The cynic in me says it will have a 6-phase VRM _again_ that's actually a 3-phase VRM _again_ with double the components _again._ Oh, and PCIE4. I absolutely wish they did though. It's like they can't be bothered to match the quality of their Intel and AMD ITX boards.
I am exited about ITX boards now. This means the market can drive the OEMs to make smart design choices on a small space rather then useing double the components on a phase design. I really hope small form factor gets as popular as RGB.
Great! BZ approved! And in the states, this asrock is $20 cheaper than strix z390-i ($200), and every now and then it goes on sale for $40 less than the strix
Related to the decoupling caps for the CPU: these function in a non-intuitive manner and I like that I see the sizes they put on the board. First, a disclaimer, my last computer build was with a Q6700 and later swapped to a Q9400 for heat reasons as it was a 24/7 server. But I dabble in electronics and computing and have gained a decent understanding over the past half decade or so; much thanks to Dave @ EEVBlog, HAD, and the Internet. Anyhow... They "decouple" a power rail from their source rail. They do this by buffering a fast charge change such that when a sudden increase/decrease in current demand occurs such that the high impedance of the supply rail makes it unable to compensate, that the capacitor will source or supply that charge. Now the size of the package has a dramatic effect on how this behaves, from 0201's to 1206's. The larger the package, the lower frequency it is suited for and the smaller for the higher. Similarly, larger capacitances work better for lower frequencies and lower capacitances for higher frequencies. All in all, it isn't about overall capacitance. It's about the lowest impedance over the most optimal frequency range. This is why on high speed PCBs you'll have 0805;'s teamed up with 0603's and 0402's. 500 kHz is really low and the electrolytics get the bulk of that, then they taper things off with 1206's, then 0805's on the PCB and why the CPU PCB has 0603's and 0402's and 0201's.
I have the z370 Fatality ITX which has the same VRM setup with some minor differences (capacitor bank). Regarding the capacitor endurance ratings, you'll notice that Nichicon doesn't advertise any FPCAPs with more than 5k Hours @ 105C. My suspicion is that the 10k & 12k versions found on mobo's are simply rebrands of the 5k's since it seems unlikely that Nichicon would develop but limit sales of these products to only a few manufactures. The mobo manufacturers can set their own qualification criteria meaning the endurance ratings can essentially be whatever they want it to be. Capacitor lifetimes increase exponentially with decreasing temperatures so as long as they remain relatively cool (10 years. This means the endurance rating is not a limiting factor and anything above 5k is just marketing.
asrocks bios is a bit odd. but once you learn to work with its quirks its ok. weird things like settings not applying till a second restart and voltages not being the same on offset mode even though you set the same offset, then you reset the cmos and it works the same again. wierd stuff. but its more than good enough if you can get a decent quality vrm on the board
I bought a ASRock 970 extreme3 back in 2012 and it did the same thing you describe. With my board I noticed if I changed to many settings in bios at one time I would have to reset CMOS or it wouldn't actually change anything. That being said I loved the board, maybe cuz it was my first, and it would overclock a fx6100 to 4.6ghz @ 1.45v, air, although the vrm's were roasting hot, only 4+1 phase with crappy heat sink. Still have the board and CPU but clocked at 3.9 @ 1.275v, my dad's daily for now.
Another great board with Intersil's best components. (ISL69138 with the ISL99227B SPS) eVGA z390, X299 DARK, AsRock's z370 Extreme4, Taichi, Gaming K6, Gaming i7, Gaming iTX. z390 Phantom Gaming iTX and X299 OC, along with the Gigabyte Z370 Gaming 7 and z390 M Gaming, SLI, X,UD, Designare, Ultra, Pro, Pro Wi-Fi, and Elite all use Intersil. They are all decent boards.
''Asrock Z370 extreme 4, Taichi, Gaming K6, Gaming i7, Gigabyte Z390M Gaming, SLI, X UD, Designare, Ultra, Pro, Pro-Wifi, and Elite'' . No, none of those boards use ISL99227B.
7:20 Again, my ASRock Z370 Fatal1ty Gaming K6 is COVERED in Nuvoton controllers! And the BIOS does have MANY options - though it does lack LN2 type options for the most part. The Gaming K6 is meant for, well, Gaming - abd therefore high STABLE Overclocks using Noctua NH-D15 or an AIO, even a Custom copper loop. I am confident the Gaming K6 can handle a 9900k at 5+ GHZ all core, and RAM at whatever the controller will handle. ASRock memory OC seems to be near or at the top of the pack, even if their XMP can be clumsy sometimes.
9:12 It's not the first time ASRock has used expensive powerstages on ITX. My own X99-e ITX/AC and the newer X299 version require them to power their HEDT CPU's. That said, they definitely aren't common and it's good to see them being used for "normal" CPUs. I suspect asrockrack has been using powerstages on their more powerful ITX boards as well (the X299 WSI/IPMI and EPC612D4I for example) If anyone can shove a bunch of stuff on a tiny board (and sell it to the consumer market) it's asrock. Update: I just found out they are making an LGA3647 itx board EPC621D4I-2M
I cant imagine putting a 9900k on that little board. Im hoping 1 or 2 new ryzen 3000 series chips can hit 5ghz. I cant wait till buildzoid is doing a pcb breakdown of a asus x570 rog crosshair viii formula or crosshair viii hero!!
I have mine with 9900k @ 5.0Ghz 1.33v. Board doesn't break a sweat even without any significant airflow. (The system is custom watercooled in a Ghost S1 case. So not so much airflow ;) )
You're hoping ryzen gets a 20% increase in clock speeds and a 15% IPC gain in 1 year? That would make an 8 core get a cb R15 score of 2550 ish, The 16 core version would get around 5000cb which is getting close to a 2990WX. Good luck with that.
Awesome breakdown! I have to disagree with you about the bios though. The ASRock bios's are literally the easiest I've ever used, they make overclocking just flat out simple.
Just had a question for you in regards to an older Z370 board. I have the asRock Z370 extreme4 with an 8700K. Being that it has a 12 phase VRM, can I drop a 9900K in it and get away with running it at 5Ghz all cores? I updated the BIOS to 4.0, so it is compatible with the processor now. Just wondering if I can keep this mobo… And as a side note, I have a case with TONS of airflow, 7 x Noctua 140mm 3000rpm fans, and a 280mm aio for the processor.
just started watching, you'll probably comment on it, the m.2 daughterboard will be right above that exposed chip/die. Not that nvmes are that much hot but still it looks too close for comfort :/
you are wondering about SM7341 not beeing used for Vcore? Well, the 2mΩ is way too much, in comparison the Vcore regulator also used in this board ISL99227B has Low side MOSFET resistance of 0,76mΩ.
1:20 Since the Z77 Extreme4 at LEAST, ASRock Mainstream BIOSes kick ass. I LOVE my ASRock Z370 Fatal1ty Gaming K6, I got it on sale for just $160 and even my admittedly limited understanding of VRMs, the Gaming K6 is quite great, a guy on OC3D forums made a chart of Z370 VRM, and the Fairchild variant (FDPC5030 which I do have!) of the Fatal1ty Gaming K6 is considered TOP of "Upper Middle End VRM" only the Taichi scored better due to 10 true Phases for VCore vs 5 Doubled Component Stages (2 MOSFETs, 2 Chokes, and 2 Caps per VRM "Stage" - Kinda lime Asus with their "Doubled 4 Phase" or whatever they call it on the 11...
if you have this board please submit ticket. event.asrock.com/tsd.asp IOL offset does not do anything. RFR Delay does not do anything. I do not think RTL and IOL init do anything. All you can do is correct the values if they train incorrectly. This board is quite nice but I cannot tighten my RTL and IOL timings. stuck 4133+@12-11-11 1t with around 50rtl 6iol
Could you make a multiphase vrm to convert 120-220v to 12v? I assume you just wouldn't because there are cheaper solutions, but could you make a smaller more efficient power supply by doing that?
The NAND chips don't mind being hot, the M.2 controller however... not so much. The controller will fail long before the flash memory reaches critical temperature, according to what other tech channels(that isn't the Verge) are saying anyways. Come to think of it, how much corsair thermal paste do you have left? I presume quite a lot. LMAO!
In regards to Asrock's bios, I use a z370 k6 in my PC and I have to say that its bios is as good as if not better than the ROG bios. It has even more overclocking options, and the layout makes more sense to me. However, the ab350 pro4 bios is god awful... who doesn't allow for dram voltage changes?
ASRock High-End hardware is usually good, you get what you paid for. Their bios is total shit! if only there was a way to decouple the software from the hardware like open source bios or something? then that would be heaps better than ANY ASRock bios.
Good idea. Check out the open source project Coreboot tho. We can't even put Coreboot on these new motherboards though, the customization process is very bad. We should get them to sell boards with "open source firmware supported" then we might get enough progress to get that commonplace. ASRock always was one to try new things and produce specialized products, it would be great to get them to adopt this one and try to fix their clunky bios.
Too much rambling and not enough actual beneficial information - like who cares why components they use, if in the end the whole mobo shows crap benchmarks :/
It’s a PCB breakdown not a review. Component choice is very important for longevity, efficiency, and overclocking. You know... the name of the channel?
The NAND chips don't mind being hot, the M.2 controller however... not so much. The controller will fail long before the flash memory reaches critical temperature, according to what other tech channels(that isn't the Verge) are saying anyways. Come to think of it, how much corsair thermal paste do you have left? I presume quite a lot. LMAO!
Actually the NAND has a ideal temp range and its quite warm. When you cool the NAND down from that point you actually drastically reduce the endurance of the drive. It is actually very possible to cut the endurance in half using these solutions some mobos have and companies are selling aftermarket. The controller id usually cooled pretty well with a metal strip sticker that spreads the heat out across the drive. With that it is fairly hard to peg the drive hard enough that the controller runs to hot. Not impossible for sure but for someone thats gaming or even doing some pretty heavy read and writes most end users wont stress a drive that hard. Now that wont tend to include pro/enthusiasts using their machine for work or some other stressful task.
@@Onewolfoc The only scenario that comes to my mind where the SSD would lose in endurance/lifespan is by rapidly cooling down the NAND flash down to ambient from its peak performance temp rather than let chips cool down slowly on their own. That'll definetly degrade them somewhat if done repeatedly, much like how some metals will turn brittle when heated up until white, then cooled down rapidly. Daily/regular usage won't do squat to an SSD in terms of heat tolerances, but when there's a VRM heatsink dumping let's say an extra 20 watts of heat on top of it, the controller is going to get stressed significantly more than a several hours long gaming session possibly can, and will be prone to failing first. If the heatpipe wasn't near the base of the block on the VRM side, then there'd be no real cause for concern... but this isn't the case, so buildzoid's comments on the design are, in my opinion, worthy of consideration if/when overclocking on ambient cooling only with this board.
@@tohothewriter8002 actually depending on if your writing or storing data the nand temp at the time of the action heavily effects endurance. A 20 or 30 degree drop out of a ideal range for writes can actually cut a huge chunk of lifespan off a drive. Like i said it is totally possible to half drive life with some of the heatsink assemblies people attach to their drives. The nand flash itself likes to be hot for writes while reads or long term storage its not only ok if the drives substantially cooler it is recommended for storage. A basic tiny amount of airflow and the metal sticker on the drive is all the cooling that most people should use. This would change in the enterprise or upon heavy read write cycles which most people won't end up stressing the drive enough. Prosumers , Enterprise , or some enthusiasts may fall within a scenario that they would have to adjust cooling i.e. try and cool the controller a tad better while not sinking that much energy away from the nand. Then you got the edge case where someone mashes a nvme drive on the back side of a ITX board between the tray and pcb which then it could get toasty which again is a situation where youd want to adjust the cooling accordingly or even make sure a small amount of air was circulated behind the mobo between the tray and the mobo.
Everything PCB is greek to me, but I did enjoy watching you draw orange boxes around the black boxes and say letters.
I really hope asrock makes an am4 version of this itx board.
I'm gonna put a 16 core ryzen 3000 on it.
Desk mini A300 :P
The cynic in me says it will have a 6-phase VRM _again_ that's actually a 3-phase VRM _again_ with double the components _again._ Oh, and PCIE4.
I absolutely wish they did though. It's like they can't be bothered to match the quality of their Intel and AMD ITX boards.
MSI has the B450I Gaming Plus AC with an as good or better VRM setup than this.. 6+2 IR3555's with a IR35201 controller.
Asrock simply has the best ITX designs in the market. Even on AM4 they have great ITX mobos.
I am exited about ITX boards now. This means the market can drive the OEMs to make smart design choices on a small space rather then useing double the components on a phase design. I really hope small form factor gets as popular as RGB.
Pro Tip: hold the shift key to make perfectly straight lines
Great! BZ approved!
And in the states, this asrock is $20 cheaper than strix z390-i ($200), and every now and then it goes on sale for $40 less than the strix
Related to the decoupling caps for the CPU: these function in a non-intuitive manner and I like that I see the sizes they put on the board.
First, a disclaimer, my last computer build was with a Q6700 and later swapped to a Q9400 for heat reasons as it was a 24/7 server. But I dabble in electronics and computing and have gained a decent understanding over the past half decade or so; much thanks to Dave @ EEVBlog, HAD, and the Internet.
Anyhow... They "decouple" a power rail from their source rail. They do this by buffering a fast charge change such that when a sudden increase/decrease in current demand occurs such that the high impedance of the supply rail makes it unable to compensate, that the capacitor will source or supply that charge. Now the size of the package has a dramatic effect on how this behaves, from 0201's to 1206's. The larger the package, the lower frequency it is suited for and the smaller for the higher. Similarly, larger capacitances work better for lower frequencies and lower capacitances for higher frequencies.
All in all, it isn't about overall capacitance. It's about the lowest impedance over the most optimal frequency range. This is why on high speed PCBs you'll have 0805;'s teamed up with 0603's and 0402's. 500 kHz is really low and the electrolytics get the bulk of that, then they taper things off with 1206's, then 0805's on the PCB and why the CPU PCB has 0603's and 0402's and 0201's.
I have the z370 Fatality ITX which has the same VRM setup with some minor differences (capacitor bank). Regarding the capacitor endurance ratings, you'll notice that Nichicon doesn't advertise any FPCAPs with more than 5k Hours @ 105C. My suspicion is that the 10k & 12k versions found on mobo's are simply rebrands of the 5k's since it seems unlikely that Nichicon would develop but limit sales of these products to only a few manufactures. The mobo manufacturers can set their own qualification criteria meaning the endurance ratings can essentially be whatever they want it to be. Capacitor lifetimes increase exponentially with decreasing temperatures so as long as they remain relatively cool (10 years. This means the endurance rating is not a limiting factor and anything above 5k is just marketing.
asrocks bios is a bit odd. but once you learn to work with its quirks its ok. weird things like settings not applying till a second restart and voltages not being the same on offset mode even though you set the same offset, then you reset the cmos and it works the same again. wierd stuff. but its more than good enough if you can get a decent quality vrm on the board
I bought a ASRock 970 extreme3 back in 2012 and it did the same thing you describe. With my board I noticed if I changed to many settings in bios at one time I would have to reset CMOS or it wouldn't actually change anything. That being said I loved the board, maybe cuz it was my first, and it would overclock a fx6100 to 4.6ghz @ 1.45v, air, although the vrm's were roasting hot, only 4+1 phase with crappy heat sink. Still have the board and CPU but clocked at 3.9 @ 1.275v, my dad's daily for now.
Another great board with Intersil's best components. (ISL69138 with the ISL99227B SPS) eVGA z390, X299 DARK, AsRock's z370 Extreme4, Taichi, Gaming K6, Gaming i7, Gaming iTX. z390 Phantom Gaming iTX and X299 OC, along with the Gigabyte Z370 Gaming 7 and z390 M Gaming, SLI, X,UD, Designare, Ultra, Pro, Pro Wi-Fi, and Elite all use Intersil. They are all decent boards.
''Asrock Z370 extreme 4, Taichi, Gaming K6, Gaming i7, Gigabyte Z390M Gaming, SLI, X UD, Designare, Ultra, Pro, Pro-Wifi, and Elite'' .
No, none of those boards use ISL99227B.
Thanks for this. I would be very interested to see your actual overclocking results with this board!
bought this mobo based on your recommendation, buildzoid. and bought your socks!
There is a bios by JZ out for the Phantom ITX, which disables vrm throttling completely, so you can grill them if you want to.
7:20 Again, my ASRock Z370 Fatal1ty Gaming K6 is COVERED in Nuvoton controllers! And the BIOS does have MANY options - though it does lack LN2 type options for the most part. The Gaming K6 is meant for, well, Gaming - abd therefore high STABLE Overclocks using Noctua NH-D15 or an AIO, even a Custom copper loop. I am confident the Gaming K6 can handle a 9900k at 5+ GHZ all core, and RAM at whatever the controller will handle. ASRock memory OC seems to be near or at the top of the pack, even if their XMP can be clumsy sometimes.
amazing board, I have my 9900ks direct die cooled at 5.3ghz avx0 stable.. love it.
9:12 It's not the first time ASRock has used expensive powerstages on ITX. My own X99-e ITX/AC and the newer X299 version require them to power their HEDT CPU's. That said, they definitely aren't common and it's good to see them being used for "normal" CPUs. I suspect asrockrack has been using powerstages on their more powerful ITX boards as well (the X299 WSI/IPMI and EPC612D4I for example)
If anyone can shove a bunch of stuff on a tiny board (and sell it to the consumer market) it's asrock.
Update: I just found out they are making an LGA3647 itx board EPC621D4I-2M
I cant imagine putting a 9900k on that little board. Im hoping 1 or 2 new ryzen 3000 series chips can hit 5ghz. I cant wait till buildzoid is doing a pcb breakdown of a asus x570 rog crosshair viii formula or crosshair viii hero!!
@hytriarch that's awesome! This board looks way better than I thought it would be with a 5 phase vrm.
I have mine with 9900k @ 5.0Ghz 1.33v. Board doesn't break a sweat even without any significant airflow. (The system is custom watercooled in a Ghost S1 case. So not so much airflow ;) )
You're hoping ryzen gets a 20% increase in clock speeds and a 15% IPC gain in 1 year? That would make an 8 core get a cb R15 score of 2550 ish, The 16 core version would get around 5000cb which is getting close to a 2990WX.
Good luck with that.
www.techpowerup.com/255099/possible-listings-of-amd-ryzen-9-3800x-ryzen-7-3700x-ryzen-5-3600x-surface-in-online-stores
@@HyunMoKoo Custom cooled? Would you be so kind to share some shots and part choices? I'm planning the very same.
Awesome breakdown! I have to disagree with you about the bios though. The ASRock bios's are literally the easiest I've ever used, they make overclocking just flat out simple.
Also the z370 Aorus Gaming 7 has the same iGPU VRM. With the UHD630, you can run HALF-LIFE 2 at 1.6GHz..
Concerning M.2's: Flash chips like it warm, like 40 to 55°C warm. The controller on the other hand would rather have it below 50°C
Just had a question for you in regards to an older Z370 board. I have the asRock Z370 extreme4 with an 8700K. Being that it has a 12 phase VRM, can I drop a 9900K in it and get away with running it at 5Ghz all cores? I updated the BIOS to 4.0, so it is compatible with the processor now. Just wondering if I can keep this mobo… And as a side note, I have a case with TONS of airflow, 7 x Noctua 140mm 3000rpm fans, and a 280mm aio for the processor.
Just out of curiosity, Which case do you use?
just started watching, you'll probably comment on it, the m.2 daughterboard will be right above that exposed chip/die. Not that nvmes are that much hot but still it looks too close for comfort :/
11:57 - The connection is actually ground. Verified on VCore and iGPU
even if the caps are not TEH BEST you can always swap them out, as a straight up upgrade, or to repair them. Maybe. I would consider it at least.
you are wondering about SM7341 not beeing used for Vcore?
Well, the 2mΩ is way too much, in comparison the Vcore regulator also used in this board ISL99227B has Low side MOSFET resistance of 0,76mΩ.
Please do a pcb breakdown of the MSI Lightning Z 2080ti board if you haven't already!
The z370 Aorus Gaming 7 uses the same SPS. From what I can see. Do all the ISL99227B IC's have the little metal tab on it (that looks like a little 1)
1:20 Since the Z77 Extreme4 at LEAST, ASRock Mainstream BIOSes kick ass. I LOVE my ASRock Z370 Fatal1ty Gaming K6, I got it on sale for just $160 and even my admittedly limited understanding of VRMs, the Gaming K6 is quite great, a guy on OC3D forums made a chart of Z370 VRM, and the Fairchild variant (FDPC5030 which I do have!) of the Fatal1ty Gaming K6 is considered TOP of "Upper Middle End VRM" only the Taichi scored better due to 10 true Phases for VCore vs 5 Doubled Component Stages (2 MOSFETs, 2 Chokes, and 2 Caps per VRM "Stage" - Kinda lime Asus with their "Doubled 4 Phase" or whatever they call it on the 11...
iol offset and associated latency timings don't accept input/don't do anything in this bios.. pls ask them to fix it
if you have this board please submit ticket.
event.asrock.com/tsd.asp
IOL offset does not do anything. RFR Delay does not do anything. I do not think RTL and IOL init do anything. All you can do is correct the values if they train incorrectly.
This board is quite nice but I cannot tighten my RTL and IOL timings. stuck 4133+@12-11-11 1t with around 50rtl 6iol
I have an updated BIOS with functional IOL Offset and RTL INIT. It is based on P4.00, not P1.60.
More Buildzoid = more better.
Could you make a multiphase vrm to convert 120-220v to 12v? I assume you just wouldn't because there are cheaper solutions, but could you make a smaller more efficient power supply by doing that?
PSU already kinda do that.
I was wondering if you could do a follow up video about the mods that worked and didn't work on this board.
The NAND chips don't mind being hot, the M.2 controller however... not so much. The controller will fail long before the flash memory reaches critical temperature, according to what other tech channels(that isn't the Verge) are saying anyways. Come to think of it, how much corsair thermal paste do you have left? I presume quite a lot. LMAO!
Hey can you do one of your analysis and breakdown videos on the ZOTAC AMP EXTREME 2080TI?
Dude! The golden cylinder at the bottom left of this board, is bent on my board. Is that going to be a problem? It came like that from Amazon.
I wish More motherboard manufacturers made high end ITX motherboards for ryzen.
Is there any significant difference (aside from the heatsink and extra m.2 slot) between this one and the Z370 Fatal1ty ITX?
is the iGPU power just left unused with CPUs that don't have integrated graphics?
Is Z390M-ITX/ac the same quality and vrm wise?
Do the asrock phantom Gaming X z390 atx board
Thank you
Have you checked the supermicro one? It uses TDA21740 but has a shitty heatsink.
I kinda forgot that they've been trying to get into making consumer boards because last I heard their BIOS is atrocious.
In regards to Asrock's bios, I use a z370 k6 in my PC and I have to say that its bios is as good as if not better than the ROG bios. It has even more overclocking options, and the layout makes more sense to me. However, the ab350 pro4 bios is god awful... who doesn't allow for dram voltage changes?
What RAM would you buy for a 32GB B-Die Kit?
I have a Samsung 970 Pro, and I will go all the way to saying that the M.2 will cook the VRM, not the other way around. Hahahahahaha
The controllers on M.2s are evil confirmed.
anyone know what that round black thingy is in front of the wifi antenna??
Maybe I'm just insensate, but I don't think you identified the controller.
Too much reflected light on the picture, we can't read the chipnames.
first
trifs
Wait, did you skip the main VRM controller? Or do I need more coffee...
@@Qyngali I might've forgotten to mention it
i agree Asrock bios are the worst, they have always been. i dont know why they refuse to invest in a bios upgrade or overhaul
This board is better than asus strix z390i gaming ?
ASRock High-End hardware is usually good, you get what you paid for. Their bios is total shit! if only there was a way to decouple the software from the hardware like open source bios or something? then that would be heaps better than ANY ASRock bios.
Good idea. Check out the open source project Coreboot tho. We can't even put Coreboot on these new motherboards though, the customization process is very bad. We should get them to sell boards with "open source firmware supported" then we might get enough progress to get that commonplace. ASRock always was one to try new things and produce specialized products, it would be great to get them to adopt this one and try to fix their clunky bios.
like like like....
Argh so close......third!
Last!
Just put actual finned heatsink like Aorus X470 gaming 7 on all those boards please....damn stupid
Second
First what?
fourth
Why are you so poor?
Too much rambling and not enough actual beneficial information
- like who cares why components they use, if in the end the whole mobo shows crap benchmarks :/
It’s a PCB breakdown not a review. Component choice is very important for longevity, efficiency, and overclocking. You know... the name of the channel?
The NAND chips don't mind being hot, the M.2 controller however... not so much. The controller will fail long before the flash memory reaches critical temperature, according to what other tech channels(that isn't the Verge) are saying anyways. Come to think of it, how much corsair thermal paste do you have left? I presume quite a lot. LMAO!
Actually the NAND has a ideal temp range and its quite warm. When you cool the NAND down from that point you actually drastically reduce the endurance of the drive. It is actually very possible to cut the endurance in half using these solutions some mobos have and companies are selling aftermarket. The controller id usually cooled pretty well with a metal strip sticker that spreads the heat out across the drive. With that it is fairly hard to peg the drive hard enough that the controller runs to hot. Not impossible for sure but for someone thats gaming or even doing some pretty heavy read and writes most end users wont stress a drive that hard. Now that wont tend to include pro/enthusiasts using their machine for work or some other stressful task.
@@Onewolfoc The only scenario that comes to my mind where the SSD would lose in endurance/lifespan is by rapidly cooling down the NAND flash down to ambient from its peak performance temp rather than let chips cool down slowly on their own. That'll definetly degrade them somewhat if done repeatedly, much like how some metals will turn brittle when heated up until white, then cooled down rapidly.
Daily/regular usage won't do squat to an SSD in terms of heat tolerances, but when there's a VRM heatsink dumping let's say an extra 20 watts of heat on top of it, the controller is going to get stressed significantly more than a several hours long gaming session possibly can, and will be prone to failing first. If the heatpipe wasn't near the base of the block on the VRM side, then there'd be no real cause for concern... but this isn't the case, so buildzoid's comments on the design are, in my opinion, worthy of consideration if/when overclocking on ambient cooling only with this board.
@@tohothewriter8002 actually depending on if your writing or storing data the nand temp at the time of the action heavily effects endurance. A 20 or 30 degree drop out of a ideal range for writes can actually cut a huge chunk of lifespan off a drive. Like i said it is totally possible to half drive life with some of the heatsink assemblies people attach to their drives. The nand flash itself likes to be hot for writes while reads or long term storage its not only ok if the drives substantially cooler it is recommended for storage. A basic tiny amount of airflow and the metal sticker on the drive is all the cooling that most people should use. This would change in the enterprise or upon heavy read write cycles which most people won't end up stressing the drive enough. Prosumers , Enterprise , or some enthusiasts may fall within a scenario that they would have to adjust cooling i.e. try and cool the controller a tad better while not sinking that much energy away from the nand. Then you got the edge case where someone mashes a nvme drive on the back side of a ITX board between the tray and pcb which then it could get toasty which again is a situation where youd want to adjust the cooling accordingly or even make sure a small amount of air was circulated behind the mobo between the tray and the mobo.