msi gaming plus wifi POST issues

[phi159402e4]

I have had my system for 3 or 4 years and it has been working fine however I've recently had an issue where the red and yellow LEDs have been stuck for random amounts of time which can range from a few minutes to hours. Once the system does pass POST it has been working fine. The last couple of days the system won't boot at all, or is taking so long that I've run out of patience and power it down. I assumed the issue was RAM related so I've tried swapping the sticks around, running with just 1 stick, clearing the CMOS, updating the bios etc but it didn't seem to make a difference. I have 2x16gb corsair dominator 5600mhz cl36 and am reluctant to replace it due to current pricing and I'm not sure it will fix the problem anyway, any other suggestions?

 

[phi159402e4]

The system just came up! I think the super long boot this time may have been due to me clearing the CMOS so it reverted to default settings which were worse than the A-XMP auto settings.
Anyway I've now managed to access the bios and wanted to try and set up the ram overclocking manually but there are a plethora of settings so I'm not sure what to do. The ram is 5600mhz cl36 as I said and I believe the target voltage should be 1.25v but there are a lot of voltage settings in the bios, does anyone know which ones need to change?

 

[citay]

What's the exact board model and the rest of your components including PSU? Because "msi gaming plus wifi", there's many boards like that, at least ten different ones for AM5 alone.

Yes, after a Clear CMOS, the first boot will take longer as it's doing memory training (finding good electrical parameters etc. for the installed RAM). It will start with safe settings for the RAM at first, it's normal. But also, the system shouldn't one day become unable to drive your RAM at the proper speed it's advertised with. All the more because 2x 16 GB DDR5-5600 should not pose a bit challenge for the memory system. So I would be hesitant to blame it all on the RAM here.

Let's first see which components you have in your system.

 

[phi159402e4]

Thanks for the quick response. It looks like I somehow left out the b650 of my board model, it's a b650 gaming plus wifi, my bad. I'll have to get back to you with the exact psu as I don't have access to my full email account from the laptop I'm using here but I know it's a corsair, I think it's an rmx750 but I need to check to be sure. I remember ordering a slightly more powerful one than I actually needed at the time for upgradeability reasons. In case you ask, the cpu and gpu are both amd 7600

EDIT: I can confirm the PSU is a Corsair RM750X Shift 80+

 

[citay]

Ok, B650, I had a feeling, because sometimes on the MSI AM5 600-series boards, that CPU+DRAM LEDs issue can start to appear out of the blue, for no apparent reason. Usually the RAM is not really the issue there, rather perhaps some kind of contact issue between the CPU and the CPU socket of the board, if I had to guess.

You can access the BIOS now? Is it the latest version?

 

[phi159402e4]

Ok, B650, I had a feeling, because sometimes on the MSI AM5 600-series boards, that CPU+DRAM LEDs issue can start to appear out of the blue, for no apparent reason. Usually the RAM is not really the issue there, rather perhaps some kind of contact issue between the CPU and the CPU socket of the board, if I had to guess.

You can access the BIOS now? Is it the latest version?

Yes the PC is running atm but it took a few hours to get through POST, it's not quite the latest version but is fairly recent, N2 I believe.

If it was a contact issue, wouldn't I expect some instability issues? The system has always behaved so far once POST has completed. Yes my issue is that the red+yellow LED thing has appeared out of the blue and seems to be lasting longer and longer although it has been kind of random. ftr I have had the PC for almost exactly 3 years and the problem first started appearing a couple of months or so ago

 

[citay]

I'd update to the very latest version, it's always an easy thing to try first. What's your CPU cooler? If you can get to your full list of exact hardware now, it would be good to post everything.

 

[phi159402e4]

I'd update to the very latest version, it's always an easy thing to try first. What's your CPU cooler? If you can get to your full list of exact hardware now, it would be good to post everything.

Ok I'll try that but it seems like a shot in the dark given that the current bios is recent and didn't appear to make a difference when I last updated it.

Mobo: MSI b650 gaming plus wifi
CPU: AMD Ryzen 5 7600
GPU: AMD Radeon RX 7600
RAM: 2xCorsair 16gb 5600hz cl 36
SSD 2TB Seagate Firecuda 530 M.2 PCIe
PSU: Corsair RM750X SHIFT 80+
Case: Lian Li Lancool 216
Fan: Cooler Master Hyper 212 Air Cooler

The PC was kinda custom built in that I took a number of options offered by the manufacturer on one of their off the shelf systems and ended up replacing just about everything with a view to improve the overall spec a little without stretching my budget too far. So basically I believe the psu and cooling should be better than what they recommended. I should also add that I moved a couple of 1TB HDDs from my previous PC but I wouldn't have thought that would impact on power too much and it was a while before my current problem anyway

 

[citay]

Ok I'll try that but it seems like a shot in the dark given that the current bios is recent and didn't appear to make a difference when I last updated it.

Even if it's a long shot, it's low-effort and takes ten minutes, so I'd always try this no matter what.

The components are all pretty decent quality. What you can try next after the BIOS update is a CPU re-seat. Remove power from the PSU, take off the heatsink and take out the CPU. I don't know how experienced you are in this, seeing how you got a pre-built PC, so maybe watch a few videos on Youtube first. Be especially careful when handling the CPU, because the CPU socket (in which the CPU sits) has lots of tiny and fragile pins, which are prone to damage. And at this point, it's also a good opportunity to take some macro photos of the socket if possible, upload them to an image hoster (for size) and link them here.

Afterwards - unless you already notice some bent pins in the socket - you can install the CPU again and mount the cooler on top, re-using the thermal paste for the time being. Tighten the cooler mounting screws in a cross-pattern, a couple turns each side until it's fully tightened. Should you get the issue again, loosen the cooler screws by half a turn or a turn each.

Since I recommended to re-use the thermal paste for now (because it's not sure it will all be solved already): If you see that they absolutely caked the CPU in thermal paste, and it's even on the sides where you'd grip it, use some thin gloves perhaps, which still allow you a good grip. Thermal paste can be difficult to get off your hands.

 

[phi159402e4]

Even if it's a long shot, it's low-effort and takes ten minutes, so I'd always try this no matter what.

The components are all pretty decent quality. What you can try next after the BIOS update is a CPU re-seat. Remove power from the PSU, take off the heatsink and take out the CPU. I don't know how experienced you are in this, seeing how you got a pre-built PC, so maybe watch a few videos on Youtube first. Be especially careful when handling the CPU, because the CPU socket (in which the CPU sits) has lots of tiny and fragile pins, which are prone to damage. And at this point, it's also a good opportunity to take some macro photos of the socket if possible, upload them to an image hoster (for size) and link them here.

Afterwards - unless you already notice some bent pins in the socket - you can install the CPU again and mount the cooler on top, re-using the thermal paste for the time being. Tighten the cooler mounting screws in a cross-pattern, a couple turns each side until it's fully tightened. Should you get the issue again, loosen the cooler screws by half a turn or a turn each.

Since I recommended to re-use the thermal paste for now (because it's not sure it will all be solved already): If you see that they absolutely caked the CPU in thermal paste, and it's even on the sides where you'd grip it, use some thin gloves perhaps, which still allow you a good grip. Thermal paste can be difficult to get off your hands.

I have zero experience of doing something like that! I'm happy to upgrade things like RAM and hard drives etc but that's about as far as it goes. I have considered building my own custom PC in the past but figured out that you can pretty much get someone else to do it for relatively little extra cost so decided it wasn't worth the risk.

I've seen from other similar posts to mine that bent CPU pins might result in similar problems but we're mainly talking about systems built from scratch. In my case the system has not exhibited problems for nearly 3 years and it still works fine once POST completes so I very much doubt that is the problem. I also don't think it's a problem with the RAM as such because there is no instability.
The issue seems to be very much restricted to whatever goes on during POST so I suspect either the mobo or maybe some interaction with the RAM, I've read that Corsair DDR5 can be a bit finicky and more than one source has suggested that I should maybe bypass the A-XMP settings and manually configure overclocking. I'm not entirely sure how to do that but I've given it a go. Tbh I find the msi BIOS very confusing as there are a lot of parameters and I'm not a hardware engineer! So:
A-XMP -> Disabled
DRAM speed -> 5600
DRAM voltage -> 1.25v (this also auto changed the VDDQ voltage to the same value but the other voltage values remain on auto)
I also tried setting MCR to enabled once the system had booted
Is that all I need to do?

Interestingly I've just flashed the bios to the latest 1.O2 version and this has resulted in the above settings reverting to defaults, is that supposed to happen? I haven't disconnected the battery so I'm puzzled. The battery is a brand new one btw

 

[citay]

I've seen from other similar posts to mine that bent CPU pins might result in similar problems but we're mainly talking about systems built from scratch. In my case the system has not exhibited problems for nearly 3 years and it still works fine once POST completes so I very much doubt that is the problem.

Most likely there are no bent pins, correct. This step of taking photos would only be to make extra sure of that, since the CPU is out for the re-seat anyway.

You see, the CPU also houses the memory controller, hence the speculation that random CPU+DRAM LEDs may have to do with contact issues between the CPU and the 1718 tiny pins in the socket. How they can develop after three years? Beats me. But, there is no official explanation as to why this keeps happening on MSI AM5 600-series boards to a certain percentage of users. Small percentage, but still, we have a couple cases each month.

To me, there should be some kind of physical cause here, because sometimes it happens after the PC has been moved, or after the CPU cooler has been changed. Consequently, taking out the CPU and installing it again, letting the socket mechanism apply the torque properly and evenly again, might help in such a case. So regarding the physical things to try with the PC to solve it, this would be the most promising one to me.

I have zero experience of doing something like that! I'm happy to upgrade things like RAM and hard drives etc but that's about as far as it goes. I have considered building my own custom PC in the past but figured out that you can pretty much get someone else to do it for relatively little extra cost so decided it wasn't worth the risk.

We all started from zero at one point. Ok, way back in the day, when CPUs consumed 5W of power and had a tiny heatsink (if any) without a fan, and things were more robust, it was probably a bit easier to start messing with it. On the other hand, there was no internet, no Youtube videos explaining things, nothing.

But, yes, this is probably not the best task to start with. It seems you can reliably get the PC to POST now again? Was there ever another delay or being stuck on those LEDs?

Interestingly I've just flashed the bios to the latest 1.O2 version and this has resulted in the above settings reverting to defaults, is that supposed to happen? I haven't disconnected the battery so I'm puzzled. The battery is a brand new one btw

Yes, each BIOS update will reset the main settings to defaults. Only the fan settings are kept sometimes. Batteries last 5-10 years, no need to replace it prematurely unless it loses BIOS settings on each boot.

If you can boot normally now, we should just see how the case develops and not do much of anything. Yes, you can set DDR5-5600 manually, of course A-XMP would also set the timings to something tighter, which would improve performance. But if you want to stay more conservative now, set DDR5-5600.

The BIOS offers all the timings etc. of course, but you don't need to set them. I change maybe a quarter or less of the BIOS settings overall, the rest is good on defaults. It's also ok to only change a handful. The default settings can be too conservative sometimes, but at least they should be stable.

 

[phi159402e4]

Most likely there are no bent pins, correct. This step of taking photos would only be to make extra sure of that, since the CPU is out for the re-seat anyway.

You see, the CPU also houses the memory controller, hence the speculation that random CPU+DRAM LEDs may have to do with contact issues between the CPU and the 1718 tiny pins in the socket. How they can develop after three years? Beats me. But, there is no official explanation as to why this keeps happening on MSI AM5 600-series boards to a certain percentage of users. Small percentage, but still, we have a couple cases each month.

To me, there should be some kind of physical cause here, because sometimes it happens after the PC has been moved, or after the CPU cooler has been changed. Consequently, taking out the CPU and installing it again, letting the socket mechanism apply the torque properly and evenly again, might help in such a case. So regarding the physical things to try with the PC to solve it, this would be the most promising one to me.



We all started from zero at one point. Ok, way back in the day, when CPUs consumed 5W of power and had a tiny heatsink (if any) without a fan, and things were more robust, it was probably a bit easier to start messing with it. On the other hand, there was no internet, no Youtube videos explaining things, nothing.

But, yes, this is probably not the best task to start with. It seems you can reliably get the PC to POST now again? Was there ever another delay or being stuck on those LEDs?



Yes, each BIOS update will reset the main settings to defaults. Only the fan settings are kept sometimes. Batteries last 5-10 years, no need to replace it prematurely unless it loses BIOS settings on each boot.

If you can boot normally now, we should just see how the case develops and not do much of anything. Yes, you can set DDR5-5600 manually, of course A-XMP would also set the timings to something tighter, which would improve performance. But if you want to stay more conservative now, set DDR5-5600.

The BIOS offers all the timings etc. of course, but you don't need to set them. I change maybe a quarter or less of the BIOS settings overall, the rest is good on defaults. It's also ok to only change a handful. The default settings can be too conservative sometimes, but at least they should be stable.

Thanks for the update.

I changed the battery as a precaution as it was easy to do.

I'm still reluctant to try removing the cpu tbh as I think it's unlikely to be the source of the problem and there is a risk that I cause damage that isn't there atm.
On balance since you're saying that my particular board has had a few similar cases reported each month, doesn't that suggest it might be a problem with the board albeit a marginal one? I mean if these problems are all down to clumsy cpu installations then wouldn't you expect these issues to be more random across all of the msi boards?
Presumably you've recommended the cpu reinstall to these customers too, do you know if this has resolved their issues? I get that it might work for new installations sometimes but not so much for previously stable installations like mine. Thinking about it, if I were to check the CPU as you suggest I'd probably be inclined to replace the mobo at that point since I think that is the primary suspect here

ftr I'm currently waiting for the PC to boot, probably 10-15 mins so far so I think the issue is still there, I expect it to boot eventually but it could take all day based on my experience

 

[citay]

I'm still reluctant to try removing the cpu tbh as I think it's unlikely to be the source of the problem and there is a risk that I cause damage that isn't there atm.
On balance since you're saying that my particular board has had a few similar cases reported each month, doesn't that suggest it might be a problem with the board albeit a marginal one? I mean if these problems are all down to clumsy cpu installations then wouldn't you expect these issues to be more random across all of the msi boards?

My suspicion is that there could be some kind of flaw in the design of these boards, maybe in regards to the socket, which can then manifest in such problems in certain scenarios. Since I suspect a physical cause, rather than an electrical defect of some part, it would make sense to try the CPU re-seat. What is the alternative? To RMA the board. So viewed against that, I recommend a CPU re-seat. You can also do a milder version, just loosen the cooler mounting screws by a bit, see if that helps.

The board is likely to be the cause here in one way or another, the only question is, can you fix it yourself by ensuring proper contact across all of the 1700+ pins of the socket to the CPU contact pads, or is it something else on these boards which cannot permanently be fixed?

Trying to get on top of this issue is like fishing in muddy waters. Whenever we have such a thread, people tend to have tried a multitude of things, but there is no single good fix for it yet. We sometimes have surprising things like, it works fine when the PC lies on its right side (with the CPU tower cooler upright), but not when the PC stands normally, with the tower cooler parallel to the floor and thus somewhat pulling on the socket.

If anything, only MSI will know the detailed cause, but there is no word from them about the issue. The MSI AM5 800-series boards not suffering from this issue anywhere near this much might be a hint that they redesigned something for those boards. But as is often the case, they don't communicate anything about this, they prefer to silently RMA boards and to keep such things under wraps. But I'm pretty sure they investigated this issue by now, I mean, if there are enough cases posted here on the forum to say "Here we go again...", then they will get at least as much and more.

 

[phi159402e4]

My suspicion is that there could be some kind of flaw in the design of these boards, maybe in regards to the socket, which can then manifest in such problems in certain scenarios. Since I suspect a physical cause, rather than an electrical defect of some part, it would make sense to try the CPU re-seat. What is the alternative? To RMA the board. So viewed against that, I recommend a CPU re-seat. You can also do a milder version, just loosen the cooler mounting screws by a bit, see if that helps.

The board is likely to be the cause here in one way or another, the only question is, can you fix it yourself by ensuring proper contact across all of the 1700+ pins of the socket to the CPU contact pads, or is it something else on these boards which cannot permanently be fixed?

Trying to get on top of this issue is like fishing in muddy waters. Whenever we have such a thread, people tend to have tried a multitude of things, but there is no single good fix for it yet. We sometimes have surprising things like, it works fine when the PC lies on its right side (with the CPU tower cooler upright), but not when the PC stands normally, with the tower cooler parallel to the floor and thus somewhat pulling on the socket.

If anything, only MSI will know the detailed cause, but there is no word from them about the issue. The MSI AM5 800-series boards not suffering from this issue anywhere near this much might be a hint that they redesigned something for those boards. But as is often the case, they don't communicate anything about this, they prefer to silently RMA boards and to keep such things under wraps. But I'm pretty sure they investigated this issue by now, I mean, if there are enough cases posted here on the forum to say "Here we go again...", then they will get at least as much and more.

I agree it's likely to be a board issue, hopefully something that can be fixed with a sneaky bios update but I'm not super confident. I think that I'll continue to monitor the situation for a while before attempting anything drastic as the PC is for the most part still useable although the 40 minutes or so it took to boot today is highly inconvenient. I sincerely hope that msi can come up with some kind of fix, it won't help their reputation at all if they just ignore issues like this.

Many thanks for your time and assistance on this matter

 

[citay]

If it's really something physical with the board, for example regarding the contact, then it's nothing they could solve with a BIOS update. As mentioned, I'd still update the BIOS as you have done, because it's so easy and low-effort that it doesn't matter how low the chances of helping are. But don't hold your breath waiting for a new BIOS to fix all this. If it's the board, then an RMA of the whole board will be the way to go for MSI, unless you or someone you know is willing to try the CPU re-seat first, or at the very least, loosen the mounting screws of the cooler.

The latter is pretty easy, here, I looked up the cooler manual, it's about loosening these four screws by half a turn at a time:

Up until a couple turns each. First you'd try with them only slightly loosened.
The fan might have to be taken off first, it uses these simple metal clamps:

 

[phi159402e4]

If it's really something physical with the board, for example regarding the contact, then it's nothing they could solve with a BIOS update. As mentioned, I'd still update the BIOS as you have done, because it's so easy and low-effort that it doesn't matter how low the chances of helping are. But don't hold your breath waiting for a new BIOS to fix all this. If it's the board, then an RMA of the whole board will be the way to go for MSI, unless you or someone you know is willing to try the CPU re-seat first, or at the very least, loosen the mounting screws of the cooler.

The latter is pretty easy, here, I looked up the cooler manual, it's about loosening these four screws by half a turn at a time:

View attachment 212141

Up until a couple turns each. First you'd try with them only slightly loosened.
The fan might have to be taken off first, it uses these simple metal clamps:

View attachment 212142

I get what you're saying, the weight of the cooler pulling on the socket seems more plausible than damaged pins tbh. I appreciate you finding those cooler diagrams and I might consider trying what you suggest but I wanted to try a few more things first.
Ftr I've had 2 straight days now where the system has taken about 5 mins to boot, I can probably live with that if it's consistent. The only thing I've done is to manually overclock the ram as I explained, if it remains stable at that for a few more days I'll try enabling the MCR bios setting to see if that helps. The thing is I've had a couple of different sources suggest that corsair ram sticks can be a bit finicky (something to do with the samsung chips they use) and that manual overclocking might be more reliable than A-XMP/EXPO so that seems worth a shot. If that does turn out to help then that might point at something that can be done with the bios

 

[phi159402e4]

So far the changes I've made to override the auto dram overclock settings in the bios seem to have improved things, it's not perfect but it has been significantly better the last few days. The next day took about 10 mins so I tried enabling MCR, after that the timings have been about 35 mins and 5 mins. Although over half an hour on the one occasion is not great, it is still better than the several hours I was experiencing before

 

[citay]

Still far from satisfactory if you ask me. Something has to be done.

 

[tfbisqui156f02d4]

Ah the hyper 212, refitted that cooler every year with the 2600k. Filled the gaps between the copper pipes to avoid any air bubble, although method of applying paste was regularly discussed. I would say after 3 years a repaste is due for that hsf. And it's an easy job, which takes less than 30 minutes. Quicker than your posting times. You do need quality paste though.
As for why after 3 years? Who knows what kind of paste is between your cpu and hsf, and since you don't, it could be cheap paste also. Which dries out over time causing heat to rise etc.
I suggest you watch some yt clips about this subject and start getting brave to do it yourself. Or else someone that does it for a fee.

 

[phi159402e4]

Ah the hyper 212, refitted that cooler every year with the 2600k. Filled the gaps between the copper pipes to avoid any air bubble, although method of applying paste was regularly discussed. I would say after 3 years a repaste is due for that hsf. And it's an easy job, which takes less than 30 minutes. Quicker than your posting times. You do need quality paste though.
As for why after 3 years? Who knows what kind of paste is between your cpu and hsf, and since you don't, it could be cheap paste also. Which dries out over time causing heat to rise etc.
I suggest you watch some yt clips about this subject and start getting brave to do it yourself. Or else someone that does it for a fee.

Thanks for the suggestion. I wasn't entirely sure that could be the problem since I only have issues on power up when the system is cold, I mean wouldn't overheating be more of an issue after running a while? That said, I'm having trouble posting again so I powered down the pc and checked the heat sink expecting it to be hot but it is surprisingly cold, is that right?