Ranting about X870 VRM/power stage current ratings AGAIN

Поділитися
Вставка
  • Опубліковано 21 лис 2024

КОМЕНТАРІ • 116

  • @Hardwareunboxed
    @Hardwareunboxed 13 днів тому +73

    BZ getting in before we have exclusive 210A power stages with no spec sheet information.
    Also Asus are big on #trustmebro.

    • @danieloberhofer9035
      @danieloberhofer9035 11 днів тому +2

      As if anyone trusted Asus right now... #thankssteve (even if it's the other Steve - we like both of you!)

  • @rielleleaf
    @rielleleaf 13 днів тому +71

    "I'll try to make this one shorter"
    *nervously looks at the **49:50*

  • @pranze3484
    @pranze3484 14 днів тому +50

    Those mobos are sold for 500€ in EU... Absolutely INSANE

    • @ChrisM541
      @ChrisM541 13 днів тому +8

      Doubly INSANE because there is practically zero difference with the X670.

  • @Wasmachineman
    @Wasmachineman 13 днів тому +34

    bz rant about VRMs.
    Yes.

  • @tkgg
    @tkgg 13 днів тому +14

    Thanks BZ. Information is power and you're supplying it.

    • @Demrezel
      @Demrezel 13 днів тому

      What can you do with this information? Will go and buy a replacement for that part? Will you be able to convince the manufacturers to use a better component (probably much more expensive)? The only thing that matters is overall performance and price other than that this is just information overload.

    • @jm8080ful
      @jm8080ful 13 днів тому

      @@Demrezel Real answer is no, you can't do anythng. But knowing is always the start of something

    • @tkgg
      @tkgg 13 днів тому +1

      @@Demrezel It's a rant that also provides information. I know jack shit about motherboard components so any information I recieve helps me make purchasing decisions moving forward. It's also the reason I usually watch motherboard breakdowns before making a purchase.

  • @larsjrgensen5975
    @larsjrgensen5975 13 днів тому +13

    60 amp powerstage curve goes to 55 Amps
    80 amp powerstage curve goes to 80 Amps
    100 amp powerstage curve goes to 40 Amps
    For those who dare.

  •  13 днів тому +10

    So to sum it up, ASUS advertisers bigger and better powerstages, sells crappier and weaker powerstages and charge more for the motherboards then ever.
    Btw the video was VERY interesting! It would be very cool if you tested the efficiency off different power stages, good for the consumer if anything to know what they actually get for there money!
    Im not surprised if the new stuff is worse, just feel like we pay more and more for less and less quality and features like no debug display on low end boards, its just not fun anymore to get shafted.

    • @N-A762
      @N-A762 7 днів тому +1

      That scares me I just bought an Asus ProArt X870E. Inspent good money for it to.

    • @plinyvicgames
      @plinyvicgames 4 дні тому

      shrinkflation

  • @wewillrockyou1986
    @wewillrockyou1986 13 днів тому +7

    I remember a while back seeing Vishay come out with the SiC840 DrMOS, they were probably the first to be rated at 100A, but they were like 3-4% less efficient than most 70A power stages out at the time from Infineon or Ti (using normalised voltages and Fsw)... Hell the 840 is less efficient than most of Vishay's 50A power stages. The 840s are kinda oddball parts because their peak efficiency is quite low but they can maintain it until something like 20-25A current and taper off less in the 25-40A range, most powerstages and DrMOS parts will be losing efficiency quickly by the 15A mark, even 70/90A rated ones. Also, rest in peace directfets 😔

  • @brovid-19
    @brovid-19 10 днів тому +1

    I don't care how many rant videos you make, when it's about the electrical aspect, I could listen all day

  • @asm_nop
    @asm_nop 13 днів тому +9

    "Rated for 110A"*
    *PCB layout requires 24 layers of vias and planes, Vo > 2V, fsw < 20kHz, Ta = fucking antarctica.

    • @SkylerAk
      @SkylerAk 13 днів тому +3

      Realistically on the full datasheet the nominal maximum current is probably closer to 60 amps around 40 degrees Celsius due to derating. It’s all marketing at this point.

  • @TheCaptainklotz
    @TheCaptainklotz 13 днів тому +1

    I just recently stumbled upon your account, it was your very comprehensive video abour overclocking- i actually enjoy the technical nerdiness about what makes those tiny chips fast way more than the dragrace part of overclocking :)
    besides your likeable voice, from a podcast point of view, i really enjoy your knowledge of electronical engineering :)
    If someday your channel grows and you dont know what to do content for, do a channel that simply reviews pc compents on the aspect of coil whine :P it should become the most valuable channel for pc enthusiasts )
    keep on the good work :)

  • @ivantsipr
    @ivantsipr 13 днів тому +10

    "I'll try to make it shorter" "49:50" 🤔🤣

  • @Iscandelt
    @Iscandelt 10 днів тому +1

    41:28 :D
    I liked the rant, have the same opinion on the powerstages. They USED TO mean something, now it's yeat another "trust_me_bro" number.

  • @albertlong3492
    @albertlong3492 12 днів тому

    17:30 what if the manufacturers put the mosfets on the back, leave blank space on the front and stick the heatsinks on that? will that be expensive or are the different layers in the pcb massive thermal insulators? just a question I had for a while and I wanted to ask

  • @aviciiaf3099
    @aviciiaf3099 13 днів тому +3

    Vishay manufactured these power stages for marketing, and board vendors actually did use them also for marketing😅

  • @arrow207
    @arrow207 9 днів тому

    I didn't see a rundown on the older ROG Strix X670E-I mini ITX board, would that be preferred over the X870-I . I was about to pull the trigger on the X870 mini-ITX board but now I am hesitant.

  • @tali9908
    @tali9908 13 днів тому

    Hey man. I stumbled across your channel yesterday, and have since been watching your videos like crack 😂 Glad to see another’s just been uploaded.
    Quick question for you, without factoring in cost, do you have a favourite X870/E motherboard for OCing?

  • @WillFuI
    @WillFuI 14 днів тому +9

    Dang he must have taken down the 9950x 8000 video.

    • @sengan2475
      @sengan2475 13 днів тому

      saw that. any clue why?

    • @meda7291
      @meda7291 13 днів тому

      @@sengan2475iirc he mentioned not being able to disclose where he got the cpu from, so might be related to that

  • @laxnatullou
    @laxnatullou 11 днів тому

    Is this issue only specific for X870I board? I'm planning to get X870E Hero..

  • @Ronin-Wilde
    @Ronin-Wilde 11 днів тому

    Just curious, is there any improvement to transient response with these higher amp power stages?

  • @willjohnsonjohnson
    @willjohnsonjohnson 13 днів тому +1

    For a test setup, you'll probably end up measuring variations in the soldering rather than differences between the mosfets. I would like to see a no heatsink test for the motherboards. It still might not answer which VRM is more efficient, but at least we'll know which boards are least likely to overheat in worst case scenarios, like in a hot closet or shoved under a desk.

  • @DavidRain007
    @DavidRain007 13 днів тому

    You also have the Hero, right? Could You compare? It seems is uses Infineon’s TDA21570. At least those are the visible ones (memory?).

  • @piscanc1
    @piscanc1 13 днів тому

    darn...
    you are making me dig true my parts bin of CNC power stages looking for spec's

  • @MeakerSE
    @MeakerSE 13 днів тому +1

    The 70 Amp power stages on my B650-I (TDA21570) have a peak efficiency of 95% for 1.8V output, 5V input 800KHz at 20AMPs and only drops to the peak effieciency of these at 55 amps

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +1

      I guess that could be an intresting test. B650i vs X870i with no heatsinks and see which one ends up cooler.

  • @SnowGirl7890
    @SnowGirl7890 13 днів тому

    Could you do video about the X870e taichi or the x870e nova wifi, please?

  • @Chicken-o5e
    @Chicken-o5e 13 днів тому

    How efficient is the x870e nova board? Did they cheap out on anything on it?

  • @Anders357
    @Anders357 13 днів тому +1

    Also "bigger number better" which is why we get 24+2+1 vrms, instead of 10+2+1 good ones. It sells.

  • @roadkill11000
    @roadkill11000 13 днів тому +1

    This type of video is of zero practical use to me, but I watch anyway because BZ rant is entertaining.

  • @derfacecrafter1869
    @derfacecrafter1869 13 днів тому +3

    105A powerstages don't deliver 105A?

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +7

      Give the right operating conditions they certainly can but in any real world application the answer is no.

    • @derfacecrafter1869
      @derfacecrafter1869 13 днів тому +2

      @ActuallyHardcoreOverclocking i know. I have watched the Kingpin mobo Video.
      Should be a joke.

  • @Cinnabuns2009
    @Cinnabuns2009 12 днів тому

    This actually sounds completely in line with a company like this. 10-20% cheaper components, double the mobo price.... right in line with the 'number go up' thinking when you look at stock price. If you asked them a pointed question about it they'd just dismiss it and say, better ones weren't available.

  • @ruudtswaanen5369
    @ruudtswaanen5369 12 днів тому

    If possible compare powerstage prices, that might tell you why they use certain powerstages?

  • @Dave20VT
    @Dave20VT 13 днів тому

    could it be, like with CPU generetions, where you get a new gen cpu, with 10% more performance, but also 10% more power draw.
    so its like the same stuff, and performs better only because of higher power draw
    so i would think, the powerstage has only the benefit of handling more current, but loosing efficiency
    that would explain why a 100/110A powerstage is not better (or more efficient) than a 60A powerstage. It may only be capable of handling more current. what do you think?

  • @yossarianification
    @yossarianification 13 днів тому +3

    I am looking for an X870 motherboard for 9700X CPU, that would be most power efficient, especially on low CPU load. My typical use case is 8-10 hours a day with computer power on with web browser and video playback (I work professionally on another computer, and this one is for personal background tasks). At most a few hours a week I do more power consuming tasks, like gaming. The price of the motherboard does not matter. I just like it to be most power efficient in browser/playback scenarios (yet still provide reasonable power for gaming). Could you recommend anything? :)

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +10

      Do you have an approprietly size titanium PSU? because that'll have way more effect on your idle power draw than the motherboard will (I guess a mobo with less RGB might also help)

    • @ChrisM541
      @ChrisM541 13 днів тому +1

      If you're looking at a 9700X then price DOES matter. Irrespective of pricing, there is virtually ZERO difference compared to the X670. Saver yourself a lot of money and give yourself some self respect by purchasing an X670 for your 9700X instead of this shocking cash grab! You should even consider a 7000 series CPU, especially considering the use case.

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +7

      @@ChrisM541 I'm pretty sure B650 is better for power efficiency because it doesn't have to power a second chipset.

    • @atirta777
      @atirta777 13 днів тому

      @@yossarianification For the lowest power draw aside from the 80+ Titanium PSU, get any single chipset mobo (B650E/X870 etc.). Then either run your RAM at JEDEC settings or in 2:1 mode at like 7600MT/s so you can run your SOC voltage at 0.9V reducing idle/low load power even further + don't forget negative Curve Optimizer/Curve Shaper

    • @yossarianification
      @yossarianification 13 днів тому

      @@ActuallyHardcoreOverclocking I was thinking about SeaSonic Prime TX-850 850W titanium PSU - it is not ATX 3.x, but for 4070 with old-style PCIe power should be enough. I know that power efficiency of PSU is crucial in my objectives.

  • @stormrider01
    @stormrider01 14 днів тому +6

    12:05 Infineon is a German company, they make good stuff.

    • @nicholasvinen
      @nicholasvinen 13 днів тому

      I don't know a lot about power stages but I buy quite a few MOSFETs and Infineon makes some of the best.

    • @BeefIngot
      @BeefIngot 13 днів тому +1

      The nationality of the company does not determine the quality of the good

  • @seanoconnor8843
    @seanoconnor8843 6 днів тому

    Stability is important when these multi core processors burst into action. They maybe traded efficiency for current handling

  • @lolakyle8
    @lolakyle8 10 днів тому

    They need bigger number to ignite sales, just like polling rate on mouse and keyboard. This kind of video needs to be preached everywhere on the PC forum, or even better, present it in some kind of Reels/Shorts/TikTok/other short form video

  • @MikrySoft
    @MikrySoft 11 днів тому

    /Remember PMPO audio power ratings? Thos 5000W stereos? Also known as ILS (If Lightning Strikes) power?

  • @ziffer90
    @ziffer90 13 днів тому

    you building a testing setup for power stages seems inevitable

  • @bayanzabihiyan7465
    @bayanzabihiyan7465 9 днів тому

    My frustration is that we are in a world where the 16c pulls maybe 150A under full load whereas the 8C which most people are buying is like 100A.
    And we got these idiotic boards making 24phase VRMs of “110A” power stages because bigger numbers sell more.

  • @mattjvideos3663
    @mattjvideos3663 12 днів тому +1

    I recommend this if it's stable for you but get a kit of ddr5 8400mhz cl40 F5-8400J4052G24GX2-TZ5RW
    For me ryzen 7900x and asus x870e hero is the go to with this

  • @rrekki9320
    @rrekki9320 13 днів тому

    *smacks lips* Hey guys! :D
    Love your rambles man ^^

  • @TheDude50447
    @TheDude50447 12 днів тому

    Im at 6min now and theres a bit more to it. Peak efficiencies arent all that relevant unless you have a reliable way of consistently running at that effciency. So what manufacturers do for a actually good product is sacrificing some of that peak efficiency to make the part more efficient across a wider band of amps. Its hard to say if thats the case here cause the graph for the 820 stops at 40 Amps. Maybe it is garbage maybe it isnt.

  • @Flank.Sinatra.
    @Flank.Sinatra. 13 днів тому +1

    that 1.8 V output is the CPU PLL 1.8 V?

  • @prashanthb6521
    @prashanthb6521 12 днів тому

    I like to listen to your views because I am just an amateur. But it will be nice if you divide your videos into timestamped chapters.

  • @thaddeus2447
    @thaddeus2447 14 днів тому +1

    is msi x870 tomahawk wifi good board ?

    • @atirta777
      @atirta777 13 днів тому

      Yes. Great IO, very reliable DDR5-8000, M.2 versatility, ALC4080 audio, clear CMOS button on the back, postcode display
      Those were some of my reasonings to buy the board. I recommend seeing Buildzoid's PCB breakdown of it

    • @Slyons89
      @Slyons89 13 днів тому

      I bought one because it was the cheapest I could find with 7 segment debug display, and all 4 m.2 slots can be used without affecting the bandwidth on PCIe 16x graphics slot. I think it’s a reasonable board for the price with attractive feature set.

  • @andrewzravenz2444
    @andrewzravenz2444 13 днів тому

    Wouldn't the solution just be a standardized third party efficiency rating for complete VRMs rather than just the power stages? The Cybenetics database has certainly made it easier to find out if a power supply is efficient and quiet, should be the the same for motherboards and graphics cards.

  • @tristankordek
    @tristankordek 9 днів тому

    This problem affects all semiconductor components.
    👍🙂

  • @nicholaspeyton5460
    @nicholaspeyton5460 13 днів тому +3

    why don't you just email them and ask?

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому

      Because ASUS doesn't provide component information. Would kinda defeat the purpose of having special order parts if you just told people what the parts actually are.

    • @nicholaspeyton5460
      @nicholaspeyton5460 13 днів тому

      @@ActuallyHardcoreOverclocking How badly is ASUS misleading us?

    • @nicholaspeyton5460
      @nicholaspeyton5460 13 днів тому

      @@ActuallyHardcoreOverclocking Would this be a reasonable *analogy* ?
      Imagine a power strip that advertises "super clean power" for your high-end electronics. But instead of high-quality components to filter out fluctuations, it’s built with lower grade capacitors that can't quite keep up. So, while it technically _provides power_ , it has little spikes and dips-enough that your devices constantly flicker or even crash when you plug them in.
      ASUS’s "110 amp" VRMs are like that power strip. They promise high current, but with those cheaper components, they deliver power with extra ripple, meaning your CPU is getting a choppier, more erratic flow. It’s technically there, but not quite the smooth, stable power you'd expect from quality VRMs.

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +1

      @@nicholaspeyton5460 powerstages generally do not affect voltage regulation. Also I kinda expect the Strix X870i to be one of the best boards for voltage regulation(but not because of the power stages).

    • @nicholaspeyton5460
      @nicholaspeyton5460 12 днів тому

      Ideal = Very hot and extremely clean. Gives me an excuse to tinker/experiment:
      She cannae take much more, Captain! The VRMs are overheating, the MOSFETs are pushin' their limits, and the capacitors are startin' to strain under the load! If we don’t give her a breather soon, she’ll blow faster than a warp core breach! I’m givin' ye all she’s got, but this motherboard just wasn’t built for these kinds of amps!

  • @osterpenpen9379
    @osterpenpen9379 13 днів тому +1

    Is it possible that the efficiency curve is lower, but the power coming out of them is cleaner?

  • @concinnus
    @concinnus 13 днів тому

    Back in the Z390 days (on the Asrock ITX Phantom specifically), the power stages had a top metal tab for better thermal conduction. What ever happened to that idea -- surely it's too obvious to be patented?

  • @mightymartin6290
    @mightymartin6290 13 днів тому

    I still dont understand why the Asus ROG Strix X870-A Gaming Wifi motherboard has 90A power delivery vs Asus Rog Strix X870-F Gaming Wifi which has 110A power delivery the boards are literraly the same! Why asus?... Anyway I already have X870-A variant waiting for the 9800X3d price to come down.

  • @TheOneGhost12
    @TheOneGhost12 13 днів тому +2

    Can anyone explain why my 9800x3d has noticeable more input lag on mouse compared to 7800x3d? Do timings automatically change when you change cpu? I will literally hire buildzoid to fix this please? My 7800x3d felt like literally air at 8k polling and now my 9800x3d has slight delay moving the mouse. I already updated bios with newest bios before 7800x3d and fresh windows then got the 9800x3d

  • @fy7589
    @fy7589 13 днів тому

    How is 200 watts and 90 percent efficiency costs 22.222 watts of heat loss? It should be 20 watts man. I don't get that math can you explain how or why? 6:52

    • @parodie6666
      @parodie6666 13 днів тому +4

      The efficiency denominator is the input power, not the output power. 200/0.9-200=22.222

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +7

      200 / 220 = 0.909
      200 / 0.9 = 222.2222
      when talking about VRM efficiency the output power is fixed. IF you have a 200W CPU it will pull 200W regadless of the VRM being 80% or 99% efficient so the only thing that changes is the input power. So you take the output power and divide it by the efficiency to get the input power.

    • @fy7589
      @fy7589 13 днів тому

      @@ActuallyHardcoreOverclocking I just got bamboozled about it like 10% less of 200 is just 20 less and I thought linear not logarithmic, just silĺy me. Thanks

  • @toject8080
    @toject8080 13 днів тому

    10:30
    Even ignoring the methodology difference you are looking at the wrong line. The 100A plot has red line for 1.8V, 800kHz. The corresponding one in the graph you are looking at is the one in the middle, not the top one. 60A peak efficiency is ~93.5%@17A@800Hz@250nH, meanwhile 100A is ~92.8%@23A@800kHz@100nH

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +1

      no consumer motherboard runs at 800KHz because 800KHz gives you worse VRM efficiency and offers no voltage regulation benefits compared to 500KHz.

  •  13 днів тому

    97% seems of the chart.

  • @crashbug4343
    @crashbug4343 13 днів тому +1

    All marketing 😂

  • @EvilMcSheep
    @EvilMcSheep 13 днів тому +2

    >I'll try make this one shorter
    >50m runtime
    lmao

  • @james...cardinal
    @james...cardinal 12 днів тому

    if you ever wanna piss of bz go put a vishay ic on his asus board

  • @maksonamex
    @maksonamex 13 днів тому

    one thing understood from last 5 years , im never buying asus products again

    • @pcoverthink
      @pcoverthink 13 днів тому

      I wish there was something better. Msi boards are falling off a cliff in the budget segment in terms of memory ox capabilities. Perhaps gigabyte got better in the past 5 years but it was ass before. Gpus are sometimes a miss from everyone but they're mostly fine

    • @rci-tf2zc
      @rci-tf2zc 13 днів тому

      @@pcoverthink asrock taichi

    • @TheTastefulThickness
      @TheTastefulThickness 13 днів тому

      Love my msi z490 ​@@pcoverthink

    • @pcoverthink
      @pcoverthink 13 днів тому

      @@rci-tf2zc dealing with ASRock bios was always one of the worst. Hopefully it's better these days and perhaps lightning is worth it price/perf wise

    • @pcoverthink
      @pcoverthink 13 днів тому

      @@TheTastefulThickness my b650 tomahawk had recovery feature up until 1y ago. In the newer bios versions they just removed so it never recovers with unstable timings and you have to clear cmos which gets super annoying. Even when it was available, it was disabled by default

  • @ChrisM541
    @ChrisM541 13 днів тому

    X870 - THE most pointless release in the history of pointless releases!

  • @spewp
    @spewp 13 днів тому

    "Today I want to complain about something that annoys me"
    This could be the subtitle for all of your videos tbh.
    Gonna delete/hide this video like the last one?

  • @rielleleaf
    @rielleleaf 13 днів тому +1

    10:36
    The 840 is leading by a bit at 30A.
    ---
    at 1.8V 800Mhz
    20A: 620 has ~93.3%, 840 has ~92.6%
    30A: 620 has ~92.2% 840 has ~92.8%
    40A: 620 has ~90.6%, 840 has ~92.1%
    At least the 840 is leading at and above 26A by a bit. Too bad the 620 spec sheet does not show 1.2V Vout graph.

    • @SkylerAk
      @SkylerAk 13 днів тому

      On amd anything greater than 25 amps is seriously unrealistic, we are talking over 250 watts of power draw average on the cpu not to mention the 30 or so watts of heat from the power stages alone. Realistically the best choice would have peak efficiencies in the 10-20 amp range.

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому

      motherboard VRMs typically run at 300-500KHz. The X670E Gene which uses the SiC850 runs at 500KHz by default.

    • @rielleleaf
      @rielleleaf 12 днів тому

      @@ActuallyHardcoreOverclocking that was the switching frequency I compared as that was the only common SF between the two charts shown, IIRC!

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  12 днів тому +1

      @@rielleleaf yeah I know but efficiency at 800KHz isn't really useful for motherboards

  • @ConeJellos
    @ConeJellos 13 днів тому

    Don't care.

    • @ActuallyHardcoreOverclocking
      @ActuallyHardcoreOverclocking  13 днів тому +13

      then why are you here?

    • @Slyons89
      @Slyons89 13 днів тому +1

      You posting this comment boosts the engagement metrics on the video lol

    • @BeefIngot
      @BeefIngot 13 днів тому +1

      You realize UA-cam now thinks you want more of these videos because you commented right?

  • @fracturedlife1393
    @fracturedlife1393 14 днів тому +7

    rant. INFO DUMP BOYS!!!