ALWAYS the hotter core and core 1 always coolest, then did another 400->800->polish and got it back to 83/82/83/83 at 4.7 ghz and 1.270v. So I'll leave it like that.

Maybe the BGA CPU is not completely flat. That is what sucks about BGA trash @Papusan if this were a LGA CPU, I could easily hold it up to a mirror and test for flatness of the slug. But VRM's, motherboard, chokes, everything's in the way. But I got the temps back to where they should be. Maybe 1C Better than before, too. I'm done messing with it. Apparently me and iunlock have two of the best clocking 7820hk's.

Tokyotim has one which can only prime at 4.6 ghz max, at 1.30v, but his chip runs cooler at 1.295v (same at 1.30v) than mine does at 1.27v (at 4.6 ghz tested, not 4.7 ghz. And his power draw at same exact vcore is less than mine, and temps much lower than mine. BUT he needs more volts to get stable). But I need a LOT less volts to get 4.6 stable than he does. And I can do 4.7 ghz at 1.27v. He will BSOD at 4.6 ghz 1.275v in prime95 (no AVX, small FFT)

I think that's the old "low VID high VID" thing discussed years ago.
LOW VID chips require less voltage for clocks, but run hotter (more "leaky"), and don't scale as well on subzero. High VID chips require more voltage, but run cooler than low VID chips (if tested at the same voltage and speed), and are better for subzero cooling

 

I could. I am just waiting for tomorrow to get master nano and try with that. What can I get with that? What do you think? Could master nano help me?

 

All depends on heatsink fit on die. Nothing will really help if the HS fits is bad. I want to see how the heatsink fits without pads(imprint on paste).

Edit. And use very little paste on die. 3/4 long thin line.

 

That's exactly what I told you in PM, @Ivan994
can you please read my PM again and do what i said, and take a picture of the CPU 'spread' like i mentioned, for me and Papusan ? Read my PM again. Dont spread the paste. Put a rice sized dot in the very middle of the CPU, then attach heatsink, screw it down, then unscrew and release, and take picture.

Then, do the same thing, but remove the pads (save the thermal pads for later), for Papusan.

 

Or, you could use a consumer grade vertical " vernier callipers". I still remember when that technician measured a chip in my tv with that!