b850m mortar wifi wont POST

[ser42ru157f02e0]

Hi all,

After rebuilding the PC into another case, it no longer POST. B850M mortar Wifi with KF556C40BBk2-64 RAM installed
Please, help

 

[citay]

Please list all the hardware, as well as which of the EZ Debug LEDs stay on when you try to POST.

 

[ser42ru157f02e0]

Please list all the hardware, as well as which of the EZ Debug LEDs stay on when you try to POST.

9800x3d, yellow and red lights (drum + cpu)

 

[citay]

No, I meant, all the hardware, down to the PSU model.

But ok, CPU+DRAM LEDs. That can sadly be a common issue on these MSI AM5 boards, it can sometimes happen even from simply moving the board, the reason is not quite known. I think you should start by flashing the latest BIOS version on there, using the Flash BIOS Button:

1. Download the latest BIOS from the MSI support site for your board and extract it.
2. Take a USB stick and format it to FAT32 (it's best for it to have a size of 32 GB or less). You can also try Rufus for formatting, as FAT32 non-bootable.
3. Rename the extracted (!) BIOS file to MSI.ROM and save it to the root of your USB flash drive / USB stick (not in a subfolder).
4. Plug the USB flash drive / USB stick that contains the MSI.ROM file into the marked Flash BIOS Port (!) on the rear I/O panel.
5. With the PC off, press the Flash BIOS Button, and the LED should start flashing for about 5-8 minutes and then the board reboots.

Under a minute of flashing, or more than ten minutes, and it didn't work. After ten minutes it will never finish, you can just turn it off then and try with a different USB stick.

 

[ser42ru157f02e0]

No, I meant, all the hardware, down to the PSU model.

But ok, CPU+DRAM LEDs. That can sadly be a common issue on these MSI AM5 boards, it can sometimes happen even from simply moving the board, the reason is not quite known. I think you should start by flashing the latest BIOS version on there, using the Flash BIOS Button:

1. Download the latest BIOS from the MSI support site for your board and extract it.
2. Take a USB stick and format it to FAT32 (it's best for it to have a size of 32 GB or less). You can also try Rufus for formatting, as FAT32 non-bootable.
3. Rename the extracted (!) BIOS file to MSI.ROM and save it to the root of your USB flash drive / USB stick (not in a subfolder).
4. Plug the USB flash drive / USB stick that contains the MSI.ROM file into the marked Flash BIOS Port (!) on the rear I/O panel.
5. With the PC off, press the Flash BIOS Button, and the LED should start flashing for about 5-8 minutes and then the board reboots.

Under a minute of flashing, or more than ten minutes, and it didn't work. After ten minutes it will never finish, you can just turn it off then and try with a different USB stick.

I tried doing this with all official BIOS firmware versions, in every RAM slot, using only one memory stick. I forgot the model of my PSU(I can check, if it’s really need), but it’s around 1500W. GPU is a Radeon RX 7900 XTX.

I suspect that my RAM might be dead.

 

[citay]

I tried doing this with all official BIOS firmware versions,

And it flashed for 5-8 minutes each time, followed by a reboot, signifying a successful BIOS flash?

in every RAM slot, using only one memory stick.

One module in A2 is the best for troubleshooting, if it previously worked in there. The RAM will not go bad from you shifting your board into another case. This is a weird issue that happens sometimes with MSI AM5 motherboards, we haven't found out why it is happening, it's probably some sort of contact issue somewhere. What you can try is, re-seat your CPU, and while you're at it, take some photos of the CPU socket to see if there are any bent pins in there. Then upload them to an image hoster like https://imgbb.com/ and link them here.

Other RAM doesn't tend to solve this issue. Don't go out and get the temptation to order anything, especially not RAM at the moment. At the very most, you can borrow other DDR5 for a test, or check your RAM in a different PC to confirm it's good. About the PSU, the model is more important to know than the rated wattage, so I can check its quality. But with a CPU and GPU like that, I suppose the PSU will be quite a high-end model.

 

[ser42ru157f02e0]

For some time, I couldn’t run the memory in dual-channel mode — only the A2 slot was working. I decided to slightly increase the voltage, but made a mistake and set VSOC to 1.4. After that, the A2 slot stopped working as well.
Regarding the PSU — it is definitely not a budget unit, it’s a high-quality model.
I spent the whole day today testing with cooldown periods of 10–15 minutes between attempts, so I’ve already gone through the entire forum and realized there’s nothing left to do except write here.

I increased the voltage after message here, and I’m afraid this might have killed my components.

Basically, I did everything you mentioned. I was planning to go to a service center tomorrow to test the RAM.

Thanks for the support. I’ll write an update tomorrow - maybe it will help someone.
Keep your fingers crossed for me… considering current RAM prices, I feel like crying.

 

[citay]

For some time, I couldn’t run the memory in dual-channel mode — only the A2 slot was working. I decided to slightly increase the voltage, but made a mistake and set VSOC to 1.4. After that, the A2 slot stopped working as well.

When you have CPU+DRAM LEDs, you can't set anything as you can't get into the BIOS. So this was all before you put the system into the new case?

With a RAM slot not working at all, what you should do first is to check for bent pins in the CPU socket, because that is a common symptom of bent pins. When you set too high of a voltage for something, you do a Clear CMOS to reset it to defaults. The RAM doesn't die so easily, the CPU can be a bit more delicate. I think SOC voltage has a limit now on AM5 anyway, after they had some problems with too high default voltage.

A slot doesn't stop working from too high voltage, what has most likely happened is that the memory system didn't agree with the settings and now it wouldn't work, but the single module not working in A2 was perhaps coincidence, it wouldn't have worked elsewhere either with those settings.

I increased the voltage after message here, and I’m afraid this might have killed my components.

Again, how can you increase any voltage when the CPU+DRAM LEDs stay on and POST doesn't continue to the BIOS or boot?

Basically, I did everything you mentioned. I was planning to go to a service center tomorrow to test the RAM.

Yes, you can do methodical troubleshooting there to narrow it down. But I wonder what's up with the voltages you mention.

 

[ser42ru157f02e0]

When you have CPU+DRAM LEDs, you can't set anything as you can't get into the BIOS. So this was all before you put the system into the new case?

With a RAM slot not working at all, what you should do first is to check for bent pins in the CPU socket, because that is a common symptom of bent pins. When you set too high of a voltage for something, you do a Clear CMOS to reset it to defaults. The RAM doesn't die so easily, the CPU can be a bit more delicate. I think SOC voltage has a limit now on AM5 anyway, after they had some problems with too high default voltage.

A slot doesn't stop working from too high voltage, what has most likely happened is that the memory system didn't agree with the settings and now it wouldn't work, but the single module not working in A2 was perhaps coincidence, it wouldn't have worked elsewhere either with those settings.



Again, how can you increase any voltage when the CPU+DRAM LEDs stay on and POST doesn't continue to the BIOS or boot?



Yes, you can do methodical troubleshooting there to narrow it down. But I wonder what's up with the voltages you mention.

The sequence of events was as follows:

Initially, I powered on the PC in a new case with two RAM sticks at default frequency without EXPO enabled, but it worried me that the system only booted on the third attempt, with each attempt taking about 10 minutes.

Obviously, my previous overclocking settings had been reset, so I tried to restore them based on screenshots I had. After that, neither RAM stick worked no matter what I did. Resetting via the button didn’t help, and clearing CMOS by shorting the pins didn’t help either.

Reflashing the BIOS via USB helped, but only one RAM stick would work in the A2 slot. Any slightest change in the BIOS would break its operation, and eventually it stopped booting at all.

Then I posted here on the forum, and I immediately managed to get one stick to boot again, but only in A2, by disconnecting the battery contact and waiting for 15 minutes.

After that, I found a post where it was suggested to increase the voltage to make dual-channel mode work — and after that, you already know what happened.

I should have explained things more precisely, but my head was a mess and I was in despair at the time.

 

[citay]

Initially, I powered on the PC in a new case with two RAM sticks at default frequency without EXPO enabled, but it worried me that the system only booted on the third attempt, with each attempt taking about 10 minutes.

That's odd, it seems to have gone through memory training each time, which on AM5 can sometimes take ages. So the board definitely thought that enough had been changed to warrant another memory training. Then why did it do it another two times, I don't know.

Obviously, my previous overclocking settings had been reset, so I tried to restore them based on screenshots I had. After that, neither RAM stick worked no matter what I did. Resetting via the button didn’t help, and clearing CMOS by shorting the pins didn’t help either.

What overclocking, RAM overclocking? With DDR5, you can simply set tREFI to the maximum value (~65K), because tuning this value gives the most benefit for DDR5. If you were to tweak other settings like CAS Latency etc., then you would also most likely need to increase certain voltages and do lots of stability testing. That's why raising tREFI (Refresh Interval) is the best method for the easy gains. Not much stability testing required and it gives a nice improvement, because the RAM spends less time refreshing itself. Usually this is no problem.

After that, I found a post where it was suggested to increase the voltage to make dual-channel mode work — and after that, you already know what happened.

But why would you need to increase a voltage manually when it seemed to have worked fine previously without requiring that? Something else must be wrong here. I would probably re-seat the CPU now and take socket photos as I mentioned. You have to start somewhere, and this you can easily do at home for free.

 

[ser42ru157f02e0]

Это странно, похоже, каждый раз происходило обучение памяти, которое на AM5 иногда занимает целую вечность. Значит, плата определенно посчитала, что изменений достаточно, чтобы оправдать еще одно обучение памяти. Тогда почему это повторилось еще два раза, я не знаю.



Что за разгон, разгон оперативной памяти? С DDR5 вы можете просто установить tREFI на максимальное значение (~65K), поскольку настройка этого значения дает наибольшую выгоду для DDR5. Если вы будете изменять другие параметры, такие как задержка CAS и т. д., то, скорее всего, вам также потребуется увеличить определенные напряжения и провести множество тестов на стабильность. Вот почему увеличение tREFI (интервала обновления) — лучший способ легко добиться прироста производительности. Не требуется много тестов на стабильность, и это дает хорошее улучшение, потому что оперативная память тратит меньше времени на обновление. Обычно это не проблема.



Но зачем вам вручную увеличивать напряжение, если раньше все работало нормально и без этого? Должно быть, проблема в чем-то другом. Я бы, наверное, сейчас же переустановил процессор и сделал фотографии сокета, как я уже говорил. С чего-то нужно начинать, и это легко можно сделать дома бесплатно.

http://deleted

Hi again. In the end, it turned out to be an issue with the CPU - at least it doesn’t work in multiple test setups. I’m going to send it in under warranty, and for now I’ve bought another 9800x3d and it works. Thanks for your replies, the RAM is fine, you were right

 

[citay]

Ok. Yeah, maybe something was already a bit flaky about the CPU and you just put it over the edge with that voltage tweaking.

As for the photos, I can't see anything wrong, but usually the pads on the CPU rarely show signs of damage. If anything, the pins inside the CPU socket of the board are more interesting. For the CPU underside to show signs of damage, something must've seriously gone wrong, and then the board would be dead as well, also with socket damage. But if the old CPU is confirmed dead now and the new CPU works fine, the CPU socket will also be ok.

And the most important thing is that the RAM is ok, yes. This is not the time to buy RAM. Luckily, a lot of brand-name RAM has a lifetime warranty, which then might take a while to get the replacement RAM from them, but at least you could keep both kidneys that way.