The “six gear power shifting” thing is a software feature on these boards. Where you can manually control and/or see how many phases the board uses to deliver the power to CPU. It mainly is a power saving feature, by default it uses all phases.
Post codes are another frustration with motherboard makers. The post codes are already built into the BIOS so it's dirt cheap to implement yet only the highest and motherboards get it. You'd figure every motherboard by now would have it because it makes troubleshooting so much easier. Yet they pretend it's a premium option.
Market segmentation. It's worse now than it was back then. I remember boards like the ASRock P67 Pro3. 8+2 phase VRM, postcode/debug LED, power and reset switch onboard and it was $99.
@@-eMpTy- My decade-old, entry Biostar board has a 2-digit display. Ever since I started looking to build my own PC again a couple years ago, I can't believe that almost no manufacturer puts decent debugging tools on their mobos anymore.
Furthermore I'm surprised OEM motherboards (HP, Dell, acer) don't already have them for their business motherboards so their techs can fault diagnose quicker
the more useful version of the rant can be: how do we come up with an actually good, quantitative measurement that can describe how good a vrm design is. That could really benefit the consumers.
This is similar to how can we objectively measure a quality of a PSU? All the manufacturers brag about is the 80+ spec, but that only covers your efficiency, not quality in other regards. See the recent GN PSU disaster videos as an example... And that's where reviewers like BZ or GN come in to tell you a product is a garbage even if it claims to have more of a thing in its design.
@@emperorSbraz The POST-code display is used as a countdown timer for how much more training it needs. But for real, I'd buy that for Zen4 or Zen5, if it works for ECC DIMMs.
Thanks for reviewing my motherboard from 12 years ago! Got some quality time pairing this board with i7 920 and GTX 295. Interesting to see back in the day high-end enthusiast set up only coast me around $2k and now.... cant even afford a video card sigh...
We literally didn't know how VRM worked at the time. I used to maintain a database on AMD board VRMs on OCN because so many were absolute garbage. Im not an EE or anything of the sort, so our method to quickly evaluate a board was basically to count inductors. Naturally we know now this is not accurate, but at the time it gave a bit of guidance on to which boards would burn themselves to the ground (i.e. MSI 790/890FXA-GD70). Especially because as BZ notes reviewers were similarly clueless and often worse. I apologize for perpetuating the issue, but it was born out of necessity after we put holes in 10+ 790/890FXA-GD70s
there are double step-down two-phase buck converters, today there are some tripple step-down three-phase buck converters available. by looking at a component you will not know how it works.
22:20 Buildzoids 2nd amendment "A well regulated VRM, being necessary to the stability of a system, the right of the boards to keep and bear Amps, shall not be infringed"
This is just "enhanced" Dynamic Energy Saver feature which was present on even older Gigabyte's motherboards. Dynamic 6-Gear Switching is mentioned on some LGA775 motherboards they manufactured.
18:50 that's not what diode emulation does; at light loads if you leave the low side MOSFET on like you normally would, due to low inductance and current(so low stored energy) the current flow can reverse, flowing from the output, through the inductor, through the lowside MOSFET to ground killing efficiency. Diode emulation mode senses the voltage across the MOSFET changes from negative to positive (indicating the change in current direction) and turns the MOSFET off so the output current doesn't flow to ground.
@@sophiethemasochisticninja7655 In case my links get filtered out. Look for Vortez's review of the Asus Z590-E where they call it a 14 phase, when it isn't.
@@Stashmyash Specification says 14+2. Looked at a picture of the motherboard without vrm cooling. it has 14 + 2 power stages. so not sure why you say it is false marketing.
If you care about efficiency at low load, as in laptops, it's also important that the VRM have low stored energy. When the CPU transitions from boost to idle state, that Δ(CV^2) is wasted..
No...you weren't the first person. The forums were full. Not even close. :) But I think you were one of the first on Yt. And a lot more exposure than everybody else. :)
It do be like that. A proper rant for an isolated topic takes half an hour at the very least. There's the context of the rant, the theory crash course to understand the rant, the rant itself, the humours roasting as to why the rant exists in the first place, the rant itself episode 2, the first attempt at rounding off the rant, the rant itself episode 3, the second attempt at rounding off the rand, the first, second and third random semi related anecdote to the rant, and finally the somewhat staccato/abrupt ending which includes something about subscribing and merch.... It just doesn't feel right if it doesn't go like that.
I did get one of these UD5 boards at launch, flashing it to Extreme after I waterblocked it. I liked Gigabyte's skt.1366 boards so much, I got an X58A-UD7 later and at the time, I remember quite distinctly people already knew the phases were "doubled" in the boards. The UD5 and Extreme having 6 real phases (and the UD7 having 12 real phases), was well know among the PC DIY enthusiast community as soon as 2009. Say whatever you want about these boards' phases, the truth is those were the real deal. My UD7 was retired last year, after powering an i7-920 D0 @4.5Ghz for 8 years straight! And was retired not because it died but, because it was time to move on. My UD5 later flashed to Extreme, is still working 24/7 with an i7-920 C0 @4.0Ghz since November 2008.
Encouraging to know that MB's can have such durability - I notice part of the blurb for this Gigabyte X58 is that it uses "50,000 hour lifespan Japanese solid capacitors", which is 5.7 years. Gulp. According to CrystalDiskInfo my Asus P8Z77-V has been powered on for just shy of 47,500 hours (5.42 years) since I built it in Nov. 2012. The P8Z77-V does actually seem to have the 8 CPU VRM phases advertised BTW.
Yeah. There's a theory going arround amongst seasoned PC builders that leaving the PC turned ON keeps the hardware free of problems for longer. From my personal experience, I certainly agree with that theory. I know it's not the most politically correct thing to say in current year but, if I can leave my PC turned ON 24/7, I just do it.
@@johnny_rook I'm well over being PC(!) and in recent years it's been left on 24/7. I discovered after buying a UPS about 2 years back that it's idling at only 52w at the wall (kudos to Asus), which sealed it for me. But yep, in electronics and engineering generally, power/thermal cycling is reckoned to account for a big chunk of the life of components.
@@arbiterodie7685 :D Had to adjust voltage from time to time but, yeah, it hold. It was a D0 step, though. MY 920 C0 only did 4.2Ghz max but not sustainable for 24/7
3 ปีที่แล้ว +3
Im getting old, still remember. And yes people knew about the phase scams. I still remember the 24 phases boards XD Usually they just had two choke per phase. I did remember looking at just how the choke where hooked up on boards when people wanted high phase boards to try and figure out what boards where scames because that was really a thing in early i5/i7 era boards.
Yup and it still goes on everywhere. Just look at Nvidia with Ampere architecture. "Double the CUDA cores!" Except it's not double, they just have double the FP32 throughput compared to previous SM design. Seems the PC market has been riddled with false advertising for decades. I guess the motto is "well nobody caught us so we did nothing wrong."
3 ปีที่แล้ว +1
@@The_Man_In_Red if you ask some people an ALU is a core XD There are fundamental parts you need to be able to call it a CPU or a Core or well a phase a phase in a vrm design. If I recall Ampere they just added an extra FPU so making the core wider but calling it two cores or cuda cores. I suppose a Pentium 54C is a tripple core then with two pipelines with two alus and a fpu XD it was just simpler back when engineers decided the marketing and not the marketing people.
"They've all been lying about them, for YEARS. Nobody cared until I came along, and actually, I'm still convinced that the only person who actually cares is me!" I had a good laugh with you over that. I love the rambling, as that's where the depth comes from. It's a shame that I haven't been availably online during another one of your live streams, I caught one maybe a year ago, and somehow the conversation ended up going into places we've lived, where you studied, the origin of your accent(even though I can hardly tell there's an accent). It was good fun.
I feel totally schooled! lol nice job and explanation, I remember the early VRM's that came out for the 3.3v versions of '486 which appeared in their twilight, back then they were a plug in module that went into a slot next to the cpu, they were required if not running a 5v part. thanks again I learned a lot .subbed!
At that time it was a phase count race between manufacturers. The marketing was doing it's job because bigger number obviously "is better than lower". Check the X58A-UD9 the "top" from that era with "Smart 24 power phase design with mutual back-up to each 12 phase"
I have this exact motherboard in my daily rig, paired with a 6C/12T Xeon clocked at 4GHz & 24GB of triple channel RAM. Happy to loan it to you if you can pay for shipping from Australia!
Been doing pc since 2000 "Comments are saying.. x brand is bad because they're lying about phase counds" [uploads videos pointing those who lie about phase counts or components] Its not the brands, they all do the same things, they're in the same boat. its also not a new problem, but I dont remember it being so common in the mid range (£170 today is £110 then) Back in the day high end motherboards would say 8+2.. when its 4+1. But if you bought a cheaper board you kinda knew it would be a 4+1 because there's no way the gigabyte 970a-ud3 has the 8+2 phase it claims at £90. Now people see a huge heatsink and it could be a triple teamed 4+1, or double teamed 6+1 or doubled 6+1 or an xdpe132g5c.... The higher phase controllers and doublers and teaming leaves allot more room for (mis)interpretation. Thats why ahoc and HUB is so important for people to learn, Sunlight is the best disinfectant so videos saying what it is and what it isn't, what matters and what doesnt are invaluable. Going parallel for more current handling is fine, the problem is the marketing. I'm wondering how long it is before we see "16+2 stage *BFVRM*" with 4c06n/4c010n parts using a RT8894 or something (turns to look at asrock) atleast it would have surface area. Fun trivia: msi 970 gaming and krait am3+ were both 3+1 with "teamed" niko mosfets, the boards literally cried with tears running over the audio section.
I have a Question! How come some traces on the circuit board do a bunch of S turns and squiggles instead of going in the shortest possible path? hope that makes sense
I think the reason for that came from that the media (computer magazine, online reviewer) always educate the DIY fans how to count the number of „phases“ on the motherboard by counting the number of solid capacitors around the CPU socket. Which is not always true but however historically always useful because no one used more than one solid capacitor for each phase. However the manufacturer got that in mind and pushing the number of solid capacitors to the limit, so they can confuse buyers and be fancy on the advertisement. Either the marketing department is doing this willingly or they were as uneducated as the casual audience as well.
I remember seeing an interview with an Intel employee several years ago who said 4-phases was enough and double digit VRM setups weren’t needed unless they were using poor quality components. I guess this is the answer to why.
The correct way to advertise imo is: how many amps the board can delivery while run non-stop without shortening the lifespan of the components, throttling or critically failing (i.e. by not overheating). If they wanted to go above and beyond, they'd publish an entire curve profile of how it behaves temperature-wise as the amperage increases
ferrite core is a cost and size effective way to build an inductor with the downside of having more distortion. the other way would be air core, a normal thing to see in 90s and early 2000s boards.
@@BrosBrothersLP The video acts as if nothing has ever been used before. PC building did not burst out full formed in 2010 or whenever. It is painful obvious at the start of the rant and at that point about chokes the author functions as if it did. That is fine, but for some of us, 2008 isn't the long long ago.
I like this video. It proves my point done ~20 years ago - at each CPU generation the corresponding motherboards are just cheap spare elements except 2 or 3 models in high-end/workstation class.
Mine also has trouble with its RAM slots, wonder if it is a particular desease of the board ? Also she (yeah it is alive) refuse to boot for days some time until the deamon goes away.
the CPU connection pins are still bent, the knowledge of this person is simply underground for what he is presenting here. He doesn't even know that there are double step-down two-phase buck converters and so the specification of 12 phase count is probably correct.
Yes you are right, forgot his bent pins. May he have missed them then, I think he has enougth knowledge to cover that (think it is its field). I don't. But even with doblers wich where used a lot back in these days, a true 12 independant phase wich lead to a better ajustement of the load. Maybe it is more relevent now with all CPU states ans boost.
@@mini-pouce i'm a trained microelectronics technician and nothing of what he tells in the video comes from a corresponding education. Not only are there double step-down two-phase buck converters (in one housing and visually indistinguishable from single step-down single phase), today there are even triple step-down three-phase buck converters in use and the next 4-phase design is under development (Die stacking comes from this industry, by the way). He has no idea which firmware is running in the Buck converter controller or how a single phase buck converter can switch two phases without a time difference at the input of the CPU. VRMs are not so expensive components for nothing. The mainboard manufacturers have their "secrets" and only trained microelectronics are able to really understand the circuitry of a mainboard. There are a number of physical tricks that have been used on mainboards since the 90s to create the same functionality with fewer components. Without weeks of measurements and precise knowledge of the tracelength, it is not possible to make statements on highly integrated technology as he makes it about this board, after all it is a 7-layer mainboard and he has no idea how the traces are laid within the PCB, he sees only the external power connections and this leads to pure nonsense from an electrical engineering point of view. he can neither show nor name the name on the converters, so it is not possible to understand what kind of component it is. I do not believe that a manufacturer made false information about his product in 2008, but simply that it is a two-phase buck converter and the guy has no idea about it. Here in europe, incorrectly specifying product properties would lead to gigantic lawsuits and in the 90s there were some through which hardware manufacturers dug their own grave with a lie. The fact is, I have a mainboard here where 12 phases are mentioned in the manual and there are only 4 converters on the board and the reason is that there are three-phase components.
@@mini-pouce if you want to learn something about phase regulators and how it is possible to put more than one in one housing look at this video.(there are still 8 possible in one housing but not used by computer electronics because there are other problems with more than 3 or 4 at the moment) th-cam.com/video/dqv2UoSEmCY/w-d-xo.html also the maximum current you can put at the output pin at one housing is limited to 40 amps (breakdown current, some use 44 amps but this converters/regulators are the ones that dying on Graphiscards at overload) that's why you have more than one housing and some more DCR Sense circuits behind the converters.
I had a Gigabyte GA-X58-UD7 rev 1.0 board and that was advertised as 24+2+2 phase design but i think that was only 12 phase. I dont have that board anymore so cant check or send pictures.
and that is why, people like you are so important. To educate consumers about these things. It now only needs to get out of this bubble by some bigger YTer and maybe the Manufacturers will change the marketing.
Hello there, since you brought up some older stuff, I think taking a look at some Gigabyte GA-8KNXP would be interesting. Maybe the Ultra-64 would be the best one among them.
Can i ask one more thing. I have no knwldge abt electronics. But, is the inductor placed near the cpu power socket important ? For power delivery ? I'm seeing lower end boards having those but comes with "poor"(to my uneducated eye) power phase, while some other boards without inductors near the cpu 12v powersocket but, have some beefy power phase. Any advise ? You can check recent Intel H510 boards(Asus, Msi, Gigabytes).
A simplified expression to indicate a capacitor with a solid conductive polymer electrolyte. Also, saying "It looks like a smt electrolytic" is kind of broad. For your information: - smt just indicates the mount type, ergo smt/smd = surface mount, - all capacitors have some sort of electrolyte.
@ Actually Hardcore Overclocking Phases ... this means that there is a phase-shift during the switching for each MosFET-Gate in the high-side-driver and also in the lowside-driver (this MosFET is the replacement of the diode). What benefits do we get from this? The input current is more even. If all Switching-Regulators would pull current to the same time, then the input-capacitors would have a lot to do. This phase-shift makes it possible to reduce the numbers of the input-capacitors and the output-capacitors. There is a lower ripple-current and this is good. 6 phases is very good. In reality people don't care, this things have to work. What people want to know are the Japan-Capacitors, because this Chinese motherboards had the problem in the past, that the capacitors are blowing up. The Motherboard-Producer did buy very often cheap Elkos from unknown source or crate a cheap design. Normally you could decrease the stress on the electrolytic capacitors (heat production because of the internal resistance) if you add some ceramic capacitors, but they did not and save the money.
One additional point is, just like when you put two resistors in parallel, putting two MOSFETs in parallel also lowers the resistance. So even though its a 6-phase, its actually appears to be a pretty good 6-phase. -Matt
I've got this board from back in the day. I didn't know enough about VRMs to know they lied, but I remember there were many "features" of this board that seemed a bit over the top. Still, it was a good board. Still works.
I watch you and gamers nexus BECAUSE, you tell the truth about companys. And their mistakes/bad products, with no hands/fingers in betwean. So yes, it is a good idea to tell someone they did something wrong, cos if no one does they will never learn to be honorst and do the rigth thing with the build and markedting of their products.
Being Gigabyte reminded me of this Motherboard with a weird power design GA-8N-SLI Royal ( has a separate add on card for power / and a PCB to switch between SLI / Crossfire ) I had one but RMAed because power card went pop in a month or so
Oh man the whole "six gear power shifting" stuff was absolute bs nonetheless fun times I feel like the X58 boards were completely insane and wild with the wacky colours from each manufacturer lol
Can you do the same with GA890FXA-UD5 rev 2.0? I am running FX 8350@4.8 with beta bios and works fine since i have started to use FX cpu. Max, stable oc is 5.0 for my cpu on watercooling. What we use is an unofficial bios from gigabyte worker. Funny thing is, officially this board doesn't support FX cpu's. :D
outside of msi or ecs, in the late 2000s enthusiast motherboards were typically overbuilt, but dumb for the cpu's at the time. The problem is they became too efficient with their designs.
I've been building computers as needed (and mostly as a hobby, but also for my business) since the early 1990s. I don't remember ever hearing anything about VRMs or phase counts or anything of the sort for my first couple decades of doing so. It wasn't until I looked into building a gaming PC a few years ago that I noticed this new terminology. I had to research what it even was because it was entirely foreign to me. I'm not an electrician. I'm not a physicist. It was like they decided to overwhelm us with technospeak in order to sound better than the competition. I guess that sells.
Motherboard manufacturers started marketing phase counts because more numbers more bettar. Especially for Intel customers. Used to be "more MHz more bettar". Then when Pentium 4 architecture hit its temperature ceiling, it became "more FSB MHz more bettar". And so forth.
Had a Gigabyte 1366 motherboard, and ooh god it sucked. Maybe I was just unlucky but RAM support was a maybe, sata controllers was more like erasers, and as soon as anything voltage or stepping related was touched it would just not post at all.
Man, I remeber having a MSI motherboard with a nForce chipset. It was the first motherboard that I bought with my own money, it was for a Athlon XP CPU and that's all that I remember ( it was around 2001 or 2002 ). Back then I didn't even care about power delivery and how it worked, I'm not sure hardware reviews for motherboards even existed lol. It would be interesting to analyze old power delivery systems, if there were any special/insane ones :)
This board looks like it has the same VRM as my fully working UD3R v2.0 that i got from a friend for 100czk with an i7 960 D0 and an original big 1366 box cooler :D
@@n0rie9a I have the rev. 2.0 and it has 12 inductors around CPU. It also has 2xUSB3.0 ports on the back and 2xSATA 6G ports from a separate controller on the board+another 2xSATA 3G from another controller on the board-GBT kinda overdid the board with extra controllers besides chipset.
please great GURU of HW can you tell me if the new asus tuf b450 gaming plus 2 as vrms enought power to suport an ryzen 5600X on OC??please. i switch myne msi b450 gaming plus to the tuf an got an ryzen 2600X and pretend switch to the 5600X in september.i have an aio from artic frezeer 2 240mm and im very satisfeid with this tuf but i trust your brain opinion better then any other guru out there. please just take a look at the tuf and post someting about it. i already know the board as not the 8+2 vrms they say. is 4x2 + 2 phases. but they really look well done and i notice better performance out of the box just from switching from msi to this tuf.so im confused.
FX 5200 taught me about problem how a hot GPU die and VRAM do not work very well and how it can be easily fix with something like a Iceberg kit. FX 9590 taught me about need for more power and the problem with hot Vrm and CPU. well here I made my first DIY fan Bracket and Frankenstein together my first open loop cooling system from parts of AIO's and other parts I could find in your local hardware store.
I had a asus P5Q Deluxe and it advertised 16 phases, It doesn't matter how much it really had because it overclocked like crazy in comparison to my older P5N-D board :D
the '00s were the dark ages for the pc DIY segment (there were darker ages, but mostly talking about late '00s as fairly modern times). almost all reviews were blurb copy-paste jobs, even the so-called hardcore/expert sites/outlets. information like chips/controllers manufacturers/models of a mobo/ram module or an actual examination of a mobos power delivery/components wasnt publicly available/easy to find. youd have to dig deep and hard to come up with even a small part/any of the above and very rarely some obscure outlet/review site would hold any info. any useful info about a mobo/ram modules was like a top CIA secret. but us hardware guys knew about gigabytes 6 phase/asus's 8 (high-end mobos)/4 phase. OC/technical forums were full with discussions about that. ram chips info was even harder to get a hold of. the guys who knew were guys like you, but "unpublished" cause no one was doing this type of content back then
I have a gigabyte board here that has an add in card to add more phases ... it’s like a early 2000’s. Interesting board . It’s wild. Gigabyte something something ultra
I had this motherboard back in the day. It actually overclocked fairly well from memory I was able to push 4ghz out of my i7 920 with a fairly basic oc.
comparing to other brands in same era, gigayte does run really hot. turns out most asus boards survived to today (p5q and p6) and only a few gigabyte boards survived.
God I miss X58 chipset. I still have my i7-920. I used to overclock that thing from 2.66ghz all the way up to 4.0ghz on all cores stable. That shit was wild! I had an MSI X58 Platinum SLI. I had 12GB of 1366mhz Runing in triple channel.
My Gigabyte Z270x Gaming 9 apparently had a 22 phase VRM. That board was awesome though, the only thing I didn’t like was it was Terrible for memory OCing.
I always wondered why they started advertising the phase count at all, let alone such a ridiculously high count as 12. You get diminishing returns for each one you add and it's really only so that you can use less capacitance to maintain low ripple. With a big enough capacitor you can get by with just one phase.
Well, but your average buyer doesn't know this and simply sees "bigger number = more better". So bigger numbers in marketing sells your shit and that's all the job a marketing department has to do. That's your why.
What is the technical or economic reason for having "phases" in the VRM at all. Why not just use components of sufficient quality and power handling ability to manage MB power? As a graduate electrical engineer from back in the dark ages, phase to me normally means power line phase in AC whereas MBs tend to operate from DC power unless I have missed something in the last few years (I probably built my 1st PC in about 1995 and swapped components since the late 1980s).
something yada yada sine curves, phasing them means more curves but they aren't 100% overlaid on each other but "pushed apart" so the "dips" in the power is removed which gives smoother power (you can look it up for more correct terminology)
BZ: 20:28 "That's it for the video."
Me: looks at 29:44 run time, not surprised at all.
0:02
Pretty par for the course, honestly. I think id have felt disappointed otherwise.
This is classic zoid.
Listening to Buildzoid rage at mainboard manufacturers really never gets old.
Right
The “six gear power shifting” thing is a software feature on these boards. Where you can manually control and/or see how many phases the board uses to deliver the power to CPU. It mainly is a power saving feature, by default it uses all phases.
So does it say that it is using 6 phases in software?
Post codes are another frustration with motherboard makers. The post codes are already built into the BIOS so it's dirt cheap to implement yet only the highest and motherboards get it. You'd figure every motherboard by now would have it because it makes troubleshooting so much easier. Yet they pretend it's a premium option.
Couldn't agree more. Two segment LED displays can't cost more than a penny or two?
Market segmentation. It's worse now than it was back then. I remember boards like the ASRock P67 Pro3. 8+2 phase VRM, postcode/debug LED, power and reset switch onboard and it was $99.
@@-eMpTy- My decade-old, entry Biostar board has a 2-digit display. Ever since I started looking to build my own PC again a couple years ago, I can't believe that almost no manufacturer puts decent debugging tools on their mobos anymore.
Because we (who want post codes) are are tiny minority compared to the masses who don't care. They, those masses, just want all the RGB in the world.
Furthermore I'm surprised OEM motherboards (HP, Dell, acer) don't already have them for their business motherboards so their techs can fault diagnose quicker
"the board is a 12-phase design"
"you'd have like 11 or 12"
*ACCIDENTALLY CORRECT PHASE COUNT FTW!*
This made me laugh.
Headline *BZ accidentally proves incorrect reviewer correct*
And then it says the board has additional phases, sooo... Still incorrect.
BZ: "that's it for the video"
also BZ: Continues to ramble for another 9 minutes.
the more useful version of the rant can be: how do we come up with an actually good, quantitative measurement that can describe how good a vrm design is. That could really benefit the consumers.
Energy wasted (W) when powering the flagship CPU at a standardized CPU stress test benchmark.
The way all voltage supplies are quantatively tested.
Outout ripple. Output spike. Dynamics. Regulating accuracy, efficiency
This is similar to how can we objectively measure a quality of a PSU? All the manufacturers brag about is the 80+ spec, but that only covers your efficiency, not quality in other regards. See the recent GN PSU disaster videos as an example... And that's where reviewers like BZ or GN come in to tell you a product is a garbage even if it claims to have more of a thing in its design.
Motherboard manufacturers have only been lying about phase counts for as long as they've been advertising phase counts on the box
just patiently waiting for the day a buildzoid edition board comes out
It'll be 50% VRM, 50% memory channels. The CPU socket will be left as an implementation exercise for the buyer.
@@MoraFermi 😂😭 swappable controllers so you can do AMD OR intel imagine 😭
mem training at first boot takes 1 literal day to post BUT it's already maxed on all timings. :D
@@emperorSbraz okay that's funny 😂
@@emperorSbraz The POST-code display is used as a countdown timer for how much more training it needs.
But for real, I'd buy that for Zen4 or Zen5, if it works for ECC DIMMs.
Thanks for reviewing my motherboard from 12 years ago! Got some quality time pairing this board with i7 920 and GTX 295. Interesting to see back in the day high-end enthusiast set up only coast me around $2k and now.... cant even afford a video card sigh...
“& I took that personally” - Buildzoid, 2021
We literally didn't know how VRM worked at the time. I used to maintain a database on AMD board VRMs on OCN because so many were absolute garbage. Im not an EE or anything of the sort, so our method to quickly evaluate a board was basically to count inductors. Naturally we know now this is not accurate, but at the time it gave a bit of guidance on to which boards would burn themselves to the ground (i.e. MSI 790/890FXA-GD70). Especially because as BZ notes reviewers were similarly clueless and often worse. I apologize for perpetuating the issue, but it was born out of necessity after we put holes in 10+ 790/890FXA-GD70s
there are double step-down two-phase buck converters, today there are some tripple step-down three-phase buck converters available. by looking at a component you will not know how it works.
22:20 Buildzoids 2nd amendment
"A well regulated VRM, being necessary to the stability of a system, the right of the boards to keep and bear Amps, shall not be infringed"
Hey Buildzoid, this is guys here!
"I'm still convinced that the only person who actually cares is me" - hahaha :)
The rest of his thousands of subs...
are we a joke to you?!
@@bookworm8415 If you watch the video you'll see that I just quoted what Buildzoid said because it made me laugh ...
"6 gear" sounds like they knew it wasnt a true 12 phase but didnt have a clue how it worked. (16:06)
Actually it's something what's stated on gigabyte website. It's on sixth line(huh) in marketing material. So unfortunately they just copied it :(
This is just "enhanced" Dynamic Energy Saver feature which was present on even older Gigabyte's motherboards. Dynamic 6-Gear Switching is mentioned on some LGA775 motherboards they manufactured.
"Hopefully I keep it kinda short" 30 minutes later.
30 minutes is kind of short for buildzoid, but I, too, was amused by this.
When he says short he means 30-40 minutes. "Normal" is 45-60 maybe. Never paid much attention before to notice trends
The Gigabyte marketing team are preparing to take it to the next level.
Triple digit VRM's.
18:50 that's not what diode emulation does; at light loads if you leave the low side MOSFET on like you normally would, due to low inductance and current(so low stored energy) the current flow can reverse, flowing from the output, through the inductor, through the lowside MOSFET to ground killing efficiency. Diode emulation mode senses the voltage across the MOSFET changes from negative to positive (indicating the change in current direction) and turns the MOSFET off so the output current doesn't flow to ground.
That it is for the video. Video continues for another 10 minutes...gotta love this guy.
That’s why we have you, we need your guidance, VRM, phases and all the things we need to understand.
Intel’s 14nm process is extremely new.
I learn new things every day!
Your rants are epic and I love how many times you wrap up but then restart the rant again. Thanks for sharing.
I finally beat buildzoid at something. I knew what a motherboard was before he did.
Appreciate your effort to keep pointing them out to hopefully keep them honest
Some motherboard reviewers still take (fake) phase counts from the manufacturer at face value. Nothing has changed since 2008.
Example of something new. Source?
@@sophiethemasochisticninja7655 There are several on youtube and the web, it shouldn't be difficult to find one.
@@Stashmyash Well im not the one trying to prove something or claim something. So source on your claim is something you will have to give.
@@sophiethemasochisticninja7655 In case my links get filtered out. Look for Vortez's review of the Asus Z590-E where they call it a 14 phase, when it isn't.
@@Stashmyash Specification says 14+2. Looked at a picture of the motherboard without vrm cooling. it has 14 + 2 power stages. so not sure why you say it is false marketing.
Hey look, that board has QPI that does 6.4 GT/s. What a neat unit to measure by.
If you care about efficiency at low load, as in laptops, it's also important that the VRM have low stored energy. When the CPU transitions from boost to idle state, that Δ(CV^2) is wasted..
Chevrolet Chevelle
- 6 cylinders
- 4 doors
Yugo
- 4 cylinders
- 3 doors
Koenigsegg Gemera
- 3 cylinders
- 2 doors
Which one has the best performance?
No...you weren't the first person. The forums were full.
Not even close. :)
But I think you were one of the first on Yt. And a lot more exposure than everybody else. :)
so why didn't any of the reviewers ever point this out?
At least german reviewers never mentioned such things. Most of them do not even know how to count phases.
> "hopefully I manage to keep it kinda short"
> looks at 29:44 duration
Perfection
It do be like that. A proper rant for an isolated topic takes half an hour at the very least. There's the context of the rant, the theory crash course to understand the rant, the rant itself, the humours roasting as to why the rant exists in the first place, the rant itself episode 2, the first attempt at rounding off the rant, the rant itself episode 3, the second attempt at rounding off the rand, the first, second and third random semi related anecdote to the rant, and finally the somewhat staccato/abrupt ending which includes something about subscribing and merch.... It just doesn't feel right if it doesn't go like that.
I did get one of these UD5 boards at launch, flashing it to Extreme after I waterblocked it. I liked Gigabyte's skt.1366 boards so much, I got an X58A-UD7 later and at the time, I remember quite distinctly people already knew the phases were "doubled" in the boards. The UD5 and Extreme having 6 real phases (and the UD7 having 12 real phases), was well know among the PC DIY enthusiast community as soon as 2009.
Say whatever you want about these boards' phases, the truth is those were the real deal. My UD7 was retired last year, after powering an i7-920 D0 @4.5Ghz for 8 years straight! And was retired not because it died but, because it was time to move on. My UD5 later flashed to Extreme, is still working 24/7 with an i7-920 C0 @4.0Ghz since November 2008.
Encouraging to know that MB's can have such durability - I notice part of the blurb for this Gigabyte X58 is that it uses "50,000 hour lifespan Japanese solid capacitors", which is 5.7 years. Gulp. According to CrystalDiskInfo my Asus P8Z77-V has been powered on for just shy of 47,500 hours (5.42 years) since I built it in Nov. 2012. The P8Z77-V does actually seem to have the 8 CPU VRM phases advertised BTW.
Yeah. There's a theory going arround amongst seasoned PC builders that leaving the PC turned ON keeps the hardware free of problems for longer. From my personal experience, I certainly agree with that theory.
I know it's not the most politically correct thing to say in current year but, if I can leave my PC turned ON 24/7, I just do it.
@@johnny_rook I'm well over being PC(!) and in recent years it's been left on 24/7. I discovered after buying a UPS about 2 years back that it's idling at only 52w at the wall (kudos to Asus), which sealed it for me. But yep, in electronics and engineering generally, power/thermal cycling is reckoned to account for a big chunk of the life of components.
A i7-920 doing 4.5 Ghz for 8 years? Nice OC, my man
@@arbiterodie7685 :D Had to adjust voltage from time to time but, yeah, it hold. It was a D0 step, though. MY 920 C0 only did 4.2Ghz max but not sustainable for 24/7
Im getting old, still remember. And yes people knew about the phase scams.
I still remember the 24 phases boards XD
Usually they just had two choke per phase.
I did remember looking at just how the choke where hooked up on boards when people wanted high phase boards to try and figure out what boards where scames because that was really a thing in early i5/i7 era boards.
Yup and it still goes on everywhere. Just look at Nvidia with Ampere architecture. "Double the CUDA cores!" Except it's not double, they just have double the FP32 throughput compared to previous SM design.
Seems the PC market has been riddled with false advertising for decades. I guess the motto is "well nobody caught us so we did nothing wrong."
@@The_Man_In_Red if you ask some people an ALU is a core XD
There are fundamental parts you need to be able to call it a CPU or a Core or well a phase a phase in a vrm design.
If I recall Ampere they just added an extra FPU so making the core wider but calling it two cores or cuda cores.
I suppose a Pentium 54C is a tripple core then with two pipelines with two alus and a fpu XD
it was just simpler back when engineers decided the marketing and not the marketing people.
@
Yep. "CUDA cores" definitely don't fetch & decode. Yet another marketing gimmick haha.
you mean 3, cause at "24 phases"...god damn
"They've all been lying about them, for YEARS. Nobody cared until I came along, and actually, I'm still convinced that the only person who actually cares is me!" I had a good laugh with you over that.
I love the rambling, as that's where the depth comes from.
It's a shame that I haven't been availably online during another one of your live streams, I caught one maybe a year ago, and somehow the conversation ended up going into places we've lived, where you studied, the origin of your accent(even though I can hardly tell there's an accent). It was good fun.
I feel totally schooled! lol nice job and explanation, I remember the early VRM's that came out for the 3.3v versions of '486 which appeared in their twilight, back then they were a plug in module that went into a slot next to the cpu, they were required if not running a 5v part. thanks again I learned a lot .subbed!
When buildzoid says short, I expect at least an hour
@20:00 "There is not much more for me to say" . Total length 29min 44sec. But we love it every minute of it
low rds is low resistance between the drain and source.
Thanks!
At that time it was a phase count race between manufacturers. The marketing was doing it's job because bigger number obviously "is better than lower". Check the X58A-UD9 the "top" from that era with "Smart 24 power phase design with mutual back-up to each 12 phase"
Please go to one of these companies and make your own mobo with them. Imagine the quality of it 😍
Imagine the cost lol
Kingpin card exists
Imagine the bankruptcy
I have this exact motherboard in my daily rig, paired with a 6C/12T Xeon clocked at 4GHz & 24GB of triple channel RAM.
Happy to loan it to you if you can pay for shipping from Australia!
Do you want it fried? Because that's how you get it fried! :P
Been doing pc since 2000
"Comments are saying.. x brand is bad because they're lying about phase counds"
[uploads videos pointing those who lie about phase counts or components]
Its not the brands, they all do the same things, they're in the same boat.
its also not a new problem, but I dont remember it being so common in the mid range (£170 today is £110 then)
Back in the day high end motherboards would say 8+2.. when its 4+1.
But if you bought a cheaper board you kinda knew it would be a 4+1 because there's no way the gigabyte 970a-ud3 has the 8+2 phase it claims at £90.
Now people see a huge heatsink and it could be a triple teamed 4+1, or double teamed 6+1 or doubled 6+1 or an xdpe132g5c....
The higher phase controllers and doublers and teaming leaves allot more room for (mis)interpretation.
Thats why ahoc and HUB is so important for people to learn, Sunlight is the best disinfectant so videos saying what it is and what it isn't, what matters and what doesnt are invaluable.
Going parallel for more current handling is fine, the problem is the marketing.
I'm wondering how long it is before we see "16+2 stage *BFVRM*" with 4c06n/4c010n parts using a RT8894 or something (turns to look at asrock) atleast it would have surface area.
Fun trivia: msi 970 gaming and krait am3+ were both 3+1 with "teamed" niko mosfets, the boards literally cried with tears running over the audio section.
I have a Question!
How come some traces on the circuit board do a bunch of S turns and squiggles instead of going in the shortest possible path? hope that makes sense
legth matching for signals that get sent in parallel.
@@ActuallyHardcoreOverclocking thanks
"Have a 16 phase using garbage MOSFETS from AliExpress" sounds like the AsRock strategy for VRM.
Especially on their entry level B560 boards... holy shit Steve from Hardware Unboxed did not like those....
Gigabytes marketing was always kinda weird with the VRM. I remember the "Quad Triple Phase" bs from their X38/X48 Boards.
Hello there. thanks for clearing and bringing this topic up. Its been a long while, rare.
I think the reason for that came from that the media (computer magazine, online reviewer) always educate the DIY fans how to count the number of „phases“ on the motherboard by counting the number of solid capacitors around the CPU socket. Which is not always true but however historically always useful because no one used more than one solid capacitor for each phase. However the manufacturer got that in mind and pushing the number of solid capacitors to the limit, so they can confuse buyers and be fancy on the advertisement. Either the marketing department is doing this willingly or they were as uneducated as the casual audience as well.
I remember seeing an interview with an Intel employee several years ago who said 4-phases was enough and double digit VRM setups weren’t needed unless they were using poor quality components. I guess this is the answer to why.
Truth in advertising is something we should all demand.
Buildzoid should make a tutorial on what the benefits are of such many phase power delivery in motherboards. would be great to watch
The correct way to advertise imo is: how many amps the board can delivery while run non-stop without shortening the lifespan of the components, throttling or critically failing (i.e. by not overheating). If they wanted to go above and beyond, they'd publish an entire curve profile of how it behaves temperature-wise as the amperage increases
ferrite core is a cost and size effective way to build an inductor with the downside of having more distortion.
the other way would be air core, a normal thing to see in 90s and early 2000s boards.
yes. but also you cannot get anywhere close the inductance on any comparable size. theres a factor of 1000-10000 between air and ferrite inductance
@@BrosBrothersLP I know. not my point.
@@emperorSbraz what is your point?
@@BrosBrothersLP The video acts as if nothing has ever been used before. PC building did not burst out full formed in 2010 or whenever. It is painful obvious at the start of the rant and at that point about chokes the author functions as if it did. That is fine, but for some of us, 2008 isn't the long long ago.
I remember reading about that fake power phase count back in 2010. I don't fully understand it back then and still don't fully understand it even now.
I like this video. It proves my point done ~20 years ago - at each CPU generation the corresponding motherboards are just cheap spare elements except 2 or 3 models in high-end/workstation class.
Yes, they have. Maddening, in an 8=16 kind of a way, isn't it?
Mine also has trouble with its RAM slots, wonder if it is a particular desease of the board ?
Also she (yeah it is alive) refuse to boot for days some time until the deamon goes away.
the CPU connection pins are still bent, the knowledge of this person is simply underground for what he is presenting here. He doesn't even know that there are double step-down two-phase buck converters and so the specification of 12 phase count is probably correct.
Yes you are right, forgot his bent pins.
May he have missed them then, I think he has enougth knowledge to cover that (think it is its field). I don't.
But even with doblers wich where used a lot back in these days, a true 12 independant phase wich lead to a better ajustement of the load.
Maybe it is more relevent now with all CPU states ans boost.
@@mini-pouce i'm a trained microelectronics technician and nothing of what he tells in the video comes from a corresponding education. Not only are there double step-down two-phase buck converters (in one housing and visually indistinguishable from single step-down single phase), today there are even triple step-down three-phase buck converters in use and the next 4-phase design is under development (Die stacking comes from this industry, by the way). He has no idea which firmware is running in the Buck converter controller or how a single phase buck converter can switch two phases without a time difference at the input of the CPU. VRMs are not so expensive components for nothing.
The mainboard manufacturers have their "secrets" and only trained microelectronics are able to really understand the circuitry of a mainboard. There are a number of physical tricks that have been used on mainboards since the 90s to create the same functionality with fewer components. Without weeks of measurements and precise knowledge of the tracelength, it is not possible to make statements on highly integrated technology as he makes it about this board, after all it is a 7-layer mainboard and he has no idea how the traces are laid within the PCB, he sees only the external power connections and this leads to pure nonsense from an electrical engineering point of view. he can neither show nor name the name on the converters, so it is not possible to understand what kind of component it is. I do not believe that a manufacturer made false information about his product in 2008, but simply that it is a two-phase buck converter and the guy has no idea about it. Here in europe, incorrectly specifying product properties would lead to gigantic lawsuits and in the 90s there were some through which hardware manufacturers dug their own grave with a lie. The fact is, I have a mainboard here where 12 phases are mentioned in the manual and there are only 4 converters on the board and the reason is that there are three-phase components.
@@mini-pouce if you want to learn something about phase regulators and how it is possible to put more than one in one housing look at this video.(there are still 8 possible in one housing but not used by computer electronics because there are other problems with more than 3 or 4 at the moment) th-cam.com/video/dqv2UoSEmCY/w-d-xo.html
also the maximum current you can put at the output pin at one housing is limited to 40 amps (breakdown current, some use 44 amps but this converters/regulators are the ones that dying on Graphiscards at overload) that's why you have more than one housing and some more DCR Sense circuits behind the converters.
@@darkbreed Will watch thanks.
Big topic but as you said amazing physic trics.
I had a Gigabyte GA-X58-UD7 rev 1.0 board and that was advertised as 24+2+2 phase design but i think that was only 12 phase.
I dont have that board anymore so cant check or send pictures.
and that is why, people like you are so important. To educate consumers about these things.
It now only needs to get out of this bubble by some bigger YTer and maybe the Manufacturers will change the marketing.
Hello there, since you brought up some older stuff, I think taking a look at some Gigabyte GA-8KNXP would be interesting. Maybe the Ultra-64 would be the best one among them.
Can i ask one more thing. I have no knwldge abt electronics. But, is the inductor placed near the cpu power socket important ? For power delivery ?
I'm seeing lower end boards having those but comes with "poor"(to my uneducated eye) power phase, while some other boards without inductors near the cpu 12v powersocket but, have some beefy power phase. Any advise ? You can check recent Intel H510 boards(Asus, Msi, Gigabytes).
I'm watching this video on a GA-EX58-UD5.
It looks nearly identical to your GA-EX58-Extreme, except for the heatsink on the EX58 chip.
whats a "solid capacitor"? It looks like a smt electrolytic.
A simplified expression to indicate a capacitor with a solid conductive polymer electrolyte.
Also, saying "It looks like a smt electrolytic" is kind of broad. For your information:
- smt just indicates the mount type, ergo smt/smd = surface mount,
- all capacitors have some sort of electrolyte.
@ Actually Hardcore Overclocking
Phases ... this means that there is a phase-shift during the switching for each MosFET-Gate in the high-side-driver and also in the lowside-driver (this MosFET is the replacement of the diode).
What benefits do we get from this?
The input current is more even. If all Switching-Regulators would pull current to the same time, then the input-capacitors would have a lot to do. This phase-shift makes it possible to reduce the numbers of the input-capacitors and the output-capacitors.
There is a lower ripple-current and this is good. 6 phases is very good.
In reality people don't care, this things have to work.
What people want to know are the Japan-Capacitors, because this Chinese motherboards had the problem in the past, that the capacitors are blowing up. The Motherboard-Producer did buy very often cheap Elkos from unknown source or crate a cheap design.
Normally you could decrease the stress on the electrolytic capacitors (heat production because of the internal resistance) if you add some ceramic capacitors, but they did not and save the money.
Love piggybacking your knowledge BZ, thank you for your work
One additional point is, just like when you put two resistors in parallel, putting two MOSFETs in parallel also lowers the resistance. So even though its a 6-phase, its actually appears to be a pretty good 6-phase.
-Matt
I've got this board from back in the day. I didn't know enough about VRMs to know they lied, but I remember there were many "features" of this board that seemed a bit over the top. Still, it was a good board. Still works.
I watch you and gamers nexus BECAUSE, you tell the truth about companys. And their mistakes/bad products, with no hands/fingers in betwean.
So yes, it is a good idea to tell someone they did something wrong, cos if no one does they will never learn to be honorst and do the rigth thing with the build and markedting of their products.
Being Gigabyte reminded me of this
Motherboard with a weird power design
GA-8N-SLI Royal ( has a separate add on card for power / and a PCB to switch between SLI / Crossfire )
I had one but RMAed because power card went pop in a month or so
Oh man the whole "six gear power shifting" stuff was absolute bs
nonetheless fun times
I feel like the X58 boards were completely insane and wild with the wacky colours from each manufacturer lol
Can you do the same with GA890FXA-UD5 rev 2.0? I am running FX 8350@4.8 with beta bios and works fine since i have started to use FX cpu. Max, stable oc is 5.0 for my cpu on watercooling. What we use is an unofficial bios from gigabyte worker. Funny thing is, officially this board doesn't support FX cpu's. :D
Gonna need one of those GA-Z77X-UP7 boards with 32+3+2 phase design!
outside of msi or ecs, in the late 2000s enthusiast motherboards were typically overbuilt, but dumb for the cpu's at the time.
The problem is they became too efficient with their designs.
I do miss real heat sink fins on old boards
"There's not much more for me to say."
Video goes on for another 10 minutes :D
first
you too!
I'm early!
To late
trying to be the last XD
He is John Oliver of pc industry, exposing all kinda bs on hardwares for years, love you man
I've been building computers as needed (and mostly as a hobby, but also for my business) since the early 1990s. I don't remember ever hearing anything about VRMs or phase counts or anything of the sort for my first couple decades of doing so. It wasn't until I looked into building a gaming PC a few years ago that I noticed this new terminology. I had to research what it even was because it was entirely foreign to me. I'm not an electrician. I'm not a physicist. It was like they decided to overwhelm us with technospeak in order to sound better than the competition. I guess that sells.
Motherboard manufacturers started marketing phase counts because more numbers more bettar. Especially for Intel customers. Used to be "more MHz more bettar". Then when Pentium 4 architecture hit its temperature ceiling, it became "more FSB MHz more bettar". And so forth.
Had a Gigabyte 1366 motherboard, and ooh god it sucked. Maybe I was just unlucky but RAM support was a maybe, sata controllers was more like erasers, and as soon as anything voltage or stepping related was touched it would just not post at all.
Oh, another short vid from Buildzoid!
Man, I remeber having a MSI motherboard with a nForce chipset. It was the first motherboard that I bought with my own money, it was for a Athlon XP CPU and that's all that I remember ( it was around 2001 or 2002 ). Back then I didn't even care about power delivery and how it worked, I'm not sure hardware reviews for motherboards even existed lol. It would be interesting to analyze old power delivery systems, if there were any special/insane ones :)
Hard to believe someone as into this stuff as you didn't know anything about computers in 2008 which was about 5 minutes ago.
LMAO hardly; then hed be born like 12 minutes ago
This board looks like it has the same VRM as my fully working UD3R v2.0 that i got from a friend for 100czk with an i7 960 D0 and an original big 1366 box cooler :D
@@n0rie9a I have the rev. 2.0 and it has 12 inductors around CPU. It also has 2xUSB3.0 ports on the back and 2xSATA 6G ports from a separate controller on the board+another 2xSATA 3G from another controller on the board-GBT kinda overdid the board with extra controllers besides chipset.
please great GURU of HW can you tell me if the new asus tuf b450 gaming plus 2 as vrms enought power to suport an ryzen 5600X on OC??please. i switch myne msi b450 gaming plus to the tuf an got an ryzen 2600X and pretend switch to the 5600X in september.i have an aio from artic frezeer 2 240mm and im very satisfeid with this tuf but i trust your brain opinion better then any other guru out there. please just take a look at the tuf and post someting about it. i already know the board as not the 8+2 vrms they say. is 4x2 + 2 phases. but they really look well done and i notice better performance out of the box just from switching from msi to this tuf.so im confused.
FX 5200 taught me about problem how a hot GPU die and VRAM do not work very well and how it can be easily fix with something like a Iceberg kit.
FX 9590 taught me about need for more power and the problem with hot Vrm and CPU.
well here I made my first DIY fan Bracket and Frankenstein together my first open loop cooling system from parts of AIO's and other parts I could find in your local hardware store.
More Pcb Breakdowns please.
I had a asus P5Q Deluxe and it advertised 16 phases, It doesn't matter how much it really had because it overclocked like crazy in comparison to my older P5N-D board :D
the '00s were the dark ages for the pc DIY segment (there were darker ages, but mostly talking about late '00s as fairly modern times).
almost all reviews were blurb copy-paste jobs, even the so-called hardcore/expert sites/outlets. information like chips/controllers manufacturers/models of a mobo/ram module or an actual examination of a mobos power delivery/components wasnt publicly available/easy to find. youd have to dig deep and hard to come up with even a small part/any of the above and very rarely some obscure outlet/review site would hold any info. any useful info about a mobo/ram modules was like a top CIA secret.
but us hardware guys knew about gigabytes 6 phase/asus's 8 (high-end mobos)/4 phase. OC/technical forums were full with discussions about that. ram chips info was even harder to get a hold of. the guys who knew were guys like you, but "unpublished" cause no one was doing this type of content back then
I have a gigabyte board here that has an add in card to add more phases ... it’s like a early 2000’s. Interesting board . It’s wild. Gigabyte something something ultra
I had this motherboard back in the day. It actually overclocked fairly well from memory I was able to push 4ghz out of my i7 920 with a fairly basic oc.
comparing to other brands in same era, gigayte does run really hot.
turns out most asus boards survived to today (p5q and p6) and only a few gigabyte boards survived.
God I miss X58 chipset. I still have my i7-920. I used to overclock that thing from 2.66ghz all the way up to 4.0ghz on all cores stable. That shit was wild! I had an MSI X58 Platinum SLI. I had 12GB of 1366mhz Runing in triple channel.
I had an ASUS P6T Deluxe x58 with core i7 940 and 24Gb Ram and oc at 4.1ghz. I miss this time.
My Gigabyte Z270x Gaming 9 apparently had a 22 phase VRM. That board was awesome though, the only thing I didn’t like was it was Terrible for memory OCing.
i just realized youtube put this video of a channel i didn't subscribe to in my subscription feed
I always wondered why they started advertising the phase count at all, let alone such a ridiculously high count as 12. You get diminishing returns for each one you add and it's really only so that you can use less capacitance to maintain low ripple. With a big enough capacitor you can get by with just one phase.
Well, but your average buyer doesn't know this and simply sees "bigger number = more better". So bigger numbers in marketing sells your shit and that's all the job a marketing department has to do. That's your why.
Back then it was common knowledge that anything over 6+1 was a gimmick... Might have to look in print to see it now.
the highesr RDSon mosfet i know of is the BUZ 74 with a whole 4 ohms
What is the technical or economic reason for having "phases" in the VRM at all. Why not just use components of sufficient quality and power handling ability to manage MB power? As a graduate electrical engineer from back in the dark ages, phase to me normally means power line phase in AC whereas MBs tend to operate from DC power unless I have missed something in the last few years (I probably built my 1st PC in about 1995 and swapped components since the late 1980s).
something yada yada sine curves, phasing them means more curves but they aren't 100% overlaid on each other but "pushed apart" so the "dips" in the power is removed which gives smoother power (you can look it up for more correct terminology)
I had no clue what a phase count was in 08. At least knew what a mb was.
I have GA-P67A-UD5-B3, looks like 7 phase from the other side.