Roman “der8auer” Hartung lays into X299 boards’ overheating VRMs

The launch of Intel's latest X-Series processors might seem like a summertime Christmas present for overclocking guru Roman "der8auer" Hartung. Unfortunately, Hartung doesn't seem to be happy with his gift. In a blunt and scathing video on his YouTube channel, he makes the bold claim that overclocking on X299 motherboards right now is limited due to what he regards as poor VRM heatsink design.

Hartung had three motherboards on hand: the Aorus X299 Gaming 3, Asus' Prime X299-Deluxe, and MSI's X299 Pro Gaming Carbon. During his testing, he found that the VRM heatsinks on all three models reached excessive temperatures, even with a relatively modest 4.6 GHz CPU overclock. After noticing odd throttling with these boards, he started attaching thermocouples to the VRM heatsink and the PCB behind the modules in order to investigate.

MSI's board was the coolest of the bunch with its VRM heatsink at 75° C. Asus's heatsink hit 86° C, and the heatsink on the Aorus board managed to reach a positively-toasty 105° C. Hartung also noticed excessive temperatures at the Aorus motherboard's power connector, and expressed frustration with the fact that the model in question only uses one 8-pin plug. He was eventually able to tame down VRM temperatures and achieve higher stable overclocks by removing the respective VRM heatsinks and aiming a 120-mm fan at the modules.

We didn't quite have the same experience as Hartung did when overclocking X-series processors on Asus's X299 Prime-Delux, since we were able to hit a 4.7 GHz overclock without trouble. That being said, we also held off on pushing our Core i7-7900X too far, as our editor believed that the firmware and monitoring utilities were still immature. In addition, well-known PSU reviewer Jon "JonnyGuru" Gerow stated to TechPowerUp his belief that the SuperFlower PSU that Hartung apparently used could also have contributed to the odd overclocking behavior.

Despite his criticism of the boards' VRM heatsinks, Hartung expressed some sympathy for the mobo manufacturers, as he suspects that Intel didn't give them quite as much time as usual to put together these products. For those interested in overclocking on the X299 platform, his recommendation is simple: wait. Hartung believes that a new round of motherboards will be out in about a month, and suggests that folks hold off for one of those.

Comments closed
    • Prestige Worldwide
    • 2 years ago

    Just a senationalist video for view revenue, which I suspect is this youtuber’s primary motivation.

    OC3D tested numerous X299 motherboards and had no such issue.

    [url<]https://www.overclock3d.net/reviews/cpu_mainboard/asus_x299-e_strix_gaming_review/3[/url<] [quote<]Temperatures If you'll excuse us for a moment there is a can of worms that needs opening. If you pay even a passing interest in the world of the X299 chipset it can't have escaped your notice that some people - mentioning no names to spare their blushes - have been shouting loudly about how the MOSFETs and power draw of the X299 can lead to massive heat issues with even the freaking cables getting hot, and the power phases becoming a handy emergency griddle. As always with these things it is a case of someone doing something once, getting a bad result, and then shouting to the world about it before they've checked their facts. For example, we ran the Strix at maximum overclocking under the kind of loading that only OCCT can provide, and after what felt like forever we measured the temperatures and discovered that the front MOSFET heatsink reached the giddying heights of 48°C quite a bit less than the 70-80c that is bounding round the internet. We also ran the tests on the rear of the board right beneath the mosfets to mimic the other testing and we saw 61.8 degrees, this means our temps were at least 30C lower than what has been seen elsewhere online. Blimey, better call the fire brigade. Before you say "ahhhh but he didnt test the strix" we have tested all of our boards here and they none were excessively hot or overheated. We used the Prime X299-A board in our CPU reviews and haven't seen any issues at all, also we have only ever used a single 8pin CPU power connector, adding in an extra cable didnt effect the overclocks what so ever. Maybe running them in a case makes a huge difference? Well no. removing the AIO from the roof and disconnecting ALL other fans made our temps rise by no more than 3C. Myth Busted? So no, the X299 has no huge platform killing issues with temperatures. Yes the mosftet heatsinks are getting a bit warmer due to the extra load and limited board real estate for cooling, but nothing like what has been reported, not even close. We do not doubt their results, but we do question the ability of the hardware used. Don't believe everything you hear, sometimes its better to stay off the bandwagon and think for yourself.[/quote<]

    • Kougar
    • 2 years ago

    Is it any surprise when VRM heatsinks are simply chunks of metal with no fins for surface area, say the thing that actually differentiates a heatsink from a block of metal?

    • ronch
    • 2 years ago

    I sometimes wonder if those funny heatsinks so common today actually dissipate heat well enough.

    • psuedonymous
    • 2 years ago

    As der8auer (it took three attempts not to type the letter ‘b’ there) himself mentions, the VRMs work fine if the heatsinks are removed entirely. The problem is entirely down to cosmetic-only ‘heatsinks’, the underlying power delivery circuitry is fully functional at extreme loads. Who wants to be that slapping a fancy-looking insulator on components to differentiate your ‘low end’ boards from ‘high end overclocking’ boards is going to be more prevalent as chip manufacturers mandate minimum standards for power-delivery circuitry?

    On the 8-pin ‘issue’: a single EPS12V connector should be able to handle 380+W with no issue (minimum according to standard is 4 rails at 8A per rail at 12V). As pointed out by [url=https://www.youtube.com/watch?v=f7BqAjC4ZCc&lc=z121c1gptofrcbnej22vxrxolpmfupblr<]JonnyGugu[/url<], the connector overheating is down to him using a PSU with crappy cosmetic light-up connectors.

    • thecoldanddarkone
    • 2 years ago

    The vrm heatsinks (I mean blocks of aluminum) are just mounted on the VRM’s. Many of them don’t even look like they are made to dissipate heat. I’m aware that actual heatsinks are ugly, but….

      • ronch
      • 2 years ago

      Yeah. If only board makers remember that heatsinks actually have something important to do, not bling the board up.

      • Srsly_Bro
      • 2 years ago

      Finish that thought, bro.

    • Aquilino
    • 2 years ago

    The thing with youtubers is that, unlike GamersNexus and similar sites, people don’t show methodology.
    So there are a lot of questions about how those numbers appeared, i.e.: [url<]http://www.jonnyguru.com/forums/showthread.php?t=14608[/url<]

    • POLAR
    • 2 years ago

    Look, Son: a bad platform

      • Klimax
      • 2 years ago

      Incorrect. For corrections see rest of discussion.

    • willmore
    • 2 years ago

    I wonder if this issue is related to the excess power draw that many reviewers saw. Could be that the VRMs were designed based on the info that Intel provided, but that now seems to be incorrect values.

    Even with a safe design margin in thre, the CPUs could have eaten that up with their excessive power draw. That would leave no room for overclocking–or very little.

    Any news from Intel on that power draw issue, yet, Jeff?

      • POLAR
      • 2 years ago

      So you are suggesting power figures were a little lower than real? Who would have tought that..

        • willmore
        • 2 years ago

        That’s one of the common threads the reviewers are seeing–more power draw than specified by Intel.

      • eofpi
      • 2 years ago

      Intel’s TDP claims haven’t reflected maximum CPU power for many years now. Maybe this is what happens when TDP is taken at face value?

    • crystall
    • 2 years ago

    So hurray for top-down coolers (and fans)!

      • DPete27
      • 2 years ago

      No kidding. in the video he says he uses an open test bench with a liquid cooler so likely zero airflow over/around the mobo.

    • Krogoth
    • 2 years ago

    X299 and current crop of Socket 2066 chips were rushed to market.

    It is painfully obviously.

    Intel wasn’t planning on launching the platform until Q4 2017 and Q1 Q2018 but pressure from AMD pushed the schedule up.

      • jihadjoe
      • 2 years ago

      I guess that’s also why it tops out at 10 cores right now. They probably originally planned to top out at 10, with the 7900X priced about the same as the 6950X.

      Also if the VRMs are having a hard time now, just wait till the higher core count i9s come out!

      • ermo
      • 2 years ago

      What this tells me is that perhaps Intel were caught off-guard to the reception of both AMD’s ‘consumer’ RyZen chips AND its ThreadRipper chips.

      To be clear, I’m not saying anything about AMD’s chips or platforms being objectively better, just that AMD have positioned its products well enough to force Intel to respond if it doesn’t want to lose previously guaranteed sales and therefore revenue.

    • derFunkenstein
    • 2 years ago

    My unsubstantiated belief is that most people buying X299 systems are going to combine them with either large towers that blow air from front to back or large AIO closed-loop coolers. So if this is really a big problem (and not a power supply problem like JonnyGuru suggested) I’m sure it’ll show up relatively frequently.

    Relatively few people use any sort of active cooling on VRMs. Most cases today don’t have any sort of accommodations for a fan blowing at the motherboard, either—especially windowed ones.

      • RAGEPRO
      • 2 years ago

      Well, it’s (at least) two separate problems — insufficient VRM design and/or cooling and then crap power supply design. The former problem is most of what der8auer is talking about, the latter problem is just related to the 65C cable temperatures at the CPU power connector on the motherboard. Apparently using a non-crap power supply (huh, I thought SuperFlower was good) would resolve the latter issue, but the former is still a problem.

        • derFunkenstein
        • 2 years ago

        I also thought SuperFlower was good, so that’s weird. Separate or the same problem, doesn’t matter. [url=https://twitter.com/jkampman_tr/status/880374335441166337<]Jeff's practice excepted[/url<], my own belief (based on nothing but looking at tons of photos of new builds on the internet over the last 15 years) is that basically nobody actively cools the VRM.

          • RAGEPRO
          • 2 years ago

          I’m actually running an [url=http://noctua.at/en/nh-c14<]NH-C14[/url<] with the fans flipped around and blowing straight out the side exhaust on my case. Instead of down-draft, up-draft! I strongly believe that's been a major contributor to the continued stability of my overclock. 🙂

            • derFunkenstein
            • 2 years ago

            Probably true, but you’re in a very firm minority.

            • Ryu Connor
            • 2 years ago

            My X99 system is inside a Corsair Carbide Air 540. I have two 140mm fans up top pushing air through the Corsair H115 radiator and into the case. I have two 140mm fans up front pushing air into the case. I have one 140mm fan at the back exhausting air out of the case.

            In short my entire system is drowning in fresh airflow, all of which easily hits the motherboard VRM cluster.

            I’d be rather surprised if I’m doing something unique amongst our little enthusiast community.

            Or TLDR I think what RagePro and I are doing is pretty normal. Especially for a big ATX system.

            Now what those ITX guys are doing. I can’t vouch for them.

            • slowriot
            • 2 years ago

            Your configuration is about the maximum I would expect to be honest. Personally I think of cases like say a Phanteks Evolv which doesn’t have as great top air flow options (fan mounts sure, but they’re heavily choked) and other similar cases.

            • G8torbyte
            • 2 years ago

            I’m doing that similarly on my X99 except I have the inflow of air from the Corsair water block radiator on the back. Two fans up front like yours pulling in but the heat is being pulled up and out the top with another top fan exhausting out. Since heat rises through natural physics I figured go with the flow.

            • thecoldanddarkone
            • 2 years ago

            When I moved to water cooling, I started actively cooling my VRM’s. I have one of those ram coolers and it blows enough air through the heatsink (lowest possible rpm setting). The VRM’s would get toasty with the 4930k@ 4.5 on my x79ws. Just this simple amount of cooling makes it so they are reasonable.

            • bhtooefr
            • 2 years ago

            I’m thinking down-draft would get marginally more airflow over the VRMs (pushing it around the board, rather than sucking from areas that aren’t the board), especially given the C14’s airflow path, but that’s just me.

            There really is something to be said, though, for top-down cooling instead of tower coolers, if you’ve got room for it, due to the ability to cool things that aren’t the CPU.

          • slowriot
          • 2 years ago

          That’s an interesting practice choice and not at all what I’d imagine real customers doing.

            • derFunkenstein
            • 2 years ago

            To be clear, I think it’s a good idea when OCing, but I can tell you right now my two OC’d systems don’t have active VRM cooling, and like you said I don’t think most customers are doing that.

            I’d also be interested in seeing some boutique builder systems that actively cool VRMs. Most of them advertise CLCs and (I’m guessing) expect the rad fans pulling hot air out to be enough.

        • psuedonymous
        • 2 years ago

        [quote<]insufficient VRM design and/or cooling and then crap power supply design. The former problem is most of what der8auer is talking about[/quote<] Der8auer literally stated that when he removed the heatsinks to run the bare VRMs, the problem disappeared. The VRM circuitry itself is fine, it's just the 'heatsinks' that act ass insulators that are not.

          • exilon
          • 2 years ago

          More accurately, he said that he removed the heatsinks and pointed a fan at it. He then also stated in the comments that pointing the fan at the heatsinks worked as well.

      • Convert
      • 2 years ago

      This would be me. All of my x99 systems use tower coolers and the boards lay flat so I’m probably not getting any extra natural convection. When I start getting x299 systems they will be the same configuration…

    • chuckula
    • 2 years ago

    I hear der8auer is turning this issue into a musical.
    Here’s his [url=https://youtu.be/EhZba-P7R18?t=42s<]main theme[/url<].

      • morphine
      • 2 years ago

      Ah, the interesting “the uploader has not made this video available in your country” ditty. A popular one, but I’ll admit I’m growing a little tired of it.

        • chuckula
        • 2 years ago

        If it makes you feel better, you might have gotten the better end of the bargain by having it blocked.

          • derFunkenstein
          • 2 years ago

          I saw start of the video and then I immediately closed it. I think you’re right.

          • morphine
          • 2 years ago

          Thank the deities for the small blessings.

          • jihadjoe
          • 2 years ago

          Says something when [url=https://www.youtube.com/watch?v=JiH7F9oopco<]a cover done by some aged Vengaboys[/url<] is better lol

        • UberGerbil
        • 2 years ago

        It’s the first-ever [url=http://i.imgur.com/a2iMLBu.gif<]all-mime musical[/url<]

Pin It on Pinterest

Share This