HELP! MSI GT72 power issues

Right old battery has had it

Code:

DESIGN CAPACITY 86,580 mWh
FULL CHARGE CAPACITY 1,765 mWh

New battery on the other hand hasn't fully charged to 100% once, also Full charge cap Vs Design Cap its not right either, should that there isn't anything there to protect the battery aka its defo not a safe one.

Code:

DESIGN CAPACITY 86,580 mWh
FULL CHARGE CAPACITY 87,168 mWh

I bet if you look at the wires for the old battery to the new battery you will notice the new battery wires being thinner then the old one.

I suggest you purchase on directly from MSI bud.

 

Ok, will remove the new battery ASAP.

I'm happy to purchase a new battery from MSI if it will fix my issue. However, given the laptop crashes when gaming on AC power only, can this really be a battery issue? I don't really want to spend £100 on a new battery unless I'm fairly sure it will fix my problem. Do you think it will?

On a side note - i've been trying to get a quote from MSI for a new battery. On the phone they quoted me ~£80 plus shipping. However, they said i could only purchase by emailing technical support. Technical support refuse to sell me one and redirected me to ipc-computer.eu. ipc-computer show the battery as discontinued and haven't replied to my email. I'm not even sure its possible to buy from MSI!

 

Try tech support again. But ultimately it's upto you how far you want to go.

I don't want you to disable Intel ME like I did as once it's disabled there is no turning back.

Sent from my SNE-LX1 using Tapatalk

 

I'll try tech support again for the battery. Though would you agree that the battery is unlikely to be the cause of the problem (and more likely to be a symptom), given the laptop does not work correctly on AC alone?

I have to say, i'm very curious about whether intel ME is the culprit. I know nothing about it, so this is just a guess. However, i find it peculiar that somehow the weird issue that I'm having with my thunderbolt port seems to be affected by all this. The fact that thunderbolt is controlled by intel seems suspicious.

Weirdly, when no battery is connected to the laptop, the thunderbolt port works properly. When a battery is connected it misbehaves.

 

Intel ME is on the PCH if i remember and i would assume that thunderbolt port is connected to it however disabling the Intel ME can brick your device, i did mine without thinking and was lucky.

But if your having issue with the Thunderbolt port when its connected to the system with a battery connected that means its something more serious i would think.

Does the system behave when nothing is connected to the thunderbolt port?

 

No, the system doesn't behave when nothing is connected to the thunderbolt port. It seems to be another symptom, not a cause.

Is there any way i can confirm/eliminate intel ME as the culprit without actually disabling it?

If it is something more serious, what further tests can i do to diagnose it?

Could the fluctuating battery voltage be caused by a problem with the power IC unit? The power IC seems to be a common problem with the GT72, with symptoms similar to what i'm experiencing. I'm just not sure how to test it.

 

For Power IC, i don't know how to test it. Infact i don't know where that chip is fully.

 

@Khenglish - what's your view? Does this sound like a power IC problem? If so, any chance you could provide instructions on how to test it please?

 

@hacktrix2006 - I've just thought of something else that might be relevant. For years now, i've had issues with my HDMI port. Some TVs/monitors work fine with it, others don't work at all. With TVs/monitors that don't work, the device appears in my display settings and Nvidia control panel, but no picture is shown on the TV. The same devices work fine when connected through an HDMI adapter to the USB-C port.

I've just read that intel ME controls HDCP, which regulates the devices that can be connected via HDMI.

Again, I'm just speculating, but there seems to be an alarming number of intel-related problems here.

 

PM

Sent from my SNE-LX1 using Tapatalk

 

Cap the Cpu power to different lower values and do some testing. The MSIbooks used to have battery boost to feed more power. If you have a kill-a-watt then you can look at with what power draw your machine start crash at with only the AC power connected.

 

Thanks Papusan. How do you recommend i limit CPU power? Just in power settings, or is there specific software I can use?

Also, where do you recommend I set the cap?

I don't have a kill-a-watt but can buy one if needed.

 

Update: I ran some tests today and got some interesting results.

@Papusan - I limited the CPU power using Throttlestop, however, this did not prevent the laptop from crashing ( see here for HWInfo report)

GPU underclocking:

I then reflashed a custom VBIOS, so that i could underclock my GPU, to test whether this would have any effect. - With the max underclock (-544MHz), I managed to run the Heaven benchmark for over an hour with no crash (all previous attempts have crashed in < 30 mins) ( see HWinfo report here).

From here, I incrementally increased the clock speed until the laptop began to crash again, which happened at -210MHz ( See HWinfo report here).

I reduced the clock slightly from here and found that the laptop was stable at -220MHz (able to run heaven for an hour) ( see HWinfo report here).

Removing aftermarket battery:

Given the dubious quality of the aftermarket battery, I opened up the laptop and removed it. After reassembling the laptop, I performed an EC reset.

On AC power alone, I was able to game for an hour with no crash at the same stable underclock -220MHz ( see HWinfo report here). I then ran the Heaven benchmark, however, the laptop crashed around an hour in ( see HWinfo report here)

Given the instability, I underclocked the GPU further and tried to game soon after. However, the laptop continued to crash. The crashing became more and more frequent, despite continuing to underclock my GPU further, until the crash began happening almost immediately after running a game. (see HWinfo reports here, here and here)

The most interesting part - EC reset:

Just before giving up for the day, i remembered what @hacktrix2006 said about EC firmware freaking out, so i decided to hold the underclock constant and perform and EC reset, to see if that had any effect. To my surprise, the laptop went from crashing in less than 5 minutes of game time, to running the same game for over 2 hours with no crash ( See HWinfo report here).

What does all this mean?

Unfortunately i have no idea what all this means? Why does underclocking the GPU seem to increase stability, is it related to power draw, or an issue with the GPU itself? Why do the crashes become more and more frequent up until an EC reset? Why does an EC reset seem to temporarily fix the problem? What are the next steps from here (tests/fixes)?

Hopefully someone can figure this out with all of the HWinfo data that i've provided!

 

Right have you tried updating your EC firmware?

Or reflashing it, might be a screw up there.

latest one is 1782EMS1.109 , just make you sure you defo have a GT72-6QD before flashing.

Sent from my SNE-LX1 using Tapatalk


Edit: Comparing your new HWinfo to your old your GPU is only drawing 66w on the new HWinfo whilst it was drawing 80w on the old.

80w on GTX 970M is right, however with this info it again bonces to something relating to Power IC / GPU or EC

 

@Yan92 How is your laptop crashing? Black screen, turn off completely, PSU light turns off, power light turns off or amber, or just crash to desktop?

 

Hi,
If the PSU LED turns off, it means that there's a short somewhere. Usually on GT72, it's the power jack on the motherboard as the soldering joints are too thin, or sometimes almost non existant. Ask someone who has soldering skills to add a bit of soldering on the joints on the motherboard. I did that on one GT72 I had with same issue and it worked.
If it does not, you may also have a cracked power jack, in that case ebay is your friend is ebay. Order a new one (7€) and off you go!

 

Thanks @hacktrix2006, @Darker01 and @johnbb for your responses!

It depends whether the laptop has a battery connected at the time of the crash.

• If a battery is connected, then the PSU turns off (i.e. it stops delivering power to the laptop and the LED on it turns off). The laptop itself remains on and operates as normal, until the battery dies shortly after*.
• If no battery is connected and the laptop is running on AC alone, the laptop itself turns off completely while gaming. In this scenario, the PSU remains on.

*The battery dying shortly after the PSU switches off is either because my battery is dead, or represents another symptom of the wider problem. I have tried a new aftermarket battery and it still dies shortly after disconnecting the PSU.

My power jack failed completely within the last few months (laptop would not accept charge in any circumstance). I have since replaced my power jack. It is entirely possible that i messed up the installation, or that the new part is failing. However, given that this issue only occurs while gaming, it seems unlikely to me that this is related to the power jack. Would you agree?

I've now purchased a power meter (equivalent to kill-a-watt). I ran the heaven benchmark on the laptop today with no GPU underclock or any other adjustments. As expected, within 30 mins or so the laptop crashed. There was no surge in power draw detected by the meter. The max draw shown on the meter was around 155W, which was measured 10 mins or so before the crash.

I should caveat this result by saying that i'm unsure whether the power meter would register any power draw surge that might cause a crash, as the meter stops measuring when a device is disconnected (i.e. it stopped measuring when the laptop crashed).

I've not tried reflashing my EC firmware. I can confirm that its already up to date. It is looking increasingly likely, in my opinion, that this related to the EC firmware. So I will try reflashing the EC firmware now.

 

I don't think it has anything to do with EC firmware.
PSU light turns off, which means there's a short somewhere on your motherboard. I thought it came from the power socket but if you changed it and soldered it correctly, apart from a faulty spare part or a bad soldering job, the failure is not from here.
It's common on the GT72 that CPU power rail is weak. Maybe one MOSFET is dying, thus the short circuit.
Unfortunately there's no way to test them unless you unsolder them one by one.

 

Yes, a lot of what I've read on other threads seems to suggest that this is a Power IC/ MOSFET failure. Its just a shame there's no way to confirm that! I certainly don't want to put my motherboard through a risky process unless i'm fairly certain that's the problem. Hence why i'm trying to eliminate all other potential culprits, EC etc.

@hacktrix2006 - I've just tried reflashing the EC firmware. It seems to have had some effect, though the laptop is still crashing. After reflashing, I ran the heaven benchmark and the laptop crashed in under 3 minutes, which is much faster than usual ( HWInfo report here). The PSU also turned off this time, which doesn't usually happen when no battery is connected.

I find the guidance on flashing the EC firmware a bit confusing. For a start, the title of the readme file (identical to the link) suggests that this method is for Windows 8, there is no mention of any other OS.
Secondly, step 5 instructs you to remove the battery/AC adapter if the battery is removable, and to just turn on the laptop after the update/flash if the battery is not removable. I'm in a weird position where my battery is not removable, however, i have disconnected it.

The steps that i followed are:

1. extracted the zip file to its own folder on my desktop
2. closed all other windows. I also exited unnecessary programs in the task bar but left some running (windows defender, malwarebytes etc.)
3. ran the .bat file as administrator and pressed "y" as insutructed.
4. turned the laptop on after the update switched it off. I did not remove the charger before doing so, or press the EC reset button.

Did i do anything wrong? I'm also not sure if that process is for reflashing, or just for updates. I've also found this guidance, which suggests there might be another method for flashing.

 

Nope everything you did was right, PSU light going off is a defo short issue. Looks like your going for a short finding mission.

You could check resistance on the VRM or for a short to ground to try and find the issue.


Not to doubt your work but the Power Jack should be first point of call for checking and work from there all it takes is a cold joint.
Sent from my SNE-LX1 using Tapatalk

 

I'm happy to go on a short finding mission but I'm going to need detailed instruction from you guys! I know nothing about this stuff, unfortunately (though I'm learning through this process!).

Actually when my friend and I fitted the new power jack, we did scratch the motherboard (the green PCB, not any traces), which we were aware could cause a short if touched by the solder. However, we removed any solder that was near the scratch and tested it with a multimeter after and there was no short. I just don't understand why a power jack short would only occur when gaming. Surely the problem would occur on the desktop too, or the laptop wouldn't turn on at all, if it was the power jack?

 

Might be a good starting point it might not be a MSI laptop being shown however you do the same process for any laptop.

However wouldn't it be better to send it in to someone whom specializes in this field?

Personally i would.

 

Thanks, I'll take a look when i next get a chance.

I would love to send it to someone that knows what they're doing; however, i don't know anyone that can help me out. I'm also not currently in a position to spend much on this laptop. Repair shops near me charge too much for motherboard repair unfortunately. If i can fix it myself then i'll continue to have a gaming laptop, otherwise it will have to remain faulty, which seems a same. Unless there are any experts on here that could help?

 

troubleshoot 101: 1st thing to do is to remove the dedicated GPU. Don't worry the fan will be at max speed, this is normal EC Panic mode as the mxm card is not present.
Then try to bench hard and see if the laptop still shorts out. If it does, it's most probably the CPU power rail. If not, it's definitely a dying MOSFET on the MXM card.

 

I removed the card last week and the laptop didn't short at all, though i didn't try any benchmarking. I'll try that shortly.

I suspected the MOSFETs on the MXM card. However, i tested the card for a short and there wasn't one (see post quoted below). Could it be that the MOSFET only shorts when under load?

 

Absolutely. The Maxwell cards are know to have very weak power rails. For instance, 970m have only 2 power mosfets and 980m, 3. For the Pascal generation, MSI 1060 mxm have 6 and 1070, 10 IIRC... So no surprise if I tell you that maxwell cards are dropping like flies after a few years.

 

I've just tried switching from discrete to integrated graphics (didn't physically remove the card, as I didn't have the time) and running the heaven benchmark. It's been running for just over an hour so far, and hasn't crashed.

It really sounds like the GPU MOSFETs are the problem. Given that there's no way to confirm this without taking them out. At this point, should I just go ahead and try replacing them?

If so, can anyone point me towards the right part from a UK based retailer please (or somewhere that will deliver here without extortionate delivery)?

The only thing that's throwing me off is that resetting the EC allowed me to game for 2+ hours the other day, immediately after the laptop crashed 3-4 times in a row less than 5 mins into a game. How would this have happened if it's a MOSFET problem (or any hardware problem)?

 

It can happen if the MOSFET is weak and the power demand is too high. You can add a few more MOSFET on the power rail, as there are reserved space on the PCB for additional ones.
But if you benched without any crash in IGFX mode only, it means that the dGPU is working badly now.
Be aware that you can change dGPU on the MSI GT72, up to 1070 gtx... 1060 / 1070 gtx may be a good option if the 970m is dead.

 

Is this the correct part?

The MOSFETs on my card have the following written on them:

87350D
TI553
ZCRT E3

Does anyone have any experience with mouser shipping by the way? The default shipping shows as £12 which includes customs fees (even though there aren't any for low value goods such as these). There are alternate options, which don't show a price but show "bill recipient". If there are no customs fees, is this free delivery?

I'd love to upgrade to a 1060/70 but I'm not able to spend much on this laptop. Looks like they cost a fair bit

 

If you are in the EU, try TME, or Farnell, or RS.
Mouser and Digikey offer free shipping above 40/50€ orders, bellow that, you pay shipping but thats the final price, no surprise VAT or chargers on top.

 

Thanks Senso, looks like RS have it. Just to check, is this the right part?

 

I looks like it is.

 

The MOSFETs have arrived, just waiting on the hot air station now.

Does anyone have any advice on removing/installing MOSFETs? I've not done this before, so would appreciate instructions.

 

Flux, lots of flux.
Pre-heat the pcb at around 120ºC for 3-4 minutes and then put the hot air at around 380ºC and circle on MOSFET at a time, and use fine point tweezers to remove the MOSFET, take note of the rotation, add more flux and put the new mosfet, circle around for 20 seconds a bit further away so the new MOSFET heats up and then solder tension will put it in place, just make sure you dont rotate it 180º.

 

Thanks @senso. So there's no need to use a soldering iron at all? I can use the hot air gun to add the new MOSFET, as well as remove the old?

Is there also no need to add/remove solder either?

How can I check whether the MOSFET was correctly installed once complete? I'm assuming I should be checking with a multimeter before putting the card back in.

 

Maybe just a tiny bit of solder, that chip has a big solder pad under it, so you can't really solder it with just a soldering iron.

To check, the pads need to be all aligned with the component footprint, and the chip needs to be correctly oriented.
Watch this video:


But your chip will not come off that easy, due to much higher thermal pass of the power planes on the gpu pcb.

 

That all makes sense and the video is very helpful, thanks!

Don't I need to worry about the solder touching the wrong pins and creating a short? Or exposing the new part to too much heat?

 

I've just attempted to replace the existing two MOSFETs and add an additional two to the empty pads on the same phases.

Unfortunately the first one was quite difficult and I ended up blackening the PCB with the hot air gun. The other 3 were relatively easy.

Does a blackened PCB matter?

I've also gotten some solder on a neighbouring copper pad, which I'm struggling to remove. It doesn't look like it's creating any bridges. Does this matter?

Pictures here and here.

Are there any tests I should do before installing the GPU? I've tested the resistance between the two large pins on the bottom right of the card and there's 58.5k ohms resistance. This is less than before I started but, as I understand it, doesn't show a short.

 

Update: Since replacing the MOSFETs on the GPU (and adding two additional ones), I've been gaming for a week or so now with no crashes!

Apparently its rare for the GPU to work at all if its the MOSFETs are causing the problem. However, the MOSFET replacement seems to have fixed my intermittent problem!

A big thanks to everyone that helped out! In particular @hacktrix2006 @Khenglish @johnbb @senso

 

Glad it's up and running bud.

Sent from my SNE-LX1 using Tapatalk

 

Good job mate!

 

Interesting. I had this same issue with the AC Adapter shutting off during game play and the light on the AC Adapter turning off. Connected my Alienware AC Adapter and it worked fine.

Ordered a replacement AC Adapter.

 

The problem issue the symptoms for both an AC issue and a Short are the same.

I had the same issue with my old MS-16F2 where I thought it was the AC adaptor but it was the GPU that failed.

Sent from my SNE-LX1 using Tapatalk

 

Hi, i like every one else have power brick turning off and seems to be GPU related and want to give a try to changing the MOSFET, as i understand there are 3 (but can add upto 6total, is there any benifit for this) do you know what the model info for the MOSFET that i need to order?
I also see one member said you could upgrade to 10 series could some one post a link to 10 series GPU that would fit in my gt72 dominator pro should changing my MOSFET go wrong (amazon US or ali express)?

 

Take exactly same MOSFET as the one that are already in place on your GTX. Brand does not matter as long as the specs are exactly the same. Yes you can add up to 6 MOSFET. It's been done already and it does improve stability of the card, especially if the card is overclocked.

1060:
https://www.ebay.fr/itm/NEW-Nvidia-...796890?hash=item4db53786da:g:IEYAAOSwYABeCGSD

1070:
https://www.ebay.fr/itm/NEW-Nvidia-...798315?hash=item4db5378c6b:g:L68AAOSwu1FeCGKL

 

Thanks, I tried to order the MOSFET and apparently they have been discontinued so my order was cancelled I also don't see my self spending that much getting a 1060 or 70 as I'd rather put that towards a new card.

 

So managed to find s few places to buy it can you confirm which of these are correct
https://www.digipart.com/part/csd87...FYs8i1JK36XRC6M13aIZcM-fJcqE1m9hoCMoUQAvD_BwE

Thanks in advanced

 

can't tell mate. IDK which one are on your 980m. They need to be same specs

 

I went ahead and ordered I had the same model number as the previous guy just the lower numbers were different but it's the same 980 card so assume it's the same.
Model: 87350D was written on it. Any way the laptop in its current state is very useless in happy to take a chance

 

as long as the first numbers are the same, the other ones are usually the manufacturing date