I blew up my 3600

Legendary Gamer

Legendary Gamer

[H]ard|Gawd
Joined
Jan 14, 2012
Messages
1,590
I figured some of you would find this humorous.

I had achieved what I believed was a stable, all core, OC of 4,250 Mhz on my 3600. System Specs:

AMD 3600 6 Core/12 Thread Processor set to 1.42V on the core
Gigabyte X470 Aorus Ultra Gaming
32 Gigs of RAM running at 3600 Mhz, RAM set to 1.35 (unstable at 1.2)
1 TB Inland M.2 Drive
1 TB SATA SSD using as a Swap Drive
EVGA 2080Ti Black ED
Seasonic 850 Gold PS
240 Cooler Master Liquid Cooler with 4 Fans in Push pull configuration

Well ventilated Blowswill tower case with the CPU cooler mounted on top, three front 140mm fans and one rear fan. Front and air ingress / exhaust was a ventilated, clean ,screen to catch dust.

The scenario:
I was playing a modded session of Battletech (Hydes Rim + other modes and a shader pack), running it at 1080P, all options for graphics set to max. In the middle of my third mission the system blue screened "Windows has shut down unexpectedly" and the system never posted again after that.

I think I aced the motherboard. My money is currently on me roasting a capacitor or VRM.

Motherboard fires up, RGB functions, system is completely non-responsive even after repeated attempts to clear the CMOS.

I don't think the processor roasted, but I can't be certain yet. I sold my house and my only spare MB is a Tmoahawk B350 in storage, in my home server. If the 3600 survived... it's my new home server CPU.

My family members said they smelled something burning downstairs... Lol.

Thus ends my AMD saga, I put the RAM back into my Z370 MB and fired my Intel 9600 right back up with no issues. So at least the RAM is fine. I had picked up a an X570 Asus Prime-P and that MB failed to post. It has no diag lights on it so after a number of hours fighting with it I gave up and sent it back. It was a helluva deal at 135 bucks, too bad it didn't work.

Moral of the story here is... don't go too cheap on your MB or you may very well end up with shit on fire. ;)
 
Last edited:
Well i have Aorus B450 Pro... goes up to 1.5v on CPU on auto settings. Its just about luck. Even costliest mobo can fail.
 
amittalkin said:
Well i have Aorus B450 Pro... goes up to 1.5v on CPU on auto settings. Its just about luck. Even costliest mobo can fail.
Click to expand...
I suppose so, however, if I recall correctly the B450 Pro has a superior VRM. The reviews even concur on that fact. The Ultra Gaming is a piece of hot garbage, as far as I can tell. Others have said that they cannot OC on it, I didn't have that problem. I will do some close up looks at the motherboard tonight.

The 2nd Gen (7nm) Ryzen parts can draw a significantly higher amount of power than you might suspect with their "TDP of 65Watts", which doesn't mean shit after you start OC'ing them. It would appear that the low end board VRM's really aren't up to the task of pushing the new chips.

Compare that to your run of the mill Intel MB like the pair of Z370's I am rocking OC'ed Intel processors on (the 9600 at 5 Ghz) all day long with no issues and the motherboards were dirt cheap. Doesn't work so well with AMD's new gen of processors as far as I can tell. Not to mention, I think I was pushing my luck with a 3600 at 4.25 all core, all the time. Most reports put it's max OC around 4.1. We shall see when I drop the 3600 into my B350. If It doesn't post, then it wasn't the X470.
 
Last edited:
HardUp4HardWare said:
Yikes, so the MB was brand new?
Sorry, dude, has to be bad luck, I guess. That isn't a bad MB.
Click to expand...
That's good to know, yeah, it was new. Some of the other reviews out there of the board had said it wasn't all that great. I picked it up for a hundred bucks during the Amazon Prime sale so I'm not out all that much. I can see what warranty service they may provide... however, I was cranking the CPU so I'm not certain if they will fix it.
 
Legendary Gamer said:
That's good to know, yeah, it was new. Some of the other reviews out there of the board had said it wasn't all that great. I picked it up for a hundred bucks during the Amazon Prime sale so I'm not out all that much. I can see what warranty service they may provide... however, I was cranking the CPU so I'm not certain if they will fix it.
Click to expand...

A highly overclocked 3600 draws a 100watts? If that popped the board that's terrible.
 
tom_ozahoski said:
A highly overclocked 3600 draws a 100watts? If that popped the board that's terrible.
Click to expand...
Yeah, the reviews I had seen on the board said that the board would not OC. I had no issues with that at all. Overclocking was simple-ish. The BIOS is a PITA to find the proper areas where you have to enter your values. The fields you need to actually enter CPU core voltage increase are done in "mv". So you have to enter 1400 in order to get 1.4 V for example, in the OS. So, it's a but confusing and the granularity of the settings is odd. The values don't always accept certain granular ranges ...So I wonder why the hell they even control the CPU voltage in MV in the first place if you can't even set the damn voltage other than a couple locked values.

I am guessing that, under load, the CPU's power draw spiked and it took out a faulty VRM or Cap on the board (Either that or one of the CPU cores died spectacularly under too high a frequency. I say that because only one of the processors cores per die tends to actually be decent/capable of higher clocks, the rest are shit on these AMD 7nm chips). I will see if I can isolate the cooked cap/vrm tonight. I didn't see anything last night but it was a cursory inspection. I simply picked up my Z370 board I had pulled from that case to mount the X470 into and dropped the Intel board and processor into my HAF case that used to house the AMD system. Damn thing fired right up with no issues with the RAM and Video Card from the AMD build. So, now I'm looking at MB and CPU as the cuplrit(s).
 
Legendary Gamer said:
Yeah, the reviews I had seen on the board said that the board would not OC. I had no issues with that at all. Overclocking was simple-ish. The BIOS is a PITA to find the proper areas where you have to enter your values. The fields you need to actually enter CPU core voltage increase are done in "mv". So you have to enter 1400 in order to get 1.4 V for example, in the OS. So, it's a but confusing and the granularity of the settings is odd. The values don't always accept certain granular ranges ...So I wonder why the hell they even control the CPU voltage in MV in the first place if you can't even set the damn voltage other than a couple locked values.

I am guessing that, under load, the CPU's power draw spiked and it took out a faulty VRM or Cap on the board (Either that or one of the CPU cores died spectacularly under too high a frequency. I say that because only one of the processors cores per die tends to actually be decent/capable of higher clocks, the rest are shit on these AMD 7nm chips). I will see if I can isolate the cooked cap/vrm tonight. I didn't see anything last night but it was a cursory inspection. I simply picked up my Z370 board I had pulled from that case to mount the X470 into and dropped the Intel board and processor into my HAF case that used to house the AMD system. Damn thing fired right up with no issues with the RAM and Video Card from the AMD build. So, now I'm looking at MB and CPU as the cuplrit(s).
Click to expand...

Yeah, that totally sucks. I think you are 100% correct in your guess.
Having just bought the GIGABYTE X570 AORUS PRO WiFi, I hope this doesn't happen to me. Hope your CPU is ok.
 
I wonder if this pic of the VRM under OC load is indicative of what happened.

https://www.guru3d.com/articles_pages/gigabyte_x470_aorus_ultra_gaming_review,18.html




index.png
 
HardUp4HardWare said:
I wonder if this pic of the VRM under OC load is indicative of what happened.

https://www.guru3d.com/articles_pages/gigabyte_x470_aorus_ultra_gaming_review,18.html




View attachment 197947
Click to expand...
Nice find on that article! I have a suspicion that the 3600 was drawing as much if not more than this processor in achieving the stable OC I had (or at least seemingly stable until it died). My guess is that there is some, actual, reason why the VRMs of the X570 boards are quite a bit more robust on average and you even see rather strong ones on low end boards like the ASUS Prime-P and TUF boards. Overclocking on these new chips takes a serious toll on the VRM vs the previous initial Ryzen release and it's refresh.

I have to do some more digging into this tonight.
 
I didnt find mention of how hot your CPU was under heavy load.
I would do a sniff test through a straw to try and locate exactly what burned if you cant see it.
 
Nenu said:
I didnt find mention of how hot your CPU was under heavy load.
I would do a sniff test through a straw to try and locate exactly what burned if you cant see it.
Click to expand...
You know, I was totally going to mention I was going to do a sniff test tonight but I thought it would sound ridiculous in the forum. Thank you for that, I will consider this proof that I was not insane in my thought process! I like the straw idea, to focus the snorting of ozone on the MB. I will let you know if it kills me (y)
 
Replace your MB and all will be fine again and don't OC any more. In case the 3600 may be fried, take a cheap one with Zen 2 support in BIOS (if you don't have an old one). Asus seem to have good compatibility with Zen 2. For instance an A3320m-k that supports the 3900X. This costs $50. I suppose you have 2 sticks of 16GB RAM (because the MB has only 2 slots for RAM). Frankly the gain by OC is minimal especially on Zen 2 and your hardware will last forever. Put your old RAM at 2400 to test everything is working, supposing your RAM is still good. My bet is you won't see any noticeable difference.
Overclocking is kind of a stupid behaviour.
 
  • Like
Reactions: travm
like this
travm said:
Lol, if you don't have a straw a rolled up $100 bill works too.
Click to expand...
You know, I was thinking some recreational powder might add to the experience. Alas, I never really got into such things.

Mainboard shows no signs of fried components. Something I wasn't thinking of a week ago was that there seemed to be a low level electronic burning smell coming from the chassis from the top exhaust fans. I had dismissed it because five months ago I was a full time smoker. I just assumed it was a residual smell from the chassis, from that time. I suspect that the motherboard was "on fire" in a manner of speaking, and it wasn't just a single component.
 

Attachments

  • Full Shot.jpg
    Full Shot.jpg
    466.6 KB · Views: 0
  • FS Closer.jpg
    FS Closer.jpg
    430.1 KB · Views: 0
  • VRMS.jpg
    VRMS.jpg
    333.7 KB · Views: 0
  • VRMS2.jpg
    VRMS2.jpg
    539 KB · Views: 0
Legendary Gamer said:
I had achieved what I believed was a stable, all core, OC of 4,250 Mhz on my 3600.

[...]


Moral of the story here is... don't go too cheap on your MB or you may very well end up with shit on fire. ;)
Click to expand...


Hmm. Sounds to me like the real moral of the story was... that wasn’t a stable overclock.

don’t count your GHz until it’s hatched?

one man’s stable overclock is another man’s “something is burning in the basement”?

a bird in hand is worth two in the water loop?


At any rate, sorry to hear you popped metaphorical blue smoke. It’s never fun when a system bites it hard like that.
 
Legendary Gamer said:
You know, I was thinking some recreational powder might add to the experience. Alas, I never really got into such things.

Mainboard shows no signs of fried components. Something I wasn't thinking of a week ago was that there seemed to be a low level electronic burning smell coming from the chassis from the top exhaust fans. I had dismissed it because five months ago I was a full time smoker. I just assumed it was a residual smell from the chassis, from that time. I suspect that the motherboard was "on fire" in a manner of speaking, and it wasn't just a single component.
Click to expand...
Sounds like your board had a slow leak in the magic smoke seal. I bet it was manufacturing defect, although maybe not bad enough to cause issues under normal expected load.
 
Had a shitty psu kill an msi b450 vrm running a 2600x. Failed vreg was under the heat spreader so you couldn't see It from outside. Ensure you use a well fitting screw driver and not spin out of It or damage the anodisation. Push down lots :)
 
Legendary Gamer said:
I suppose so, however, if I recall correctly the B450 Pro has a superior VRM. The reviews even concur on that fact. The Ultra Gaming is a piece of hot garbage, as far as I can tell. Others have said that they cannot OC on it, I didn't have that problem. I will do some close up looks at the motherboard tonight.

The 2nd Gen (7nm) Ryzen parts can draw a significantly higher amount of power than you might suspect with their "TDP of 65Watts", which doesn't mean shit after you start OC'ing them. It would appear that the low end board VRM's really aren't up to the task of pushing the new chips.

Compare that to your run of the mill Intel MB like the pair of Z370's I am rocking OC'ed Intel processors on (the 9600 at 5 Ghz) all day long with no issues and the motherboards were dirt cheap. Doesn't work so well with AMD's new gen of processors as far as I can tell. Not to mention, I think I was pushing my luck with a 3600 at 4.25 all core, all the time. Most reports put it's max OC around 4.1. We shall see when I drop the 3600 into my B350. If It doesn't post, then it wasn't the X470.
Click to expand...

sounds like same thing that happened to me with asrock 990fx ex4. VRM wasn't able to handle my OC. Didn't know as much about VRM's as I do now. I am attatching an image of the AMD X370/X470/X570 vrm data spreadsheet. very interesting stuff.

edit: cant get image to show right here. so i linked to a copy on my onedrive. may have to click "view original" to see it properly or just give it a second to load when you zoom in. https://1drv.ms/u/s!ArP5Omaho6n3hkzpgi62x3p1x20R?e=nmPJGh

i also thought about getting an x470 for my 3800X but after looking at that i'm glad i didn't. X570 for sure.
 
Last edited:
d3athf1sh said:
sounds like same thing that happened to me with asrock 990fx ex4. VRM wasn't able to handle my OC. Didn't know as much about VRM's as I do now. I am attatching an image of the AMD X370/X470/X570 vrm data spreadsheet. very interesting stuff.

edit: cant get image to show right here. so i linked to a copy on my onedrive. may have to click "view original" to see it properly or just give it a second to load when you zoom in. https://1drv.ms/u/s!ArP5Omaho6n3hkzpgi62x3p1x20R?e=nmPJGh

i also thought about getting an x470 for my 3800X but after looking at that i'm glad i didn't. X570 for sure.
Click to expand...

I do think that the x570 is a much better choice for many reasons but for OC especially.
 
Yeah, I have had some good experiences with Gigabyte over the years and was really liking their motherboards. The Z370P D3 boards I am running a 9600 and 8600 on are essentially bulletproof. The VRMs on those designs have less than half the caps and chokes that the X470 Ulrta Gaming does and they easily push 14nm+++++++++++++++++++++++ chips into the "Hell's hot" range and do so with a smile on their faces... I think I must have had a slow leakage VRM issue, because it took over a week to cook that motherboard and I was pushing the voltage up the entire time and putting that processor under load regularly. deruberhanyok is absolutely correct. It wasn't a stable OC. Makes me wonder if I can replicate those processor speeds on a higher end Mobo with a serious VRM... But running a sub 200 buck chip on a 500+ dollar motherboard sounds ridiculous to me.

Nobu definitely magic smoke, since I never visibly saw any. It was like my system was killed by a motherboard ninja or the Spanish inquisition. Honestly, who would attribute playing Battletech to the death of their rig? :grumpy:

N4CR nice to see you again my friend! Yeah, when I get a chance I will pull the VRM sinks off and take a much closer look.

d3athf1sh thank you for that VRM Spreadsheet... That's eye opening!
 
HardUp4HardWare said:
I do think that the x570 is a much better choice for many reasons but for OC especially.
Click to expand...
I had picked up an ASUS Prime-P, it was well reviewed and carries the same VRM as the STRIX and the TUF. However, that board (X570) has no diagnostic indicators and the speaker even didn't kick out any errors. System fired, but failed to post with RAM, without RAM, with another processor, swapping video, power and different RAM. I gave up, sent it back and put the X470 back in. I probably should have gone with the TUF. However, I was liking the fact that the ASUS Prime-P ran 7-8c cooler in all tests. If it had worked, it would have been a nice mobo for 139 bucks.
 
Legendary Gamer said:
I had picked up an ASUS Prime-P, it was well reviewed and carries the same VRM as the STRIX and the TUF. However, that board (X570) has no diagnostic indicators and the speaker even didn't kick out any errors. System fired, but failed to post with RAM, without RAM, with another processor, swapping video, power and different RAM. I gave up, sent it back and put the X470 back in. I probably should have gone with the TUF. However, I was liking the fact that the ASUS Prime-P ran 7-8c cooler in all tests. If it had worked, it would have been a nice mobo for 139 bucks.
Click to expand...

Yeah, I have always gone with the cheapest ASUS MB I could get in the past. This time I got the Gigabyte Aurous X570 pro wifi which was only around $250. Granted I am not overclocking so I don't know how that would go, but it seems like a great MB.

I will say this, choosing a MB is probably the most difficult part of the build. Good MB manufacturers make some shitty MBs. I almost got an EVGA board and read about how problematic it was and I was shocked. I thought EVGA was top tier, but again, the low end of the top tier is still shitty sometimes.
 
is 4.25 at 1.42v really that good of an all core oc on Ryzen 3000? That voltage seems a bit high for being manually set.

From what I've seen, most don't recommend going over 1.35v for manually set voltage.

I'm currently at 1.325v @ 4200 on my 3700x.
 
Legendary Gamer said:
I had picked up an ASUS Prime-P, it was well reviewed and carries the same VRM as the STRIX and the TUF. However, that board (X570) has no diagnostic indicators and the speaker even didn't kick out any errors. System fired, but failed to post with RAM, without RAM, with another processor, swapping video, power and different RAM. I gave up, sent it back and put the X470 back in. I probably should have gone with the TUF. However, I was liking the fact that the ASUS Prime-P ran 7-8c cooler in all tests. If it had worked, it would have been a nice mobo for 139 bucks.
Click to expand...

yeah as soon as i get the money i'm picking up the TUF X570 it's only got 4 real phases but each one has 3 50A mosfets. not sure if you look at buildzoid's youtube channel but he goes in depth on the X570 vrm situation. Oh and if you are looking at the ~$200 X570 mb's you really need to watch both of these too:




he's got a bunch of other x570 vids on channel
 
Last edited:
Yes, those were my principal research points prior to picking up the X570 Asus Prime-P. It's all I need in a board, just a real PITA for configuration if anything goes wrong. I didn't like that the TUF runs about 8c hotter for no reason that they could understand other than the fact that (I suspect) it might have fancier but shittier coolers on it. The PRIME-P, the TUF and the STRIX all use the same VRM and it's supposed to be more than enough to power and OC the 3950 even at 4 phases. The VRM design has shown up in a number of higher end Intel boards.
 
yeah when you put them under load (OC) the TUF board holds up the best. probably due to a thicker pcb. noticed that when i blew out my asrock 990FX and then got an ASUS Sabretooth 990FX (TUF) it was a lot thicker and MB temps were way lower.
 
Xero717 said:
is 4.25 at 1.42v really that good of an all core oc on Ryzen 3000? That voltage seems a bit high for being manually set.

From what I've seen, most don't recommend going over 1.35v for manually set voltage.

I'm currently at 1.325v @ 4200 on my 3700x.
Click to expand...
I wasn't really OC'ing by posted reports
(from other people). I was over clocking by instinct. I wanted to see what this puppy was actually capable of and a lot of people don't push their review samples hard enough to really see what these can do. The fact that I was able to even get this processor to 4,250 Mhz, and it wasn't overheating. Too bad the MB couldn't sustain my OC. I have always kind of said "Opinions be damned" and pushed the envelope.

The Gaming Ultra, essentially, died horribly under this pursuit. I find it interesting because my Tomahawk B350 was pushing my Ryzen 1700 at 4.1 / 1.42 for almost a year. The Gaming Ultra died a lot faster than I would have expected it to (The B350 of mine still works).
 
Legendary Gamer said:
Yes, those were my principal research points prior to picking up the X570 Asus Prime-P. It's all I need in a board, just a real PITA for configuration if anything goes wrong. I didn't like that the TUF runs about 8c hotter for no reason that they could understand other than the fact that (I suspect) it might have fancier but shittier coolers on it. The PRIME-P, the TUF and the STRIX all use the same VRM and it's supposed to be more than enough to power and OC the 3950 even at 4 phases. The VRM design has shown up in a number of higher end Intel boards.
Click to expand...

if you look at that spreadsheet the Prime-P uses diffent config than the TUF. You are thinking of the Prime-Pro. 2 different boards. sux it's like they try to confuse people sometimes with their part numbers.
 
d3athf1sh said:
yeah when you put them under load (OC) the TUF board holds up the best. probably due to a thicker pcb. noticed that when i blew out my asrock 990FX and then got an ASUS Sabretooth 990FX (TUF) it was a lot thicker and MB temps were way lower.
Click to expand...
Yeah, the only issue is that the X570 Prime-P is 8c lower than the TUF... So, it doesn't make sense to me and it didn't to hardware unboxed either. So much so that they ran exhaustive testing on the X570 Prime-P and TUF and verified that the Prime-P ran cooler in all tests. Crazy for a cheaper, technically lower quality PCB and heat spreaders.
 
You must log in or register to reply here.
Back
Top