The Official MSI GT73VR Owners and Discussions Lounge

The battery wear value is an OCD man's nightmare, it's based on a theoretical value programed in for a battery make / model that may not even be in the laptop, and when the laptop starts being used it might show a "high" wear level that isn't anywhere near the true meaning it's assumed to convey.

And, if it starts out off a lot, it continues to try to calibrate - overshooting and coming back over days and weeks of use.

It's really not a good measure of battery "wear" in the short term even though the name suggests otherwise. Once the value settles on a number after a few weeks of use - whatever number value that is, I'd use that as the 100% capacity mark, and go from there.

Before that reading we used battery tests, running them new, and every few months afterwards. I've only caught one bad battery out of many - and it was one that was set on the shelf to "preserve it for later use", but taking it out of the charging circuit really ruined it.

There aren't a lot of batteries being replaced because they will run for a long time, many small and full recharges, and the 1st and 2nd owner - 3-5 years out won't even notice a difference in run time.

I've pushed batteries very hard and still they run solid, so I really am skeptical that the battery boost would have any negative long term effect on the battery.

The battery boost has been implemented for about 10 years now, and no wide-spread battery failures are being reported, and you know we would be hearing it there were.

 

I've seen more than enough of Alienware and battery wear the last 2-3 years, with battery replacement as the final solution. Often on cost for the owner of the laptop. When you use more of something... The shorter the life span will be. Period!! Seen it a lot here home. The winter here home kill car batteries. A few friends have old cars. See, hard to start in the winter. I'm sure you know what the end results will be.

 

Links for battery failure reports for AW laptops? I haven't seen any for Asus or MSI...maybe AW is the problematic exception?

 

You can avoid that now with the EC hack (just force 330W or 460W power limit) and disconnect the battery, but this won't work for models without 1080 (or 1070 SLI) configurations in the EC.

 

It also won't work on Alienware laptops

 

I'm sure you will find someting from @judal57 3 batteries dead? http://forum.notebookreview.com/search/18419147/?q=battery&o=relevance&c[user][0]=512313

Maybe more locked down?

 

Pretty sure Alienware laptops detect the PSU amps from the middle PSU pin so if a configuration has a 230W, maybe they can buy a 330W PSU for them, if one with the middle pin exists in that configuration. But for 7820HK+1080 users, I don't know. Someone would have to mess with the EC Ram. And messing with the EC Ram is scary if you don't know what you're doing. I succeeded in disconnecting the AC circuit once by playing whack a mole with random values before I had Phoenix's EC screenshot (hearing a light relay type sound) which lasted several seconds. When the AC "reconnected", BD Prochot was suddenly on, and i was at 800 mhz with no load on the chip. Yeah...

 

Yeah, Alienware laptops detect the PSU amps from the middle PSU pin. There are several who have failed due to broken pins. If Dell's firmware can't receive a signal from PSU, the power will be completely limited. In worst case (more as normal) no battery charging. The pin is a weak poit.

 

Not necessarily, not found yet is all

Hey, nice find on the links, you found a guy with no battery problem at all, 4% wear after a year, I think my lowest has been 8% so good find

http://forum.notebookreview.com/threads/alienware-15-battery-specs.793683/#post-10292665

"i have more than a year and my haswell battery is good .. 4% wear"

http://forum.notebookreview.com/threads/alienware-15-battery-specs.793683/#post-10292700

 

A Happy camper... Dell wouldn't replace his 25% weared down battery, but the whole system And I'm sure the battery from the second rev of AW Echobook still shouldwe live. Barely 1.5 years old. Btw I don't remember how @judal57 managed to kill 3 batteries, and now on the 4th.

From my 1.5 years old P870. Still good

 

Really man, "Battery wear" is an illusion, it's a measure of reference number of capacity vs actual, which is a number - if that number never changes it's staying at full actual capacity.

When I said mine was at 8% while true it's a joke, it's really 100% actual capacity for that particular battery. After calibration and settling after a few charge cycles over a few weeks, that's the number that stabilized, and ran for 2+ years at that same 8%.

It's possible to get a bad battery out of the box, just like you can get a bad screen, keyboard, memory, SSD, etc. Stuff can fail out of the box, then you get it RMA'd and have years of good operation.

His first battery was under spec and he had it replaced, then had no further problems. His battery wear level stabilized at 4%, and stayed there for over a year last report, and it's probably still at 4%.

That isn't an example of long term battery boost wear, and I don't think you will find any. We would have heard about it.

Laptop batteries are a lot more resilient than the worriers think. When I hear someone worried about their battery failing from wear, or failing from use, or failing from being plugged in all the time, I think of all the FUD out there that makes them worry.

There's no need to worry.

 

Still, battery Turbo boost aint't a nice feature No need for it if the PSU is big enough @Falkentyne Same if the firmware cut the power below the PSU's rated power.

 

Battery wear would probably decrease as a result of full charge/discharge cycles. If you're on AC power and getting battery drain, you would probably get one discharge cycle a month with normal use, maybe 2 with constant high loads. I don't know. I'm taking a guess.

My laptop was bought in early May and I still go between 97%-101% battery wear. (yes, 101%, it says its rated capacity is lower than its current capacity. whatever). It was at 105% a few times when brand new.

Its still a bit more complicated on the MSI jokebooks though. @Papusan
For one thing if the battery decreases below 30%, hybrid mode gets disabled and max total power draw gets limited to 60% of total AC rating (the point where hybrid starts being used). I guess that's to force the battery to be allowed to charge, but that's still an absurd system. Is the AC Adapter not powerful enough to charge the battery while you're under load or something? Maybe if you were drawing 230W (or 330W) while trying to charge the battery, I could see an issue there, but big deal...what's wrong with the battery simply waiting to charge until after you finish your high load application(s)?

I believe on some of the much older MSI systems, with 180W power adapters, NOS, IIRC from reading (maybe I'm recalling incorrectly), would allow you to draw 220W or 230W combining battery and AC (hey, why not just include a bigger AC adapter?), but a stock GT73VR can't even exceed 230W or 330W even if you tried, unless you were trying to run an Prime95 8 thread AVX load at 4.5 ghz with furmark or Firestrike ultra running at the same time...(and who does that?).

And people should be allowed to unplug their batteries and draw up to the max the AC can handle, if they so choose.

 

The older Gt70 with 780/880m come with 180w Psu. Even with unlocked mobile 4930/4940mx The MSI firmware wouldn't let the laptop use the fully power from the flimsy 180w psu. Once the battery boost was finished draining the battery down to 30% and the hardware lost the extra power and throttled down, the firmware fired up charging of the battery again. But when the battery was charging and you still run full load this time the psu pulled a bit over 200w from the wall. Aka more from the wall when charging the battery vs. when the battery boost was running. Stupidity put in system!! Why not let the psu use fully power when you run the games? Probably Not because they was scared the psu could be killed. But what when the psu run over rated power when charging up the battery due battery boost?

 

I mentioned in an earlier post about the "Black screen" problem, when some users switch from iGPU and then back to dGPU (requiring you to switch to the iGPU blind, by the GPU button ->alt tab->tab->spacebar SCM key sequence, and then loading Bios defaults, etc). Of course people without SCM installed will be required to clear the CMOS to get the screen back (power button depressed for >30 seconds, or holding down power when powering on, so it shuts off again). The black screen problem is triggered by a "warm" power reset happening rather than a cold reset, when changing some Bios options, and then switching to iGPU and back, and may happen right after a CMOS clear/Bios defaults.

The actual trigger for this bug is the PEG port settings. If the PEG port is set to "Auto", a warm reset (often by bios defaults, spread spectrum settings, etc) will trigger the black screen by 'deactivating' the PEG port after switching from iGPU and back to dGPU. it seems that "Auto" becomes the same as "disabled" after this type of switch, after a warm power reset (warm reset is not control alt delete).

The permanent fix for this is to change the PEG 1 port from Auto to Enabled, which permanently fixes this bug. And the dGPU is still disabled when you switch to the iGPU. Yet you can't access the PEG settings without an unlocked Bios or unlocking the Bios options yourself. Catch 22...
Pretty serious bug with a simple fix most users can't even access without paying for an unlocked Bios or unlocking the Bios yourself with @sirgeorge 's instructions (which can run the risk of a brick if you mess up).

 

I updated my guide slightly, added a pre modded bios with the PEG switch enabled by default and a bunch of settings unlocked. Only works on the 6820HK board though as I can't test any changes to a 7820HK bios.

 

@sirgeorge Can you link me to your guide post again?

 

About to pull the trigger on this, but not sure how feel about the 2 power bricks, does anybody have this specific model and can give me some feedback ?

https://www.amazon.com/MSI-GT73VR-S...01IVZ12W0/ref=cm_cr_arp_d_product_top?ie=UTF8

 

Bios Modding Guide

All tools in link below:
https://drive.google.com/open?id=0B_bTuwDv_P1GRm04SlJSTlloV2s

If you have a 6820HK (GT73vr 6RE) you can skip to step marked * and flash the included bios rom after disabling your bios lock using the guide as is.

1) Make sure your bios is set to disable secure boot (this is found in BIOS).
2) Get your bios from the MSI website and extract the rom file from archive.
3) Open the downloaded rom using UEFI tool. Click file -> search, search for GUID 899407D7-99FE-43D8-9A21-79EC328CAC21. This will cause a line of text to appear indicating successful search. Double click this line.

Spoiler: Searching

http://i.imgur.com/9bWhiob.jpg

Spoiler: Double click this

http://i.imgur.com/J3ETqTA.jpg

4) A DXE driver should now be selected called Setup. Right click and save this.

Spoiler: Saving setup

http://i.imgur.com/YWhOGab.jpg

5) Use IFR extractor to extract the saved setup.ffs

Spoiler: Extracting

http://i.imgur.com/IIotQUs.jpg

6) Open the new file generated called setup.txt
7) Search for 'BIOS Lock', you will come to the line displaying the address of the bit used to store the BIOS lock variable. eg) 0x5AE

Spoiler: Finding bios lock

http://i.imgur.com/VNbzOyL.jpg

This is specific to the laptop/bios version so it will vary! Note it down.

* Now format a flash drive using rufus to 'boot efi using MSR'. Extract the contents of the EFI archive in google drive to the formatted usb.
This should result in E:\EFI\BOOT\bootx64.efi

8) Boot the USB drive, this brings you to RU (it'll be tiny if you're on a 4k laptop).
9) Click return to escape the menu and do 'ALT =' to bring up search.
10) Search 'setup'. IF this brings you to a page with only two lines of hex this is not the right page, use the PageUp/PageDown keys to get to another page also called 'setup'.
11) Once on the correct page use CTRL PgDown until you reach your address, eg) 0x5AE for anyone with a 6820HK.
12) At the address should be 01. Type 00 and write this to the address using CTRL W.

I had to do this several times and kept cycling pages back and forth to make sure it stuck!

13) Once the 00 value has stuck at your address, reboot and flash your modified bios with FPT -F biosname.rom -bios

To actually modify your bios dump it in FPT using -D bios.rom -bios. Then open it in AMIBCP. That program is pretty intuitive and allows you access to ALL THE MENU's. To make all menu's accessible click the advanced tab and change all menu's from default to USER.

EDIT: I updated the archive with a working Rufus and a pre modded bios for the 6820HK. PEG1 is set to enabled to avoid black screen when switching IGP/DGP. Full voltage controls are enabled (I recommend LLC values of 15 for both AC and DC in core). To avoid throttling raise SA/CORE/RING limits.

The 75w limit of the PCIE rails can be changed using this bios, the option is found in Advanced/System Agent. It probably doesn't do anything but who knows 0.o

Overclocking performance menu has many goodies in it

Spoiler: AMIBCP

http://i.imgur.com/LRQME21.png

Set the lock to off by default

Spoiler: BIOSLOCK

http://i.imgur.com/8HJLtVP.jpg

 

I have the same model but 4k screen instead. It's good, but get the 7820HK version if you can as they clock higher and seem to be less buggy.

Also that is practically $1000 cheaper than what I paid for mine fml

 

Yeah. https://forum-en.msi.com/index.php?topic=286717.0

As for the dual power bricks, they are a PITA unless you don't move the laptop around at all.

 

yeah the value goes away if i get another model or change anything. How do handle the 2 power bricks ? I carry my laptop everywhere for work and home everyday. Even with one right now its a pain in the ass, that's my only concern really, I don't need the SLI, or the crazy amount of ram, but like you said, I'm saving around $ 1000.

 

You should go for the 1080 model witih a 7820HK CPU (better) and just TDP mod it. It will work better and you won't need two power bricks. Make sure you get the original CM238 chipset version, not the HM175 (which keeps the same name as the original one, except in other countries where it's called GT73EVR).

 

What is the TDP mod, and how do i find out what chipset these companies sell ? I think i'm going with the single 1080 and the 7820 route, will just wait to I see a sale on it. Would love to get a Optical drive with it too.

 

https://github.com/LaneLyng/MobilePascalTDPTweaker/releases

The main thread is on page 2 or 3 of the Sager/Clevo section now.
Coolane has his *1070* pulling 250W from the MXM slot. I am not sure how much the MSI slot can give even though the 1080 has an auxiliary power connector.

 

thank you for your assistance.

 

Also remember its VERY Important to get the **CM238** chipset version!!!!!!!!!
If you cannot find it, get the GT75VR. The same EC firmware RW everything power mod SHOULD work on the GT75VR, as I doubt MSI even knows about what I found with the GT73VR yet (or even if they care); the register still should not be locked down.

 

use this for chipset search

https://us.msi.com/search/GT73VR/product/1

 

@Mobius 1
The problem is, the CM238 version of the GT73VR is *discontinued*. @Donald@HIDevolution mentioned this himself. It's not being produced anymore. So anything you see for sale is old stock. It's been superceded by the GT75VR, and given the downgrade treatment of the HM175, much like the GT72VR/GT62VR relationship's downgrade.

 

Yeah, I guess the GT73 will take place of the GT72 now.

I just hope the price normalizes.

 

hi, im going to repaste CPU and GPU on my GT73vr titan pro-425US with 7820HK and GTX1080. i will replace the thermal pads as well, Do any member knows the thickness of the thermal pads?. 0.5mm or 1mm?. thanks

 

there are multiple thickness pads that are used on the GT73/GT75.

Also if you repaste you don't necessarily have to replace the thermal pads too. Just don't get them dirty or even touch them with your hands.

 

thank you so much

 

Also the GPU thermal pad for VRM and power delivery module is not connected to GPU core heatsink (for single video card models)

 

Falkentyne I have a CM238 wondering what power mod you have done if you want to share.

 

It's a powermod to allow more than an arbitrary amount of power to be drawn through the system.

Only requires you to use RW-everything and can be done inside the OS. -> http://forum.notebookreview.com/thr...cussions-lounge.794968/page-818#post-10593608

 

The main benefit for people who have NOT done the TDP vbios mods is to prevent battery drain from hybrid power, and to allow full power draw at <30% battery level or with the internal battery unplugged.
For people who TDP mod their video cards, then you can draw past the PSU preset cap without the CPU throttling (it's best for GTX 1070 users to have access to the 330W Delta PSU in that case. 1080 users doing TDP mods should be a bit better off, depending on how much sustained load the 330W PSU can handle at or past 330W).

 

@Falkentyne are there any other disadvantages to the HM175 than the loss of 4 PCIE lanes?

 

I would recommend an undervolt for those with 1070 + 230w + power mod. It's easy, free, improve performance, and decreases battery wear.

Possibly more restriction on the EC. Not entirely sure if the loss of x4 means the GPU is running at x8?

 

There wouldn't be any battery wear if you do the power mod anyway. The battery wouldn't be touched until about 260W on 330W AC. Undervolting's great for a 24/7 throttlestop profile, until you're trying to overclock for max performance, then you may need to increase the voltage.

 

I can tell you one thing for sure.
On the HM175, at least when using the EC firmware made for that chipset (17A1EMS1.112), exceeding 80W power draw on the CPU at >4.2 ghz makes the system shut down hard.
Someone with this nerfed "GT73EVR" with the unlocked HK CPU will need to test that themselves, to reproduce that, or see if that's a "bug" with using the HM175 EC with a CM238 board.
There's definitely far more power restrictions as well. Using PSYS_PMAX with a value of 2700 on a GTX 1070 version (230W) causes instant 800 mhz CPU throttling as soon as you load the GPU. Doesn't happen on the CM238 version. Unknown what would happen on the 330W AC profile of the "GT73EVR".

 

Sounds to me like MSI have looked at this forum and decided to clamp down on overclocking the newer model of laptop. If that's the path they're choosing to go down then I'll probs ditch them for an LGA when Volta / Zen2 / Intel 9th gen arrives.
It's a shame, MSI's laptops are the best looking imo.

 

always go for the most powerful single gpu imo.

 

I was thinking that I had an HM 175 but...

Is it possible that I don't have HM175 and don't have CM238 or is it an hw info bug?

Edit : Seems like the CM 236 is an old version of CM 238 and HM170 is the old HM175. Cool!

 

You have the unlocked processor so you're fine.
Virtually all of the HK processors used the CM238 or CM236.
According to @Donald@HIDevolution , some of the new "GT73EVR's" can pair a HM175 with a HK unlocked processor. I would really like to see someone who actually has this abortion of a BGAbook. Isn't the HM175 chipset completely locked down? Or is it the 7700HQ that is locked down, and overclocking is actually possible on a HM chipset (including changing package power limits?).

 

@Falkentyne Do you think there is a reason, your power mod with EC could not be ported/adapted to CM236?

BTW, I'm waiting for the last pieces for my 1070gtx mod. I finally buy the Pomona clip

(The latest MSI gt72vr has hm170 and unlocked CPU. May be they could help to find your informations)

 

Works exactly the same on CM236. It's the same motherboard. You can even flash 17A1EMS1.108 EC into a CM236 and the system will work (you may have to check strange throttling though). The offset and values may or may not be different, but that's easy enough for a couple of users with GT73VR 6rx systems to check. Just takes a couple of people to post their EC RAM screenshots.

 

The Intel® Core™ i7-7820HK is unlocked on the GT73VRs that have it.

 

GOOOGLE my fiend It support overclocing but I think it's limited to 1,25x TDP

 

I saw that link awhile ago but that still doesn't show the absolute limits. I read literally every single post on this thread and I haven't seen anyone with overclocking feedback with a 7820HK and a HM175 chipset.yet. And @Papusan you know, this still depends on what the BGAbook Great Creator MSI says can be done or not.

For example:
Phoenix first flashed the 17A1EMS1.112 EC into his CM238 GT73VR. After seeing that it didn't brick, I tried it also.

The 17A1EMS1.112 EC is for the HM175 chipset version however. Yet it works on CM238. Or rather....."works."


But I noticed problems eventually that I did not know were caused by the EC, and in my impulsiveness, I thought it was just...basic HW limitations of MY system. It wasn't at all. But I didn't put 2+2 together until I went back to 17A1EMS1.108:

1) system performance was identical with no stability issues with the "HM175" EC and CM238 EC.
2) 4.2 ghz @ -100mv showed no differences or anomalies between the two.
3) Bios package power limits (this is NOT the same as platform, aka Psys power limits) worked correctly on both, but I don't know if this is controlled by the Bios (Chipset) or the EC. So someone with a HM175 7820HK has to test this, which I can't do.

4) 4.4 ghz started showing problems:
>80W CPU power draw would cause the system to -shut off- (AC Power off) and then restart within 15 seconds.
4.5 ghz: same power draw: System would shut off and restart in less than 5 seconds.

This did NOT happen on .108 EC.

Changing PSYS_PMAX in the Bios to 2700 (= 337W) for System power Limit (which does nothing btw. This is supposed to be combined with those obscure "Platform Power Limit" settings, would cause the CPU to drop to 800 mhz on ANY video card load. (this is with the 230W AC power limit for GTX 1070). With Psys_PMAX at 2400 (300W), this did NOT happen. This is very bizarre. Why this strange behavior? **NOTE** Throttlestop DID NOT show ANY throttling or "BD Prochot". It showed nothing.

I wish I would have been able to test this with the "RW Everything" Power ID mod to see what would happen, as I'm rather curious. Since there is a GTX 1080 configuration with this "GT73EVR", I wonder what would happen with the "330W power ID". if the same thing did happen, then this is either some extra absurd limit put in by MSI, or a bug. Too bad I didn't know about RW Everything ID back then. But I'm not going to test this! I did my time! Someone else with a GT73VR (or EVR) can do it!

This did NOT happen on .108 EC. I set PSYS_PMax to 2700 with the normal AC 230W profile and the only thing that changed was HWinfo64 reporting that the system was using 300W of power at load. Yeah sure it was....but the CPU didn't run at 800 mhz when I ran Valley...

The ONLY thing I can think of, is if the GT73EVR does not have an SLI configuration (doesn't show it on the MSI site), that the absolute MAXIMUM allowed power would be 330W, however the EC firmware , from my tests with PSYS_PMAX and IMON Current negative offsets, is not fooled at all by PSYS_Pmax or by Imon Offset, so I don't know why a 330W PSYS_PMax rating causes 800 mhz throttling.
What's even more bizarre is, if you set PSYS_PMAX to 8191 on .108 EC, which reports 1000W power draw, nothing happens, except HWinfo64 reporting 900+W of power. No throttling happens, or maybe I didn't test it enough. I don't know. But that's strange. 800 mhz throttling without any flags shown in Throttlestop...

From what 'research' I did, Psys_Pmax is supposed to be used by OEM's to help set their own power limits, via the "Platform" power limit settings, from reading some Intel docs. But no idea why the .112 EC would act that way.

So someone who actually buys the "GT73EVR" with the 7820HK would have to test this to see how much of this is "EC" deliberately set, how much is "Using EC on CM238 when designed for HM175" and how much is actual chipset limitations. But why would anyone want to buy the downgraded version, without ESS Sabre, downgraded Killer Nic/Wifi, and downgraded chipset?

And what would happen if someone with a GT73EVR tried using the .108 EC on their HM175 board? Would it work? Would it 'stop' the 80W CPU package power AC shutoff?

yeah people. I spent way too much time on this in my quest to stop CPU throttling, before I found out about the RW Everything trick...