MSI MAG B550 Tomahawk + Ryzen 5600X posted once, not anymore...

[Mattsson96]

Hello! 

 

Me and my brother in law are facing a puzzle which proves hard to solve. I am honestly running out of ideas, and I would like to say I have experience in building rigs and problem solving but nothing I can find online, or in my experience seems to be able to solve this one.

 

Back story

My brother in law is building his first ever PC, we ordered the parts together, but as I live 300km away and due to covid, the build part has been via voice calls, him building and me telling him what to do and in what order. Anyway, even though challenging, trying to explain everything for a newbie over the phone, we managed it. From parts to a full build. Awesome, or is it?

 

His build:

• AMD Ryzen 5600X
• MSI MAG B550 Tomahawk Motherboard
• Kingston HyperX Fury, DDR4 3200MHz, CL16, 1.35V (2x8GB) RAM
• A-Data 512GB XPG SX8200 Pro SSD
• Seagate 2TB BarraCuda HDD
• Corsair CX 550M PSU
• KFA2 Geforce GTX 1660 Super GPU

 

Turn of events and the problem

We knew this motherboard needed a BIOS update in order to work with the new Ryzen 5000 series CPUs. We did assume that the shop had done this already for us though, as the CPU+mobo+RAM were sold in a bundle. We later realised that this was not the case, as we saw on the bundle page a disclaimer "Unfortunately this motherboard does not support the CPU out of the box blablabla... THOUGH, we have only bundled motherboards which have a flash bios button which allows you to easily update your BIOS without the need for a compatible CPU". It has to be said we did not see this disclaimer until we had put everything together and powered it on for the first time.

 

After powering it on for the first time, everything connected and installed, RAM in slots A2 and B2 (second and fourth counting from the CPU), we instantly wre greeted with a screen indicating how much RAM was installed, what CPU (it did say Ryzen 5600x), hard drive and so on. Press F1 to enter setup, awesome! We went into the BIOS and looked around. We quickly noticed that the m.2 nvme SSD was not being recognized which was weird (Installed in the top spot, closest to the CPU). We planned on checking that it was physically installed correctly of course, but as we were already in the BIOS, we opted to set the XMP profile to profile 1, which would allow our 3200MHz ram to run at 3200Mhz instead of the default 2600. So we activated XMP profile 1 and hit F10 to save and exit. At this point we still think that the latest BIOS is installed, and that our CPU which was recognized by the system, is now supported and everything in terms of BIOS version is fine. After this we powered of the system, unplugged the power and took a look at the SSD, reseated it and made sure it was installed correctly.

After plugging everything back in and turning the system back on, we were greated with a black screen, and the text "No signal" (HDMI used). The fans were spinning and motherboard RGB lit (keyboard not lit). We were baffled and could not understand what had changed since we just a moment ago, were let into the BIOS. The first though was that there is something wrong with the XMP profile with the new CPU, causing the system not to post. So next, we reset the CMOS by removing the battery for a while (first of many). Tried again, same thing.

At this point we realised that the EZ Debug CPU LED was lit and glowing constantly. This would indicate a problem with the CPU assumably, but could also perhaps be because of the CPU not being supported by the BIOS (But then how did we even post the first time at all!?). Anyway, we decided to make sure we were running a supported BIOS by using the flash bios button feature of the tomahawk. This feature would technically allow us to flash the BIOS to the latest, without the need for a CPU or RAM. As we had everything installed already though, we did the flash with all components installed. We attached the USB drive to the correct port, made sure the power supply was on, but the system was off, and clicked the button on the rear IO. The indicator started flashing, indicating it is updating, and after some 5-7 minutes it was done. Awesome, now we should have the latest BIOS, now it has to work?

We made sure that everything was plugged in and started the system, still the same. Not posting.

Next we thought that maybe something was saved from previous, lets clear CMOS again, this time we did it by shorting the CMOS jumper pins (power unplugged) (which I have read would be faster and more reliable than removing the battery for an unknown time). After that, still the same.

We were running out of ideas at this point, but as we were desperate and the CPU LED was still lit, and I had read that the bios flash button was supposed to work without a CPU in the socket (only motherboard and CPU power connected), but I had never read if it was possible to do it WITH a CPU in the socket. So as we were unsure if this would be an issue, and if the CPU was seated correctly, we took the motherboard out of the case and removed the CPU and the RAM, leaving only the 8PIN CPU and 24PIN mobo power connected. We then redid the bios flash, with only the motherboard and power cords. Everything went well, the BIOS flash indicator flashed for 5 ish minutes and turned off, everything pointed to it being successful. At this point we again reset the CMOS, by shorting the jumpers. Then reinstalled the CPU, making sure all CPU pins were looking OK and the cooler and CPU was seated correctly. We installed only ONE ram stick this time, in the second slot (A2). We now moved the motherboard back into the case in order to get sufficient support for the GPU which is needed (as the 5600x does not have integrated graphics). After putting everything back together and plugging only the display and power cord (we left the front IO unplugged, apart from the power, reset and LEDS). STILL THE SAME PROBLEM. The CPU LED still lit...

At this point we are running low on ideas and motivation. We finally try swapping the RAM for the one that is not plugged (testing them one by one), as well as testing them in different slots (A2 and B2). Still same problem. We also tried a different monitor using DVI instead of HDMI. Still the same problem.

Finally we tried bootin the system without the SSD installed (desperately), but still same. It is not even posting with only GPU+CPU+1 RAM stick + PSU...

 

At the time of writing we have the CMOS battery out, leaving it out for 1h (is the plan) and reseating it and trying a final time is the last idea we have. After that the only thing I can personally think of is that there is something wrong with some component, most likely the CPU but also not ruling out the motherboard or even GPU. As he does not have any other AM4 CPU, we are unable to test the board with another CPU at this point, we are thinking of testing the GPU in another system at a later time when we have one available for him.

 

To summarize what we are fairly sure have been done:

• BIOS flashed to latest non beta version (7C91vA4)
• CMOS reset multiple times, also after BIOS installation (both shorting pins and removing the battery).
• All components look physically OK and are correctly seated
• The system posted ONCE after the initial build, but not after (HOW IS THAT POSSIBLE!?)
• Two different monitors and two different outputs tested (DVI and HDMI)
• (Keyboard does not light up)

 

Initially I thought the XMP profile messed something up, but a CMOS reset should have solved that. Then I thought the BIOS version was to blame, but that is almost definately updated now (unfortunately no way to confirm). We are completely out of ideas. Can anyone on the internet think of something that we have not thought of? I have spent hours online, finding very similar problems, though the solutions have always been "reset CMOS", "flash BIOS", "reseat CPU", "reseat RAM" or something else we have already tested.

 

Any help at all is appreciated!  Thank you!

 

Edit 1: We finally posted! Leaving the CMOS battery out for an hour solved the problem. Why not shorting the CMOS jumpers worked is beyond me.... Seems like we will have to run the RAM on 2400MHz for now...

 

Edit 2: We installed the SSD, HDD and the last RAM stick and now we are no longer posting... WHAT IS GOING ON.

 

[voidrunner959]

To make this short I'm in the same boat buddy I'm building from scratch and had all my parts last night and I'm using the 570x tuf board. would not post and was telling me was a dram problem. I'm currently running the new 64GB in my current tower as we speak so its not that. Everything comes on and everything seems to work so my last guess is the 570x needs a bios update but the 570x cant flash bios. 

My last guess is the CUP. I have a 5900X the small thing is I run a 1600X in my current rig sooooo cant use it to update the 570X. Now i have to go buy a 3 series to update the 570X to use the 5900X (this is my hell) Ill keep you posted Im going to best buy to get the 3 series in a hour and after that will see what happens if it just post and boots to my bios i can update it and rule that part out. I feel you pain though 

[Mattsson96]

Believe it or not. I just managed to finish this beast of a forum post, and the very last thing we were trying 

Quote

At the time of writing we have the CMOS battery out, leaving it out for 1h (is the plan) and reseating it and trying a final time is the last idea we have.

, ends up working! We posted!

 

I honestly have no idea how the CMOS had not been cleared before (I even thought BIOS flashes would reset the BIOS settings?), but it looks like the shorting of pins and the not so lengthy removals of the CMOS battery earlier, had not worked. Why the shorting had not been successful is beyond me, the battery removal was probably not long enough.

 

Anyway, I am so happy we finally posted. Now we will try to get the ssd, hdd and the final ram stick installed, windows up and running and hopefully it goes smoothly. Maybe later we dare to touch the XMP profile again. This time we know that in case of trouble, removing the CMOS for an hour will be the reset solution

 

Anyone else experiencing problems with using the XMP profiles with the new 5000-series? For now we will leave it at 2400MHz...

 

I you solve your problem too@voidrunner959. Sounds like buying a 3000 series is a bit overkill? Surely the shop offers BIOS update for a smaller fee than the cost of a CPU?

[voidrunner959]

1 minute ago, Mattsson96 said:

Believe it or not. I just managed to finish this beast of a forum post, and the very last thing we were trying 

, ends up working! We posted!

 

I honestly have no idea how the CMOS had not been cleared before (I even thought BIOS flashes would reset the BIOS settings?), but it looks like the shorting of pins and the not so lengthy removals of the CMOS battery earlier, had not worked. Why the shorting had not been successful is beyond me, the battery removal was probably not long enough.

 

Anyway, I am so happy we finally posted. Now we will try to get the ssd, hdd and the final ram stick installed, windows up and running and hopefully it goes smoothly. Maybe later we dare to touch the XMP profile again. This time we know that in case of trouble, removing the CMOS for an hour will be the reset solution

 

Anyone else experiencing problems with using the XMP profiles with the new 5000-series? For now we will leave it at 2400MHz...

 

I you solve your problem too@voidrunner959. Sounds like buying a 3000 series is a bit overkill? Surely the shop offers BIOS update for a smaller fee than the cost of a CPU?

I can return it open box same day for a refund so just easy i been working on this $2500 tower for so long at this point i just want it to post and update so i can move on with my life 

[cr8tor]

2 minutes ago, Mattsson96 said:

I honestly have no idea how the CMOS had not been cleared before

Removing the battery and unplugging the power is not enough to clear the bios.

There are capacitors in the PSU and on the mobo that will still be charged enough to power the bios enough to hold the settings.

Waiting for them to discharge on their own can take many hours sometimes.

 

To remedy this, while the batter is out, and while the power cord is unplugged, press and hold the power button for 5 seconds.

This will result in a cleared bios every time as pressing the power button instantly drains all residual power.

Often when you do this, you will see fans bump or lights flicker just a bit, that is the residual power in the capacitors being used up.

 

You should be ok using XMP, i would suggest trying it again, but be ready to reset the bios if needed.

Again, rather then letting it set for an hour, press the power button to drain the residual power.

Much quicker.

[Mattsson96]

19 minutes ago, cr8tor said:

Removing the battery and unplugging the power is not enough to clear the bios.

There are capacitors in the PSU and on the mobo that will still be charged enough to power the bios enough to hold the settings.

Waiting for them to discharge on their own can take many hours sometimes.

 

To remedy this, while the batter is out, and while the power cord is unplugged, press and hold the power button for 5 seconds.

This will result in a cleared bios every time as pressing the power button instantly drains all residual power.

Often when you do this, you will see fans bump or lights flicker just a bit, that is the residual power in the capacitors being used up.

 

You should be ok using XMP, i would suggest trying it again, but be ready to reset the bios if needed.

Again, rather then letting it set for an hour, press the power button to drain the residual power.

Much quicker.

True all you say. I just figured this CMOS jumper shorting thing would work well and efficiently as described (never used it myself). 

 

Also, regarding that, I read on the MSI B350, you should have the power cord plugged while doing the jumper thingy, but for the B550 it says it should be unplugged... Not sure if they are both true or which one is... We tried both

 

Also, at the time of writing, we installed the other RAM, the SSD, the HDD and now it no longer posts... WHAT.

[cr8tor]

1 minute ago, Mattsson96 said:

now it longer posts... WHAT.

Did you mean now it NO longer posts?

 

If thats the case, clear the cmos again. Changing the memory can cause issues if its not detected properly.

Its often a good idea to reset the bios any time you make hardware changes.

It forces the bios to redetect hardware, retrain memory, ect.

 

If you clear the bios and it still wont post,

remove the items 1 at a time, clear the bios, then test. That should help you narrow it down to which exact component is causing the issue.

[Mattsson96]

14 minutes ago, cr8tor said:

Did you mean now it NO longer posts?

 

If thats the case, clear the cmos again. Changing the memory can cause issues if its not detected properly.

Its often a good idea to reset the bios any time you make hardware changes.

It forces the bios to redetect hardware, retrain memory, ect.

 

If you clear the bios and it still wont post,

remove the items 1 at a time, clear the bios, then test. That should help you narrow it down to which exact component is causing the issue.

Yes sorry that is what I meant. We are trying that now. Although maybe stupidly in the other order, unplugging everything we just added, trying again. That did not work. Now trying to clear CMOS again with your method but it does not seem to work. Not sure if 5 seconds is enough in our case. Now testing to hold it down longer while battery unplugged

 

Edit: We are going to try "our CMOS reset method", with all components installed. We are fearing that a CMOS reset fixes it temporarily, but it only works for one boot, the second fails no matter if change hardware things or not. We will try to confirm this by leaving the CMOS battery out again for abt an hour. Then reseating it with all components installed.

[cr8tor]

7 minutes ago, Mattsson96 said:

Not sure if 5 seconds is enough in our case.

It is. Never found a PC this wasn't long enough for.

When you press the power button, it tries to turn on.

Trying to turn on sucks down everything left in every capacitor.

Make sure you are leaving the PSU power switch, if it has one, in the ON position and unplug the power cord.

Another possiblity is the PSU in all of this.

If it works once each time you drain the power like this, it often points to a PSU issue.

 

But it sounds like you are moving in the right direction, let us know how it goes.

 

Honestly, im thinking it may be a bad PSU or CPU at this point from whats happening. Just a hunch, but thats my prime suspect so far in this one.

[voidrunner959]

2 minutes ago, cr8tor said:

It is. Never found a PC this wasn't long enough for.

When you press the power button, it tries to turn on.

Trying to turn on sucks down everything left in every capacitor.

Make sure you are leaving the PSU power switch, if it has one, in the ON position and unplug the power cord.

Another possiblity is the PSU in all of this.

If it works once each time you drain the power like this, it often points to a PSU issue.

 

But it sounds like you are moving in the right direction, let us know how it goes.

 

Honestly, im thinking it may be a bad PSU or CPU at this point from whats happening. Just a hunch, but thats my prime suspect so far in this one.

that's my fear is a bad 5000 series 

[Mattsson96]

Just now, voidrunner959 said:

that's my fear is a bad 5000 series 

let's hope you are both wrong and that it is something else..

2 minutes ago, cr8tor said:

It is. Never found a PC this wasn't long enough for.

When you press the power button, it tries to turn on.

Trying to turn on sucks down everything left in every capacitor.

Make sure you are leaving the PSU power switch, if it has one, in the ON position and unplug the power cord.

Another possiblity is the PSU in all of this.

If it works once each time you drain the power like this, it often points to a PSU issue.

 

But it sounds like you are moving in the right direction, let us know how it goes.

 

Honestly, im thinking it may be a bad PSU or CPU at this point from whats happening. Just a hunch, but thats my prime suspect so far in this one.

Ohh. Well I am unsure if he had the PSU in the ON position while doing it. As I said, we are 300km apart connected by voice calls I will check with him. It would be nice if that turns out to be the clearing issue at least. Identifying the actual problem still remains though. I agree, PSU or CPU most likely

[Braini]

27 minutes ago, voidrunner959 said:

that's my fear is a bad 5000 series 

That is very very unlikely unless there is a physical damage. I can give you hope as I had basically the same problem and it was solved by updating the BIOS as you can see here: 

 

[Mattsson96]

6 minutes ago, Braini said:

That is very very unlikely unless there is a physical damage. I can give you hope as I had basically the same problem and it was solved by updating the BIOS as you can see here: 

 

If only our solution would be as simple. As we got it to post once now after, we have confirmed we are running the latest (Non BETA) BIOS, released on 2020-11-04. So the BIOS version should not be our problem

[Braini]

Do I understand your current problem/situation correctly? 

You can repeatedly get your system to POST by resetting CMOS regardless of the connected hardware but only once?

[Mattsson96]

10 minutes ago, Braini said:

Do I understand your current problem/situation correctly? 

You can repeatedly get your system to POST by resetting CMOS regardless of the connected hardware but only once?

That remains to be seen. We are resetting it a second time now (by waiting, next time we will try the quicker way). But yes, I fear that could be the case, but we will soon know for sure. If not now then tomorrow. We will try to boot it once, if it works, then change nothing and boot again to make sure.

Will keep you posted

[Mattsson96]

Final update of tonight. Having installed all components again, left CMOS out for an hour, reseated and booted, we can say that it still won't boot. This indicates that it is probably something that we added that caused the issue. Yes we are aware this was not the smartest, to put everything in at once, but we are tired now so give us some slack

 

The interesting part is, after the CMOS reset, with all components installed, the CPU led is no longer lit, instead the DRAM LED is lit, and of course it is no longer posting. Tomorrow we continue with removing the RAM stick we added last, resetting CMOS and see what happens. Wish us luck

 

If anyone have any further tips please do post them. Also apologies for the double post. I wanted to make sure everyone interested are notified.

[Mattsson96]

Okay another update for you guys. Yesterday we tried some more debugging.

 

The last thing we tried last time was to have every component installed, and reset CMOS by removing the battery for a long time. This did not work and the system did not post, leading us again to think something is wrong with some component.

The first thing we did today was to remove the RAM stick we inserted last time (As last time the DRAM LED was lit), try to reset CMOS by first shorting the CMOS pins (Using the reset button Lifehack!) and then draining the power of the system with the power cord unplugged but with the PSU switched ON, as @cr8torsuggested. After powering on the system we saw no change, and the CPU LED lit. This was rather confusing but next we tried to remove the HDD and the SSD (those were the only new components since last post) one by one and performing this RESET CMOS method (we even tried having battery out + drain). Still no post, CPU LED lit.

Now we were back to the physical state of the machine, in which we had our last successful post. We again tried the above CMOS reset, confident that NOW if ever, we should post if the CMOS is actually cleared properly. But no... No post. So instead of using this method to clear cmos, we went back to basics, leaving the battery out for 1h10min but still having the exact same physical state as the last time we managed to post (trying to replicate it). After 1h10, we posted!. Now comes the interesting part: I figured that we should try restarting the system to see if posts again, without making any changes in terms of HW or BIOS settings. We went into BIOS, F10 to save and exit without changes, reboots and posts. Nice. The plan next was to try to insert the missing ramstick, so we power of the system using the switch on the PSU. At this point I was still curious, from being completely OFF, will it still post (even though we have made no changes other than switched the PSU off)? So we turn the PSU back on (no HW modifications yet), and turn on the system: NO LONGER POSTS, CPU LED LIT. 

 

Bottom line it seems like if we turn the system ON from a drained state, it posts (but for some reason it did not yesterday with all components!? Maybe we did not wait long enough) it posts once. If we then turn it completely off, using the PSU switch (maybe even using the case power button), and then power it back on, it no longer posts. 

It seems to me like this could indicate an issue with the PSU? From a drained state it works once, but trying to boot while there is still some load in the capacitors, causes a no post.

Sure it could still be something else, but I suspect the PSU more strongly right now.

 

Also, why the CMOS pin shorting + power draining OR battery out + power drain methods won't work for us is also a mystery to me... The only thing that has made a difference for us is battery out + patience...

 

You guys agree with my diagnosis?

 

Today my brother in law is taking the PC to a friend of mine that lives close by. He has access to a AMD system of his own, allowing him to quite easily test each component, one by one to pin point the issue.

[Braini]

If you switch off the PC and repeatedly press the power button for like 10 seconds the PC should be completely "drained". If this works you can do some more testing w/o wasting so much time. For example, is removing the CMOS battery change needed at all to get the PC in a postable state or is it only the "draining". If hot rebooting always keeps the system postable and a cold reboot always results in a no POST than it seems to me that the process of turning the system OFF puts it in an unpostable state which really would be a weird problem.

[Mattsson96]

4 hours ago, Braini said:

If you switch off the PC and repeatedly press the power button for like 10 seconds the PC should be completely "drained". If this works you can do some more testing w/o wasting so much time. For example, is removing the CMOS battery change needed at all to get the PC in a postable state or is it only the "draining". If hot rebooting always keeps the system postable and a cold reboot always results in a no POST than it seems to me that the process of turning the system OFF puts it in an unpostable state which really would be a weird problem.

You make excellent points and I agree to all of them. I hope my friend, who will debug this system tomorrow (I hope), will read this and try repeatedly holding the power button to test the theory. No matter what it is, it is odd... Will be interesting to get to know where the issue lies. He should be able to find it by trial and error using a known working system (something my brother in law, did not have available).

[cr8tor]

14 hours ago, Mattsson96 said:

No post. So instead of using this method to clear cmos, we went back to basics, leaving the battery out for 1h10min

It did work, i promise you. It just didnt fix the issue.

If it has to sit for an hour or more, and then posts once, and has to sit for an hour again, id bet its the power supply.

Im too poor to actually bet, but if i was rich, id put money on the PSU at this point.

[Mark Kaine]

Ok this is really hard to follow... 

 

SO AFTER THE SYSTEM TURNED ON A SECOND TIME, what did you change? 

 

There's simply no reason that you couldn't revert these changes easily to resolve the issue. 

 

I do believe you did not do that though. Is this so far correct? 

[cr8tor]

1 minute ago, Mark Kaine said:

There's simply no reason that you couldn't revert these changes easily to resolve the issue. 

Unless the issue is sporadic. 

[Mark Kaine]

1 minute ago, cr8tor said:

Unless the issue is sporadic. 

Well yes, but honestly it's completely unclear what they did, what was changed after the second time the pc posted? 

 

Or did it simply not post  again without any changes whatsoever? 

 

 

[cr8tor]

8 minutes ago, Mark Kaine said:

There's simply no reason that you couldn't revert these changes easily to resolve the issue. 

Unless the issue is sporadic. If its not happening every boot, it makes it really hard to diagnose the issue.

 

 

Also, your all caps signature is pretty tough to read. And i think you have a typo at the end.  
Windows Go Home? 

Edited November 22, 2020 by cr8tor
Didnt mean to post this twice, not sure how the first one was submitted

[Mark Kaine]

28 minutes ago, cr8tor said:

your all caps signature is pretty tough to read

I should make it RGB, shouldn't I? (tho I'm lazy)

 

28 minutes ago, cr8tor said:

Windows Go Home? 

It's not a typo, the full name is "Windows 10 Go Home And Be A Family Man" but I thought that was too long ~

 

 

28 minutes ago, cr8tor said:

If its not happening every boot, it makes it really hard to diagnose the issue.

Again, yes, but I think that isn't the issue. 

 

Like I said its really hard to follow, too much unnecessary info, as I understood it, they fixed it, *then* changed "something" and since then the pc did not boot once........ 

 

 

You see. Hence its very important to know what they changed (which isn't clear or maybe I just misunderstood, but I tried to parse the info, they repeatedly say it does not boot currently)