[SOLVED] Intermittent CPU debug LED on B450 Tomahawk w/ 3700X

Jul 8, 2019
6
0
10
Parts:
CPU: Ryzen 7 3700X
Motherboard: MSI B450 Tomahawk, BIOS ver. 7C02v18
Memory: 2 x 16GB Ballistix LT @ 3000MHz
Storage: Inland Premium 1TB NVMe
Graphics: GTX 1080 (also tried with a 780)
PSU: Corsair Vengeance 650m

Hello everyone, after building my new PC on Sunday, the motherboard's CPU debug LED comes on quite frequently when powering on, and the system won't POST. However, 1) all fans and hard drives will still power up, and 2) this issue is not consistent. In between several of these failed boots, I've gotten Windows to fully install, do all updates, and download all my essential programs. Also, I've noticed that the issue hasn't happened (yet?) when doing a Windows restart, as that's something that a lot of newly installed programs and drivers ask. This may need more testing, as I've only done a few Windows restarts. The issue has so far only happened when attempting to boot from a full shutdown. In this scenario, it happens 80+% of the time I'd wager.

I'm thinking something came defective (motherboard? PSU? CPU?), or maybe became defective over time (the case is from a 2013 build, I wonder if the front pannel connectors are causing a short). I think this may be the situation because I have since got the build working... sort of. When I don't get the LED, the computer seems to be running just fine. Windows runs fine, no weird shutdowns happening (although I haven't done any stress tests yet; I plan to do these in a few hours). However, my idle temperatures are alarmingly high, at 57C in HWinfo64 with just a couple browser tabs open. This also makes me concerned, especially for the stress testing.

Nothing in particular seems to fix the CPU debug LED. I've disconnected the reset switch on my case, as I've heard bad reset buttons can cause the issue I'm having. But that alone hasn't fixed it. For example, My 2nd most recent power on didn't work. So I unplugged the case's audio out and the power button LED +/-. Now it works. However, last night I tried the same thing and it didn't. Also, the build will often power on successfully after plugging these back in.

It's almost like just fiddling with the motherboard pins will cause the computer to randomly work or not. I'm dumbfounded as to where the problem is.
 
Solution
OK a long post I'm afraid so hope you don't go cross-eyed.

I am concentrating on the RAM here as I do really think it is not a PSU issue due to symptoms being random restarts and failed boots from Windows restarts as opposed a cold boot.
The other thing to look in to is your motherboard voltages which is a separate thing altogether but if they are too high or low you will can get failed boots or high temperatures if the motherboard voltage is too high.
There are masses of posts on these forums on that subject and no doubt other users including myself will help in that respect too but let's eliminate the RAM as the culprit firstly.

Can you please try booting your PC with just 1 stick of RAM at a time to try and eliminate that as the...

CosmicDance

Notable
Jun 11, 2019
414
83
1,040
Do you have a speaker on the motherboard as you can look up the beep codes when it doesn't boot if so?

As for the temperature yes that is quite high for idle.
Often the way thermal paste is applied makes a huge difference to temperatures.
Just a small bit of paste, smaller than a pea, dabbed in the middle of the CPU is sufficient then when you tighten the cooler it spreads out.


Andy
 
  • Like
Reactions: Sydite
Jul 8, 2019
6
0
10
Do you have a speaker on the motherboard as you can look up the beep codes when it doesn't boot if so?

As for the temperature yes that is quite high for idle.
Often the way thermal paste is applied makes a huge difference to temperatures.
Just a small bit of paste, smaller than a pea, dabbed in the middle of the CPU is sufficient then when you tighten the cooler it spreads out.


Andy

Thanks for the assistance, I do have a motherboard speaker; it's currently unattached and I'm unsure if it's still working. I'll plug it in to a working build here in a couple hours to verify, before moving it to the dysfunctional build.

I have used the pea method, yes. The thermal compound is Arctic Silver 5. I am using the stock Wraith Prism cooler. I was wondering if perhaps some combination of the motherboard, CPU, and cooler are not making full contact with each other? Perhaps the CPU is not seated all the way into the motherboard socket, causing the intermittent debug LED, and the cooler is not fully touching the CPU, causing the high temperatures? However, I don't know how plausible this is, since the cooler is clamped down tight. My only suspicion is that the tension rod on the motherboard socket is loose somehow. I suspect this because when I first reseated my cooler and CPU, the CPU came with the cooler (this was using AMD's pre-applied thermal paste -- It looked like way too much and was seeping over the sides somewhat), with the rod still down.
 
Jul 8, 2019
6
0
10
Did you do a clear CMOS?

Yes, after I first encountered the debug issue I was worried I might have borked the BIOS somehow, so I shorted the pins and took out and then reseated the battery. I did this a few times, as well as using the BIOS USB Flashback+ feature after each time to install BIOS version 7C02v18.
 

CosmicDance

Notable
Jun 11, 2019
414
83
1,040
OK a long post I'm afraid so hope you don't go cross-eyed.

I am concentrating on the RAM here as I do really think it is not a PSU issue due to symptoms being random restarts and failed boots from Windows restarts as opposed a cold boot.
The other thing to look in to is your motherboard voltages which is a separate thing altogether but if they are too high or low you will can get failed boots or high temperatures if the motherboard voltage is too high.
There are masses of posts on these forums on that subject and no doubt other users including myself will help in that respect too but let's eliminate the RAM as the culprit firstly.

Can you please try booting your PC with just 1 stick of RAM at a time to try and eliminate that as the problem by attempting to reproduce the failed boot.
If you have overclocked anything e.g. RAM etc in the BIOS then put all these settings back to default for now.
You will need to move a single stick to a different slot e.g. A1/A2/B1/B2 in accordance with your motherboard's configuration for a using a single RAM stick as it will be a different configuration to using two sticks.

If this still fails to boot intermittently then -

You can also run Windows Memory Diagnostic Test by typing 'memory' into the search bar next to the Start button and selecting the Memory Diagnostic App.
This is a low level test that will commence after a reboot and really does test the RAM thoroughly for any read/write/performance errors.

If your RAM passes the diagnostic test then -

I am taking a guess at some of your build Sydite that you haven't listed as there are aspects we all overlook sometimes but don't wish to offend you in any way!
Also on this forum no one generally knows the person's knowledge and experience in terms of PC's and building/tweaking/wrecking etc so just ignore my post if it comes across as patronising.

Is your RAM from an old system with it being 3000 MHZ and you putting it together with a new Ryzen 3700X?
Some motherboards are very fussy about which RAM they will accept and if your specific RAM type is not listed on your mobo's QVL (Qualified Vendor List) then it is not guaranteed to work.
You will be able to access this QVL from the mobo's manufacturer website usually in a PDF document within the Drivers and Utils section of your specific mobo then the RAM is not guaranteed to work or be stable.
It's not to say it will not work but it if it isn't listed but it just isn't guaranteed to or will not be fully stable as a lot of people including myself use RAM that is not in their QVL for their motherboard.
The reason I bought RAM that is not listed in my QVL is down to availability at the time.
I researched similar builds to mine with the same mobo and RAM by looking at benchtest results online of the UserBenchMark app that were from tried and tested configurations.

As for the CPU temperature you can physically push the cooler harder against the mobo while it is running in Windows to see if it is a mounting issue.
Just boot without the side case on, hold the outside of the PC firmly and push gradually firmer on the cooler by its fins.
This will close any gaps that are present to a great extent and help you hopefully eliminate that issue.
near in mind with the side of the PC off your temperatures are very likely to fall anyway initilally so observe the idle temp then proceed to push the cooler firmly.

In Windows select the Power Saving Plan to see how your idle temprature reacts.
This lowers my Ryzen 2700X to a maximum speed of 2200 MHZ from its 3700 base.
The idle temp also drops to 34c with this as the CPU voltages consequently drop but again it goes back to your BIOS voltages and how they are configured as they only stay within the set voltage ranges of the BIOS.
Balanced Power Plan is the usual all round use for Windows and gaming.
You may have to alter some advanced settings within the Power Saving Plan such as turn monitor/hdd off or on to suit your needs as they are designed with battery powered laptops in mind.

Which cooler model do you have by the way and is it new or carried over from an old build?
Check the fins for dust and observe the RPM using the HWinfo APP which is reliable monitoring program for all aspects of PC s including the temperature.
You can turn your fan curve up in your BIOS so that you have a much faster fan RPM at any given temperature.
In your case you can change your curve so that at up to 60c your fan spins much faster as you are idling at 57c.
CPU coolers are sometimes set on a low RPM by the BIOS when you look at the fan curve and don't always keep up with the heat being produced by the CPU.
This is obviously more so with stock coolers and older ones which may have worn out.
The cooler that came with my Ryzen 2700X has a small switch on the side marked L/H for low and high.
When set to high gives much higher RPM when it gets hotter so knowing exactly what your hardware is and what it can do does help you to keep things in check.

With your mobo being a B450 series the same as mine you need the very latest BIOS update in order to have full Ryzen 3700X compatibility.
These have only just appeared recently on my manufacturer's website - ASUS ready for the new Ryzen 3000 series so it will be dated around June 2019 time.

The final point is just an observation but the new Ryzen 3000 series optimal RAM speeds seem to be around the 3600 MHZ mark.
I have 3200 MHZ RAM and would run a new Ryzen 3000 with this myself but obviously sacrifice some performance if I were to upgrade from my current 2700X.
So your 3000 RAM will be holding your CPU's full potential back to a degree but not causing failed boots due to the lower clock speed if it is compatible.
These Ryzen 2000 and 3000 series really do benefit from faster RAM speeds.
Again, the RAM is obviously limited to your mobo's QVL for definite stability.

I know this is going off-topic from your 2 issues which are the failed boots and overheating.

Here is a link to show memory scaling and how it affects CPU performance for the Ryzen 2000 series but it is the same principle for the 3000 series and will no doubt be benchmarked in greater detail as and when people start building wih this new generation of chips -

https://www.legitreviews.com/ddr4-m...ryzen-7-2700x-on-the-amd-x470-platform_205154

Andy
 
Last edited:
  • Like
Reactions: Sydite
Solution
Jul 8, 2019
6
0
10
The other thing to look in to is your motherboard voltages which is a separate thing altogether but if they are too high or low you will can get failed boots or high temperatures if the motherboard voltage is too high.
I've taken some pictures inside my BIOS here to show my XMP settings, my voltages, and fan curve settings. Please take a look here.

You can also run Windows Memory Diagnostic Test by typing 'memory' into the search bar next to the Start button and selecting the Memory Diagnostic App.
This is a low level test that will commence after a reboot and really does test the RAM thoroughly for any read/write/performance errors.
I haven't tried this yet, I will do this now. I am not sure if there is an issue with the RAM however -- see next paragraph.

Is your RAM from an old system with it being 3000 MHZ and you putting it together with a new Ryzen 3700X?
Some motherboards are very fussy about which RAM they will accept and if your specific RAM type is not listed on your mobo's QVL (Qualified Vendor List) then it is not guaranteed to work.
I'm using two Crucial Ballistix LT 16GB modules purchased new about two weeks ago, model number BLS16G4D30AESB. Unfortunately, I could not find these sticks on the QVL, under "Memory by RX-3X00", "2X00", or "1X00". However, before putting these DIMMs into this new Ryzen 3000 build, I had tested them for a week in an Intel i5 build. That machine was powered on for most of the week, which included some gaming and heavy web browsing.

As for the CPU temperature you can physically push the cooler harder against the mobo while it is running in Windows to see if it is a mounting issue.
Just boot without the side case on, hold the outside of the PC firmly and push gradually firmer on the cooler by its fins.
This will close any gaps that are present to a great extent and help you hopefully eliminate that issue.
near in mind with the side of the PC off your temperatures are very likely to fall anyway initially so observe the idle temp then proceed to push the cooler firmly.
...
Which cooler model do you have by the way and is it new or carried over from an old build?
I'm using the stock Wraith Prism as my CPU cooler. I was hoping to use this one for a long time, before I decide to get into overclocking. I've double-checked the latch on the cooler, it seems secure and as tight as it can go. I have now attempted to push the cooler down to see if I can make better contact with the CPU. I did not notice any "give", it feels as though the cooler is very snug against the processor.

You can turn your fan curve up in your BIOS so that you have a much faster fan RPM at any given temperature.
In your case you can change your curve so that at up to 60c your fan spins much faster as you are idling at 57c.
CPU coolers are sometimes set on a low RPM by the BIOS when you look at the fan curve and don't always keep up with the heat being produced by the CPU.
This is obviously more so with stock coolers and older ones which may have worn out.
The cooler that came with my Ryzen 2700X has a small switch on the side marked L/H for low and high.
When set to high gives much higher RPM when it gets hotter so knowing exactly what your hardware is and what it can do does help you to keep things in check.
I have now switched from L to H on the cooler. From the sounds of things, the CPU fan is reaching higher RPMs much more often now. Also, could you look at the image album I linked above and tell me if my new fan curve settings look okay? It definitely seems more effective now, but it sounds quite erratic and the temperatures are jumping around moderately, even at idle. I feel like there's room for improvement with these fan curve settings.

In Windows select the Power Saving Plan to see how your idle temperature reacts.
...
You may have to alter some advanced settings within the Power Saving Plan such as turn monitor/hdd off or on to suit your needs as they are designed with battery powered laptops in mind.
I am now using this setting. Average idle temperatures are now much better, at about 40C. However, all cores are running at 2.2GHz. I built this machine with gaming in mind, so I am concerned that this could negatively impact performance in games.

Additionally, I have installed the latest chipset drivers, so there are also Ryzen balanced, power saver, and high performance power plans. I tested balanced and power saver. Both of these seemed to give worse temperatures and were much more erratic than the Windows defaults. I have switched back to Power saver until I can get temperature on Balanced to be acceptable.

With your mobo being a B450 series the same as mine you need the very latest BIOS update in order to have full Ryzen 3700X compatibility.
These have only just appeared recently on my manufacturer's website - ASUS ready for the new Ryzen 3000 series so it will be dated around June 2019 time.
As of the time of this posting, I have the latest BIOS version for the B450 Tomahawk, version 7C02v18. This version is dated 6/24/2019.
However, I noticed that the support page for my motherboard highly recommends updating AMD chipset drivers as well. I have now done this. Could this newer chipset driver help in my current situation in any way?
 
Jul 8, 2019
6
0
10
Additionally, I had the CPU debug LED issue happen twice during this process. Once after restarting Windows (the chipset drivers asked me to restart). So I powered down and then powered back up. This time, the motherboard went through each of the debug LEDs one by one before POSTing (I've interpreted this as functioning correctly, but I'm honestly not sure). I booted into the BIOS and made the changes as shown in the image album I linked.
I saved the changes and rebooted. CPU debug LED issue again. Power down, power up. Motherboard goes through each debug LED one by one again, and then proceeds to boot into Windows.

It seems like the CPU debug LED issue is more frequent when powering on from a cold start. I've had it happen 10 times in a row, only to work the 11th time. But when restarting after having just used the computer successfully, it's possible it either won't happen at all or will just happen once, maybe twice, before booting successfully.
 

CosmicDance

Notable
Jun 11, 2019
414
83
1,040
If you connect the mobo speaker (presuming your other mobo has one you can use to connect to the AMD one) you will get error beeps which you can then look up to ascertain the problem.
Otherwise it is elimination of possibilities without being definite.

Download HWinfo to monitor your exact temps, fan speed, voltages etc in Windows.
In your BIOS the 1.48 volts looks fine as that is the peak amount.
What the CPU uses in Windows will vary according to load.

RAM may work in your other mobo without issues but it may not necessarily work with the AMD mobo configuration you are using.

Windows Power Saver is just a temporary mode I use to lower clocks and voltages but, like yourself, I meed full power for gaming hence the Balanced plan.
Obviously at the moment this isn't feasible due to your CPU temperature.

I use a Wraith Prism on my Ryzen 2700X on the H setting and it keeps it cool under load and idle.
 
  • Like
Reactions: Sydite
Jul 8, 2019
6
0
10
If you connect the mobo speaker (presuming your other mobo has one you can use to connect to the AMD one) you will get error beeps which you can then look up to ascertain the problem.
Otherwise it is elimination of possibilities without being definite.

Download HWinfo to monitor your exact temps, fan speed, voltages etc in Windows.
In your BIOS the 1.48 volts looks fine as that is the peak amount.
What the CPU uses in Windows will vary according to load.

RAM may work in your other mobo without issues but it may not necessarily work with the AMD mobo configuration you are using.

Windows Power Saver is just a temporary mode I use to lower clocks and voltages but, like yourself, I meed full power for gaming hence the Balanced plan.
Obviously at the moment this isn't feasible due to your CPU temperature.

I use a Wraith Prism on my Ryzen 2700X on the H setting and it keeps it cool under load and idle.

Andy, firstly: my apologies on not getting back to you in a while. I've been silent as I was browsing various forums and reading others' similar experiences. I wanted to learn more before I reported back. I even went out and bought an X570 board to see if maybe my suspicion about my B450 Tomahawk was correct.

I just assembled my parts on the X570 board a couple hours ago, and so far I'm pleased to say that everything is working fully. No more failing to POST, and the temps are lower, though still a bit more than I'd like on Windows/Ryzen balanced modes. Hopefully this gets better with time, though I may simply need better airflow for my system.

My conclusion at the moment is that there is something faulty with my B450 Tomahawk, and/or the latest BIOS is not functioning correctly for my 3700X chip.

For now, I'll mark this as solved and come back if I experience these issues again. Thanks!
 
Last edited:

CosmicDance

Notable
Jun 11, 2019
414
83
1,040
Great to hear that you are up and running with your new set up!
That mobo is far superior with a colossal increase in the PCI bus speed for things like RAM, SSD drive in the M2 slot and new faster GFX cards utilising PCI 4.0 and also built with this new 3rd gen Ryzen in mind.

I use a B450 mobo for my Ryzen 2700X but the latest BIOS and board's functionality are just a BETA stage for new Ryzen CPU s.

Thanks for posting your resolve as it's good to know how to effectively get your new CPU working properly for yourself and helps many others too.

No doubt the manufacturers will be releasing updated BIOS files and AMD updating chipset drivers for things like controlling CPU temperature more effectively.

You can always change your CPU's fan profile to spin faster in your BIOS or Ryzen Master etc.
If it came with the stock Wraith Prism Cooler that AMD bundles it also has a little switch on it marked L/H.
Switchimg it to H (high) gives a huge increase in the fan's max RPM.
Mine now spins up to 3800 RPM on full 100% if I want it which helps cooling a lot.