5950x and PBO overheated something or busted custom loop?

[A Computer Guy]

So I was tinkering with PBO on my 5950x for the first time tonight, nothing special just tweaking PPT, TDC, and EDC limits, everything else auto.
Wasn't expecting anything special to happen other than a bit higher CPU temps and maybe slightly better clocks and after some time started to notice a strange smell.
I didn't notice the smell right away because my PC is sitting on a rack above my monitors and it was perhaps 15-30 minutes into an all core testing already.
I would say it did NOT smell like typical electronics overheating.

At the time I was monitoring CPU's temps and my UPS wattage. Nothing seemed particularly abnormal. CPU was popping up to 89c (which is higher than usual but not unexpected) and I was pulling under 360 watts system total (don't remember the exact number) according to the UPS. Two temps I didn't manage to get were VRM temps and Water temps before hastily turning everything off in panic mode to check for fire hazards in the house.

CPU:
- Ryzen 9 5950x

Motherboard:
- ASRock X470 Master SLI/AC (revision 1.02, BIOS/UEFI version P4.53)

Cooling:
- EK-FB ASRock X470 Gaming K4 RGB Monoblock, EK-Quantum Reflection PC-O11 D5 PWM, EK-CoolStream PE 360
(monoblock includes VRM cooling)

What's more likely?
1) Melting the thermal pads on the VRM's ( I have read the VRM's on this motherboard are basically trash, didn't know that before I bought it. I had replaced and used some leftover EK thermal pads from my GPU block in my custom loop maintenance back in June. They were maybe 2yrs old but were in very good condition )
2) Microleak in Radiator ( the most concentrated smell seems to be coming from near the ports of the radiator however that is also above to the VRM's under the monoblock, there is no indication of fluid leakage or loss so far )
3) Component burnout somewhere ( the PC didn't crash or exabit any other issues before shutdown so I kind of doubt this )
4) Overheated thermal paste, perhaps some little excess burning off?

It took a good hour to clear the smell out of my house.
So I'm on my backup PC for now while it sits in the garage airing out.
I suspect #2. EK clear fluid already has a particularly strong smell and I can only imagine what it must smell like when it becomes heated and vaporized.
I'm thinking I'm going to have to do a complete teardown to investigate to be sure.

Anyone want to place their bets.

(edit)

TLDR. No problem. Turns out I was wrong. It was #3. See the thread for some nice pictures.

 

[Toothless]

Get some thermal pads and light them up, compare smells.

I had a power supply smell like waffles and syrup once.

 

[Mussels]

So I was tinkering with PBO on my 5950x for the first time tonight, nothing special just tweaking PPT, TDC, and EDC limits, everything else auto.
Wasn't expecting anything special to happen other than a bit higher CPU temps and maybe slightly better clocks and after some time started to notice a strange smell.
I didn't notice the smell right away because my PC is sitting on a rack above my monitors and it was perhaps 15-30 minutes into an all core testing already.
I would say it did NOT smell like typical electronics overheating.

At the time I was monitoring CPU's temps and my UPS wattage. Nothing seemed particularly abnormal. CPU was popping up to 89c (which is higher than usual but not unexpected) and I was pulling under 360 watts system total (don't remember the exact number) according to the UPS. Two temps I didn't manage to get were VRM temps and Water temps before hastily turning everything off in panic mode to check for fire hazards in the house.

CPU:
- Ryzen 9 5950x

Motherboard:
- ASRock X470 Master SLI/AC (revision 1.02, BIOS/UEFI version P4.53)

Cooling:
- EK-FB ASRock X470 Gaming K4 RGB Monoblock, EK-Quantum Reflection PC-O11 D5 PWM, EK-CoolStream PE 360
(monoblock includes VRM cooling)

What's more likely?
1) Melting the thermal pads on the VRM's ( I have read the VRM's on this motherboard are basically trash, didn't know that before I bought it. I had replaced and used some leftover EK thermal pads from my GPU block in my custom loop maintenance back in June. They were maybe 2yrs old but were in very good condition )
2) Microleak in Radiator ( the most concentrated smell seems to be coming from near the ports of the radiator however that is also above to the VRM's under the monoblock, there is no indication of fluid leakage or loss so far )
3) Component burnout somewhere ( the PC didn't crash or exabit any other issues before shutdown so I kind of doubt this )
4) Overheated thermal paste, perhaps some little excess burning off?

It took a good hour to clear the smell out of my house.
So I'm on my backup PC for now while it sits in the garage airing out.
I suspect #2. EK clear fluid already has a particularly strong smell and I can only imagine what it must smell like when it becomes heated and vaporized.
I'm thinking I'm going to have to do a complete teardown to investigate to be sure.

Anyone want to place their bets.

2023 update: This kept happening to me with EK coolant, once i went alphacool radiators and coolant i've had zero problems.

I've had my custom loop go bad maybe 5 times in the last year alone
It happens

CPU hit 90C at stock with PBO off.
I was grump. Rebuilt loop, reversing in and out on radiator accidentally - crud started accumulating in CPU block. temps went back up.

I ran a primochill cleaner in my loop and within maybe 5 seconds my 'clean' loop with clear distilled water turned into green sludge
These are in order - 'clean' fresh distilled to sludge in seconds, to semi sludge to clean again, as the cleaner did its thing over ~8 hours. Drained the loop and CPU is now at 60C gaming vs the 90C - after the fact it became obvious that the microchannels in my rad were clogged or partially clogged.

I had the thermal pad on my x570 chipset randonly dry out too and in the space of a week went from ~1000RPM 60C to 80C 3000RPM
Less than a year old at the time, so yeah - new thermal pads might be a good idea - remember that monoblocks are EXTREMELY sensitive to the wrong height VRMs

The lesson here: cooling stuff goes bad in weird ways.

AIO's slowly evaporate tiny amounts of coolant through their hoses, custom looops are even more likely - heat can make tiny amounts of things smell bad briefly. thermal pads have oil, thermal paste, coolant, etc.

 

[Zach_01]

So I was tinkering with PBO on my 5950x for the first time tonight, nothing special just tweaking PPT, TDC, and EDC limits, everything else auto.
Wasn't expecting anything special to happen other than a bit higher CPU temps and maybe slightly better clocks and after some time started to notice a strange smell.
I didn't notice the smell right away because my PC is sitting on a rack above my monitors and it was perhaps 15-30 minutes into an all core testing already.
I would say it did NOT smell like typical electronics overheating.

At the time I was monitoring CPU's temps and my UPS wattage. Nothing seemed particularly abnormal. CPU was popping up to 89c (which is higher than usual but not unexpected) and I was pulling under 360 watts system total (don't remember the exact number) according to the UPS. Two temps I didn't manage to get were VRM temps and Water temps before hastily turning everything off in panic mode to check for fire hazards in the house.

CPU:
- Ryzen 9 5950x

Motherboard:
- ASRock X470 Master SLI/AC (revision 1.02, BIOS/UEFI version P4.53)

Cooling:
- EK-FB ASRock X470 Gaming K4 RGB Monoblock, EK-Quantum Reflection PC-O11 D5 PWM, EK-CoolStream PE 360
(monoblock includes VRM cooling)

What's more likely?
1) Melting the thermal pads on the VRM's ( I have read the VRM's on this motherboard are basically trash, didn't know that before I bought it. I had replaced and used some leftover EK thermal pads from my GPU block in my custom loop maintenance back in June. They were maybe 2yrs old but were in very good condition )
2) Microleak in Radiator ( the most concentrated smell seems to be coming from near the ports of the radiator however that is also above to the VRM's under the monoblock, there is no indication of fluid leakage or loss so far )
3) Component burnout somewhere ( the PC didn't crash or exabit any other issues before shutdown so I kind of doubt this )
4) Overheated thermal paste, perhaps some little excess burning off?

It took a good hour to clear the smell out of my house.
So I'm on my backup PC for now while it sits in the garage airing out.
I suspect #2. EK clear fluid already has a particularly strong smell and I can only imagine what it must smell like when it becomes heated and vaporized.
I'm thinking I'm going to have to do a complete teardown to investigate to be sure.

Anyone want to place their bets.

What was the CPU EDC/TDC/PPT during that test?
According to this your board needs a serious air flow around VRMs at 150A, and its incapable of 200A

AM4 Vcore VRM Ratings v1.4 (2019-11-07)

docs.google.com

Also 90C is the MAX operating limit and its probably throttling at that point. You shouldn't push it to hard.
Too high EDC + (on) high(est) temp can kill your CPU overtime easily. PPT(w) comes second.

You should've ask for some opinions/directions prior to this, IMHO.

 

[ir_cow]

Only time I had a small that bad was when I forgot to plug in the pump lol. Besides burning dust, smell isnt normal.

 

[A Computer Guy]

Get some thermal pads and light them up, compare smells.

I had a power supply smell like waffles and syrup once.

I ran a quick test in Ryzen Master, boom got smoke this time, shutoff.

Well I confirmed it. Melting the thermal pads is likely what happened or maybe component burnout right where the VRM's are on the side of the motherboard.

Guess it's time to do a teardown this weekend and see if there are any signs of visible component damage.
Just before shutoff I saw the VRM spike to 83 but I bet it went higher as the pc froze for a few seconds when Ryzen Master ran it's test.

Womp Womp might be time to order a new motherboard.
5950x and PBO is an obvious fire hazard with my motherboard.

 

[Zach_01]

Also want to point out that higher EDC does not mean higher performance. Higher temp for sure...

For all core loads:
Only higher effective clocks = higher performance, and this can be achieved at lowest possible temp with a reasonable EDC and most likely lower than the stock 140A(EDC).

 

[A Computer Guy]

What was the CPU EDC/TDC/PPT during that test?
According to this your board needs a serious air flow around VRMs at 150A, and its incapable of 200A

AM4 Vcore VRM Ratings v1.4 (2019-11-07)

docs.google.com

That's why I have the monoblock for better VRM heat management. I forgot that spreadsheet was out there but I thought it only applied to all core overclocks which is not really what I was aiming for.
Interestingly I didn't have this problem with PBO with my 3950x but I admit that was several bios updates and versions of ryzen master long ago. Ultimately I decided PBO wasn't worth it then.

Here is what I tried. The defaults are 1000, 420, and 460 but I adjusted these to 250, 180, 220 based on a reddit that recommended forgetting about curve optimizer and simply dialing in EDC to 100% and then adjusting the other values.

Also 90C is the MAX operating limit and its probably throttling at that point. You shouldn't push it to hard.
Too high EDC + (on) high(est) temp can kill your CPU overtime easily. PPT(w) comes second.

Isn't the motherboard or CPU supposed to prevent too much power draw?

You should've ask for some opinions/directions prior to this, IMHO.

I can't argue with that. I have little doubt I probably smoked my board. My bad. Boy do I feel stupid right now and rightfully so.

Note the CPU temp was just fine so I think my CPU is ok but I guess I will have to wait and see.

 

[Mussels]

I ran a quick test in Ryzen Master, boom got smoke this time, shutoff.

Well I confirmed it. Melting the thermal pads is likely what happened or maybe component burnout right where the VRM's are on the side of the motherboard.

Guess it's time to do a teardown this weekend and see if there are any signs of visible component damage.
Just before shutoff I saw the VRM spike to 83 but I bet it went higher as the pc froze for a few seconds when Ryzen Master ran it's test.

Womp Womp might be time to order a new motherboard.
5950x and PBO is an obvious fire hazard with my motherboard.

I mean, start by checking the pads (and the required thickness)
With pads and a block on em, even cheap VRMs should be fine... it's not a 5950x with PBO can blast to 250W



Oof you were not joking, it really does have bad VRMs.
This video does have some really wobbly camera work later on, if you get motion sickness

(The intro has the advertised specs, around 4 minutes he finds out what it really has)

Outright false marketing and claiming different VRM vendors
Claims 'Digi Power 12 phase VRM, 45A choke'
(They now claim intersil 12 phase on the website, but teardowns prove its 4)

Turns out it's a 4 phase from what he describes as a fireworks factory :/

 

[A Computer Guy]

I mean, start by checking the pads (and the required thickness)
With pads and a block on em, even cheap VRMs should be fine... it's not a 5950x with PBO can blast to 250W



Oof you were not joking, it really does have bad VRMs.
This video does have some really wobbly camera work later on, if you get motion sickness

(The intro has the advertised specs, around 4 minutes he finds out what it really has)

Outright false marketing and claiming different VRM vendors
Claims 'Digi Power 12 phase VRM, 45A choke'

Turns out it's a 4 phase from what he describes as a fireworks factory :/

I think fireworks factory confirmed. I may have heard popping. Oof.

 

[Mussels]

My first MSI B450 was like this, 4 phases without even a heatsink on them
At least it was cheap garbage and advertised itself that way

Yours is like that, only pretending its something literally 5x more capable

 

[Ferrum Master]

DO NOT USE EDC PAST 145A

Read up the threads here and other forums. If you do so you kill any kind of boost clock with the latest AGESAs.

 

[A Computer Guy]

I mean, start by checking the pads (and the required thickness)
With pads and a block on em, even cheap VRMs should be fine... it's not a 5950x with PBO can blast to 250W

I did use the recommended thickness per EK but I always had some doubts if I should go a tad thicker. When I did my first maintenance teardown on my custom loop with my monoblock it did look like there was enough pressure on the pads so when I replaced them I used the same size. I had bought a few extra pads anticipating the chance I might need to replace them.

Oof you were not joking, it really does have bad VRMs.
This video does have some really wobbly camera work later on, if you get motion sickness

He had another video on this too.

 

[tabascosauz]

I think fireworks factory confirmed. I may have heard popping. Oof.

iirc essentially it's just a 4-phase with doubled lo-side mosfets. Nearly all of the average/mediocre X470/B450 boards had that setup.

Could accommodate a stock 12- or 16-core on that VRM........but only if it had decent choice of mosfets (ie. 4C029) combined with a big heatsink. Unfortunately ASRock is not new MSI and loves bottom of the barrel and the SM4337/SM4336s on your board are about as bad as it gets. I don't think AM4 boards as a whole have seen any other discrete mosfet with a higher RDSon. As it is, a stock or (maybe PBO) 3700X is about the limit.

I had a 4-phase with much more capable older DrMOS parts (50A IR3556), and when pushing a 3700X over 100W the VRM was pushing 80+. I would trust that Gigabyte boards before the K4.

I'm surprised you made it this far, unless you usually only game and don't run any full loads that max out PPT (all-core CPU rendering). But even still, 5900X and 5950X draw like 100W in most games, which is a lot for this board, and why don't usually recommend anything > 3700X on these 4-phase SInopower boards.

Water handles it well but 200-250W is still pretty brutal. Last time I pushed my 5900X to 210-220W the VRM on my Unify-X was getting a bit warm, and that was I think a 14-phase Vcore with 90A ISL99390s.

 

[A Computer Guy]

DO NOT USE EDC PAST 145A

Read up the threads here and other forums. If you do so you kill any kind of boost clock with the latest AGESAs.

Based on what you said Ryzen Master defaults totally suck. The defaults are 1000 ppt, 420 tdc, and 460 edc.

 

[Mussels]

Based on what you said Ryzen Master defaults totally suck. The defaults are 1000 ppt, 420 tdc, and 460 edc.

That's just unlocked totally, the maximum figures - the same as if you ran an all core OC, effectively

 

[Ferrum Master]

Put some layers

Based on what you said Ryzen Master defaults totally suck. The defaults are 1000 ppt, 420 tdc, and 460 edc.

The Software team with that piece software doesn't communicate with the AGESA god tier coders. It worked on older first 5xxx AGESAs, where you could dump loads of current for basically zero performance gain, they patched it out later on firmware level that nor bios nor ryzen master can trigger those addresses.

I can recommend putting thermal pads or pad a heatsink on the other back side of the motherboard, where VRM are. Reuse the screw holes that holds top heatsinks and add some more on the other side. Some older high end boards did such things in the past. Or layers of pads or one thick one and use the PC case MB plate a heatsink itself.

 

[A Computer Guy]

I've had my custom loop go bad maybe 5 times in the last year alone
It happens

This had been my first custom loop. For 2yrs I did pretty well. Kept up on maintenance, cleaned blocks, radiator. pressure tested. etc...

The lesson here: cooling stuff goes bad in weird ways.

Interestingly as bad as this board was supposed to be it eventually hit just about every checkmark in my exploratory journey coming back to AMD with AM4.
From Zen+, Zen2, Zen3 upgrades and BIOS\UEFI updates were really no problem. I was able to do some easy RAM overclocks. ASRock responded to my inquires with BIOS\UEFI fixes several times including with ECC support. I maxed out CPU, RAM, really had a good experience (despite the AMD growing pains) with this board up until now.

iirc essentially it's just a 4-phase with doubled lo-side mosfets. Nearly all of the average/mediocre X470/B450 boards had that setup.

Could accommodate a stock 12- or 16-core on that VRM........but only if it had decent choice of mosfets (ie. 4C029) combined with a big heatsink. Unfortunately ASRock is not new MSI and loves bottom of the barrel and the SM4337/SM4336s on your board are about as bad as it gets. I don't think AM4 boards as a whole have seen any other discrete mosfet with a higher RDSon. As it is, a stock or (maybe PBO) 3700X is about the limit.

Without the monoblock I wouldn't have even considered putting anything higher than 3800x on it. When I had the 3800x on it without the monoblock it did ok.

I had a 4-phase with much more capable older DrMOS parts (50A IR3556), and when pushing a 3700X over 100W the VRM was pushing 80+. I would trust that Gigabyte boards before the K4.

I'm surprised you made it this far, unless you usually only game and don't run any full loads that max out PPT (all-core CPU rendering). But even still, 5900X and 5950X draw like 100W in most games, which is a lot for this board, and why don't usually recommend anything > 3700X on these 4-phase SInopower boards.

Water handles it well but 200-250W is still pretty brutal. Last time I pushed my 5900X to 210-220W the VRM on my Unify-X was getting a bit warm, and that was I think a 14-phase Vcore with 90A ISL99390s.

Yea I didn't realize how brutal it was going to be. In my mind, maybe perhaps because of outdated and incorrect information, I thought the CPU was supposed to prevent this situation since I was letting the CPU do it's thing to balance the voltage, power, and heat but just letting the CPU know it can try to do more if it can.


So basically what probably happed here is I blew my VRM because the CPU asked for more power and it destroyed itself trying to do so. Do I understand that correctly?

 

[Mussels]

Yea I didn't realize how brutal it was going to be. In my mind, maybe perhaps because of outdated and incorrect information, I thought the CPU was supposed to prevent this situation since I was letting the CPU do it's thing to balance the voltage, power, and heat but just letting the CPU know it can try to do more if it can.

PBO changes those settings and the ones you used in RM disabled them entirely

That said, no one would expect the board to die - just thermal throttle, even with stock cooling

 

[Ferrum Master]

Btw as a fellow EK Monoblock user... what kind of torque you use to tighten everything up? Are you aware of the numbers?

You should do heavy PBO undervolt, limit currents and you should be fine. Do not use any Windows level crap for adjusting mission critical stuff.

 

[tabascosauz]

Without the monoblock I wouldn't have even considered putting anything higher than 3800x on it. When I had the 3800x on it without the monoblock it did ok.

Yea I didn't realize how brutal it was going to be. In my mind, maybe perhaps because of outdated and incorrect information, I thought the CPU was supposed to prevent this situation since I was letting the CPU do it's thing to balance the voltage, power, and heat but just letting the CPU know it can try to do more if it can.

So basically what probably happed here is I blew my VRM because the CPU asked for more power and it destroyed itself trying to do so. Do I understand that correctly?

142W is still in the realm of "doable with more cooling". 200W on hardware like that is just considered playing with fire, lol.

ASRock makes some of the most god awful VRMs, but they are also usually very trigger-happy with over-current protection on the awful boards. Not sure why nothing happened here, but then again the conservative throttling might be more a post-2020 thing for ASRock. Yours is a rather old specimen. Usually all except the cheapest (50A Vishays SiC6xx) DrMOS parts have a thermal shutdown flag and OCP, but I wouldn't count on discrete mosfets (and the godawful PWM controllers they're using here) to have the same.

When it's stock a Ryzen 3000/5000 will draw on a lot of telemetry to protect itself, but I've not yet seen a Ryzen CPU throttle itself to protect a VRM. I don't think the CPU receives info like that, and the hardware here is certainly not new enough/ not smart enough to provide current/temp monitoring. If the power delivery is garbage and has no functioning OCP/OTP protections of its own, not much to stand in the way of a fireworks show, I'm afraid.

If you replace the board, check up thoroughly on the CPU. DDR4 VRM went kaput on my Aorus ITX board a few months ago, and crippled a lot of functions on the SoC side of my 5700G as well.

 

[Zach_01]

Ryzen 3000/5000 has very fine control over just about anything but I've not yet seen a Ryzen CPU throttle itself to protect a VRM. If the power delivery is garbage and has no functioning OCP/OTP protections of its own, not much to stand in the way of a fireworks show, I'm afraid.

I believe that's why TDC exists

Thermal Design Current (“TDC”): The maximum current (amps) that can be delivered by a specific motherboard’s voltage regulator configuration in thermally-constrained scenarios.

 

[A Computer Guy]

PBO changes those settings and the ones you used in RM disabled them entirely

That's a bit different than what I remember. But there is a good chanced perhaps I misunderstood this article many years ago. Or have the rules changed too?

Explaining AMD Ryzen Precision Boost Overdrive (PBO), AutoOC, & Benchmarks | GamersNexus

stub With the launch of the Ryzen 3000 series processors, we’ve noticed a distinct confusion among readers and viewers when it comes to the phrases “Precision Boost 2,” “XFR,” “Precision Boost Overdrive,” which is different from Precision Boost, and “AutoOC.” There is also a lot of confusion...

www.gamersnexus.net

That said, no one would expect the board to die - just thermal throttle, even with stock cooling

Yea melting to death wasn't what I expected.

142W is still in the realm of "doable with more cooling". 200W on hardware like that is just considered playing with fire, lol.

ASRock makes some of the most god awful VRMs, but they are also usually very trigger-happy with over-current protection on the awful boards. Not sure why nothing happened here, but then again the conservative throttling might be more a post-2020 thing for ASRock. Yours is a rather old specimen. Usually all except the cheapest (50A Vishays SiC6xx) DrMOS parts have a thermal shutdown flag and OCP, but I wouldn't count on discrete mosfets (and the godawful PWM controllers they're using here) to have the same.

When it's stock a Ryzen 3000/5000 will draw on a lot of telemetry to protect itself, but I've not yet seen a Ryzen CPU throttle itself to protect a VRM. I don't think the CPU receives info like that, and the hardware here is certainly not new enough/ not smart enough to provide current/temp monitoring. If the power delivery is garbage and has no functioning OCP/OTP protections of its own, not much to stand in the way of a fireworks show, I'm afraid.

If you replace the board, check up thoroughly on the CPU. DDR4 VRM went kaput on my Aorus ITX board a few months ago, and crippled a lot of functions on the SoC side of my 5700G as well.

I'm assuming I'm going to get a new board at this point. What should I look for regarding the CPU potential damage?

Btw as a fellow EK Monoblock user... what kind of torque you use to tighten everything up? Are you aware of the numbers?

You should do heavy PBO undervolt, limit currents and you should be fine. Do not use any Windows level crap for adjusting mission critical stuff.

To the motherboard it is just regular screws. on the outer edges for the VRM I have to be careful not to overtighten or it can warp the board. This was one of the reasons I think EK specification for the pads might not have been thick enough but when I pre-test the mounting and remove the block I check if I am making good enough impressions in the thermal pad to know good contact is being made. Not very scientific I know.

If you mean regarding the block reassembly? I just do it by hand being careful not to overtighten them, approximating the torque I needed to remove them. I didn't have any leaks or cracks in the acrylic of any of my blocks so far. I should probably get the right tool for it for the long term. I don't know the numbers.

 

[tabascosauz]

I believe that's why TDC exists

Thermal Design Current (“TDC”): The maximum current (amps) that can be delivered by a specific motherboard’s voltage regulator configuration in thermally-constrained scenarios.

TDC has absolutely nothing to do with VRM OCP. SVI2 and Ryzen is smart, it's not omniscient. Everything in the past 3 years just points to TDC being a long-term counterpart to EDC.

In theory, PROCHOT EXT (external) sounds like a way for external sources to throttle the CPU, but again a very murky subject and a different topic.

I'm assuming I'm going to get a new board at this point. What should I look for regarding the CPU potential damage?

Mine was a little different being an APU, most of my damage was done to stuff in the SoC domain, so Fabric, memory controller and iGPU were all seriously affected. If I had to guess for yours, probably just any crashing/lower performance/significantly lower clocks than what you'd normally expect.

 

[A Computer Guy]

I'm surprised you made it this far, unless you usually only game and don't run any full loads that max out PPT (all-core CPU rendering). But even still, 5900X and 5950X draw like 100W in most games, which is a lot for this board, and why don't usually recommend anything > 3700X on these 4-phase SInopower boards.

The only reason I made it is likely because of that monoblock and having given up on PBO until today's adventure.

TDC has absolutely nothing to do with VRM OCP. SVI2 and Ryzen is smart, it's not omniscient. Everything in the past 3 years just points to TDC being a long-term counterpart to EDC.

In theory, PROCHOT EXT (external) sounds like a way for external sources to throttle the CPU, but again a very murky subject and a different topic.



Mine was a little different being an APU, most of my damage was done to stuff in the SoC domain, so Fabric, memory controller and iGPU were all seriously affected. If I had to guess for yours, probably just any crashing/lower performance/significantly lower clocks than what you'd normally expect.

Well it's will be a learning experience to find out. I will take picture of my teardown and tests. Hopefully others will learn from my mistakes.

Mussels said:
That said, no one would expect the board to die - just thermal throttle, even with stock cooling

Yea VRM melting to death wasn't what I expected. - I should probably clarify why I had this perhaps short sighted expectation.

Up until this point in my mind I thought the biggest risk to CPU damage when overclocking Ryzen was manually adjusting the voltages especially if it's too high on all core workloads.
This is why in Ryzen master I choose not to manually set voltage and (by assumption) let the CPU do it - by excluding the voltage control in the Ryzen Master OC profile.
It didn't occur to me at the time while the CPU has means to protect itself (voltage, current, heat) it won't to jack to protect the VRM from doing crazy - or the end user for that matter.