Watch our power supply playlist, including exploding power supplies, over here! th-cam.com/video/SMn2F2wx1LM/w-d-xo.html Grab a Modmat as shown in this video! Large is here: store.gamersnexus.net/products/modmat-volt-large Or Medium here: store.gamersnexus.net/products/medium-modmat-gpu
ATX12VO is quite interesting. My PC is already powered on 12v lead acid batteries and I am using buck converters to step down the voltage to 5 and 3.3v with. The batteries are being charged by an MPPT solar power charger. I am also using a 12v power supply connected to unreliable grid that goes on and off unexpectedly. There are relays and filtering capacitors that works as a DC UPS. All of this would have been a lot easier if my PC is an ATX12VO.
Just to be clear, are all the spicy capacitors still going to remain within the PSU or is there going to be an increased risk of deadifying yourself by handling the MB?
I would prefer that they stay the same. Given: I don't know about actual/real data, but I'd be concerned about the 5V system dying on the MB therefore rendering the MB virtually useless. And since the MB a primary component of any system you'd have to RMA the MB or replace it at a fair cost. Whereas it currently is - you can go out and purchase a PS; or swap it out as some people probably have a PS that they could install, whereas most if not all would not have a MB they could swap out.
To the people who put the logo on the knife, it probably didn‘t even register. English as a second language, semantic saturation from seeing one‘s own company name everywhere and all that
I have a fear that this move will end up being problematic for a while. Motherboard manufacture tend to cheap out on components far more than a beefy PSU maker would. I can see a lot of RMAs in this conversion era happening.
@@prashanthb6521 in this case, since the power conversion happens closer to the load, and the input currents are smaller (due to higher voltage), there's actually less need for expensive tantalum or polymer or (for cheap boards) electrolytic caps. Given that those are basically the only common modern electronic components that need a "lifetime" spec (due to being chemistry, not solid state), having conversion on the board is actually a win. Caps are needed to supply power during the time it takes for the PSU to realise voltage is falling and to increase current to compensate. If the regulation happens on the board, the inductivity from the load to the regulator is much lower, meaning the regulator can react quicker, meaning there's less need for the capacitive buffer.
Same... Not to mention there's piles on top of piles of broken motherboards every time, same for PSU's and GPU's, but I'd rather have a 10 year warranty PSU going bad than having even more failure points on like a expensive 2 year warranty motherboard that will most likely only start to go bad after warranty ends... And then I'd rather buy another well chose good PSU to replace a bad one than having to buy a very outdated motherboard to maintain a old PC cause it takes me MANY years to be able to buy a whole new one. The PSU even if bought to fix a old PC can then be used on a new PC at any time. ps: F* this...😡
@@Validole You didn't thought far enough. Or better, you believe/expect to much from the Mobo Manufactures 😂 1. There need to be many capacitors, each sized for the specific need, instead of an big one in the PSU, usually high quality Japanese ones. (Which leads to lower efficiency) 2. There will be everywhere buck converters, cheap regulated buck converters, like MP1584 or even a cheaper texas instruments ones. Which leads to less efficiency either. 3. The mobo will run a lot hotter, because of the onboard conversation, which takes a lot more effort to cool, as in the PSU. So in conclusion, Manufacturers will surely save out and only convert all voltages on one place, not on multiple places for every component, like people expect. This is because, they will have only one place to put the converters under an heatsink. The capacitors and converters will be the cheapest crap. There only 2 benefits: - First, that you eliminate almost any voltage drop, because the converters are still nearer to the components, means less cable length etc... - Second, you will need an a lot smaller cable and even less cables coming from the PSU. But i don't see any efficiency benefits, it's just some marketing bullshit to me which will simply lead to no efficiency benefit. I think that efficiency will get worse tbh, because of just shifting the converters to the mobo. Only one converter for 12 -> 5v and another to 12 -> 3.3v (or even worse 5 -> 3.3v). So only 2 regulated buck converters to save costs, instead of multiple converters like people think here. To make it really efficient we need to get rid of 5v and 3.3v entirely, but this is impossible, because usb power is 5v and almost all IC's use 3.3v. I love the idea of smaller and less Cables from the PSU and i love the idea of the possibility of new SATA connectors, only with an single Cable directly from the Mainboard tho, but this 12vo spec won't be like people are expecting. It's all just marketing and reasons to make everything a lot more expensive. And Mainboards are already 3x more expensive as some years ago. Cheers
@@RamaOlama Except motherboard is already doing a lot of conversion (300W+ cpu, ram, chipset supplies), the relatively low power 3.3 and 5V conversions are going to do very little to change thermal situation.
Definitely agreed on PSUs retaining their price! They won't want to give up the current pricing, even if cost goes down, unless there is sudden and significant competitive pressure from a newcomer. Meanwhile, motherboards seem like they're just infinitely skyrocketing...
@@GamersNexus If this standard is pushing more power into the MB for conversion and step down doesn't that mean more heat there? So its more expensive overall. Takes up more space inside a build. Moves heat generation closer to sensitive components. For a benefit that won't benefit the end user and could be accomplished by other means.
Motherboards already have very high margins and are suffering from low sales due to the stupid prices that came with LGA1700/AM5 sockets. If they push it harder into the customer, they will buy even less. This also means less CPU sales. At some point Intel/AMD will get grumpy that Mobo AIBs are sabotaging their sales because of the stupid mobo prices.
I'm guessing that' true even though I don't understand it. If the cost of the PS is reduced, wouldn't PS makers try to gain market share by lowering their prices?
From experience, this type of set up introduces a whole load of different motherboard issues. HP has done something similar to this for YEARS with their desktop models and motherboards are more common failures than platter drives, and that quite frankly is impressive.
Nonsense. I have a Hp sff with 240w original power supply 12 years old, I also added a Quadro k620 discrete video card from a workstation. I haven't had any problems, no damage. I only clean the dust thermal paste I have not changed and it works perfectly with the fourth generation operating system
@@stonehardy580 You have to be the exception... I've seen this across multiple models. I'll bet a lot of the problems is the mobo manufacturer, last I checked it was Foxconn. Introducing more parts into a single point of failure inherently is going to cause more problems.
@@titan1983x The exact motherboard model is hp 1850. My model is hp 6305 sff with amd fm2 socket. The only fan is Foxconn, I replaced it with Artic 92pwm pst, not because it doesn't work, but because it's very noisy. The CPU heatsink is passive without a fan. I haven't changed the power supply fan - it's the original one, I just clean the dust periodically. Model is from 2012 or 13 I think, 100w processor a10 5800b and discrete video card from quadro k620 workstation work perfectly with 240w original hp power supply. I have had no problems, no defects or repairs so far.
My concern with the 12VO standard is that by shifting more things onto the mobo there's more points of failure on a more expensive piece of hardware. Right now if something goes wrong in power delivery then the PSU can be changed. If the power delivery system breaks in a 12VO system you need a new mobo and all the headache and cost associated with that.
That makes complete sense and saw another user post similarly! Agreed on that. I guess the question is whether or not those parts increase failure potential in a meaningful way.
@@cyko5950 Cost of a new power converter add in card-50$. Cost of a new motherboard-600$. I can tell you which one I would choose, and which one the motherboard makers will choose.
I'm surprised Intel went the 12v route for power efficiency savings, rather than extending the existing ATX standard with smarts (via sense/data pins like the NVIDIA 12-pin) to shut off rails when not in use. 12vo seems squarely aimed at OEM usage, rather than DIY. I can definitely see these standards existing side-by-side, potentially leading to fragmentation.
It's already fragmented and has long been so with OEMs having their own proprietary power delivery schemes. At least this might standardize things on the OEM side. As it is not only do you need the manufacturer's replacement or an adapter cable/voltage converter when attempting to replace an OEM power supply with an off-the-shelf part, you might need a model-specific version. But, yeah, this seems a net negative for DIY.
12VO being more energy efficient is good. But I don't like the notion of it being on the motherboard too. Board prices will go up, PSUs will NOT go down and now there is one extra reason for your motherboard to die, like they aren't the most common part to fail already.
The 5V and 3.3V rails are low current utilization now. There's already a VRM for memory (and on each DDR5 stick) , CPU; any inductor you see on a motherboard is part of a VRM. Most new PSUs convert the 5 and 3.3 from the main 12V rail (DC DC converter) anyway. I don't know why people freak out about adding two new voltage converters to the mobo. Servers and maybe cars are going to 48V, voltage conversion is very easy and compact now.
Speaking of power supplies, I think I speak for a lot of viewers when I beg of you for more power supply reviews please!! Especially the lower end stuff! Aris is one person that can only review so many units, leaving a bunch of the budget segment a complete mystery!
@@GamersNexus Please post more on noise/spikes/surges tests.....power quality essentially. How well dirty mains power gets filtered should be know before we buy a PSU.
Yes it would but the cost of power converters that can work on 48V input probably start to cost heaps more. The capacitors and indicators and MOSFETs all go up in spec and cost as well on anything that is 48V. Not sure if this also translates to increased track clearances on the PCB. Don't go putting 48V into lower voltage pins. I've seen someone plug their LAN into an old 48V phone system. The smoke and charcoal was amazing. So was the user's expectation that it would continue working after they let the smoke out.
@@vk3fbab perhaps that old 48V phone system also had an AC ringing voltage? In any case, 48v is everywhere in the telecoms world and it's great. Also for a given load in watts, your capacitors and mosfets might need higher voltage, but lower current, so capacitors can have less microfarads to do the same job. ST electronics call sub-30v mosfets "low-voltage" and the 30-350v mostfets are called "medium-voltage", so no one would need any special components.
PoE+ and newer use 48-56V to power devices over ethernet cables. USB PD3.1 goes up to 48V. I think the limit before major protective changes are required is when you’re needing 60V and above.
My problem with putting more and more on the Motherboard is, that you get more and more possible points of failure. And at the Motherboard prices often easily pass a few hundret € it gets expensive if anything on the Motherboard dies.
Agreed. Overloading the motherboard with extra power, is asking for trouble. Motherboard will need significant heat cooling methods, even so, motherboards is the most sensitive component. Its better to leave the Power sockets and power stuff on the PSU as its own standalone handling.
Also, there are plenty of quality 24-pin PSUs for reasonable prices on the new and used market. Hard to convince consumers to choose 12VO when enconomies of scale favor ATX 24 pin PSUs. Motherboards are a different situation, since you are more likely to be swapping it out. It may be better for DIY consumers to save on motherboard costs if they already have a 24 pin PSU. That being said, OEMs have pushing their own proprietary versions of 12VO before the standard was created to save on costs to hit power efficiency targets. However, OEMs largely haven't moved to ATX 12VO, and having a standard for all OEMs to use will improve consumer choice and reduce e-waste IMHO. OEMs also need to stop using proprietary board form factors. If they somehow save costs by doing something like soldering front USB connectors on the board, make it a ratified standard! Otherwise, they are just making more e-waste!
@@jay_ta708 motherboards won't need more cooling, you're running at most like 50 watts through the 5v and 3v3, and vrm are 95%+ efficient, so you have to dissipate like 2.5W in like 4 components at worse. dc-dc is also very reliable, it won't be a major increase in failure rate. will motherboard manufacturers add 10-50 bucks to the mobo for a $1 increase in cost? most likely, but that's capitalism for you.
A few extra things to note: - Backwards compatibility to older motherboards is going to be a nightmare. Having to convert down to the other voltages. - Nowadays, a good PSU can last over 10 years, but people usually change the rest of the components earlier than that, keeping their current PSU. That means that you are going to end up spending more money due to the incresed cost of the motherboards and the more regular rate of swaps.
I don't see why it would be a nightmare, would be trivial to design a 24pin adapter that takes 10pin + PCI-e or 8pin power and uses step down regulators to make 5v, 3.3v, I'm not 100% how the -12v pin would work but is that even used anymore?
@@KriLL325783 even a negative voltage can be generated. No problem. Every current PSU does this already. They are already 12V only designs for years now, except the dedicated 5V stand-by supply.
@@ronakparikh Watch the video and actually pay attention, and maybe you'll figure it out. But if you're too lazy for that, the short of it is: If it got completely adopted in just a single city, it could remove not just kilowatts, but MEGAWATTS off of the total power consumption. Linus tested these ages ago, and found a solid 30 watt reduction in idle power use.
@@ronakparikh not sure where you live but I live in a city of 4.5M people, the regional centres are larger than your 100k number. It adds up very quickly, also Steve mentioned it more than once but it went over your head.
Backwards compatibility won't matter, you'll just use an older PSU. It's not like a change is made and you can't by old stuff any more, this will take years to get into all home PC's. If they can get it into all government and big businesses in the next 3-5 years that would be great, most of those things are left on so IT departments can run upgrades after hours, so if they aren't using as much standby power it's a big win for the company/grid/planet. 😁
You mean like the front panel connectors? It is definitely one of the few things that has remained the same for over a decade. Still frustrating to plug them in when there's poor light or space!
@@GamersNexus yeah that was it.....was thinking at this point would have like one snap in connecters that links all the front panel ports and switches, I know somethings have slightly changed like the block u can connect the cables to and then connect the block to motherboard
@@GamersNexus Don't most cases have all of the front panel connectors together now? Rather than having a bunch of 1 and 2 pin connectors, my case all of them are put together into a single connector for all of them together. It's a lot easier to plug that in.
I really appreciate these deeper dive videos. Like I had heard about the 12VO stuff a little bit, but haven't really heard anything recently. So it was nice to get the run down on it in more detail.
I feel 12VO would be most useful as an OEM standard. From what I've gathered, the biggest excuse for oems using proprietary psu's is exactly because of the lower power draw. Lenovo has a 14 pin connector they started using in like 2014 for this reason. It would be nice for oems to use the 12vo standard so you could actually buy an off the shelf replacement for it.
They do it for a reason. Big companies like Dell, HP and Lenovo now, IBM, Fujitsu and Compaq back then, use proprietary stuff is because they sell to other companies. Those companies don't necessarily care about upfront cost as much as speed of servicing. So, of course, they made it so the only people who could service a unit would be that company and that only. No one could undercut them then.
I do like the efficiency aspect of 12VO, but I was also hoping that mATX/ ITX boards would benefit from reduced cabling rather than more expensive and bigger boards.
A lot of the voltage conversion stuff can be placed on the back side of an ITX board. Those are not gonna be as cheap as they used to be because they're no longer a budget option. SFF is seen as an enthusiast space alas. But I do think 12VO will be a net positive for ITX because that 24-pin takes A LOT of real estate and I think most could probably do without that extra 6-pin connector too.
You don't actually need +3.3V. You can remove the +3.3V pins from the cables of your regular PSU, and the PC will probably work just fine. 10:52 The additional pins on the PSU side are sense pins, hence the doubled cables on the motherboard side. As for +5V, it doesn't need to be created on the mobo, you can use an external buck converter, that you can stash somewhere else inside the case. Corsair already makes one ("+5V Load Balancer"). The big advantage of ATX12VO is that the PSU is much smaller (SFF, anyone?), more efficient, and quieter. Such PSUs will also provide much better performance at equivalent price, since there's only one rail that needs to perform well. Switching to ATX12VO would clear the market from all the garbage group-regulated PSUs.
Finally a sane comment here. Here in India I've only ever seen the garbage group-regulated ones. Pull too many amps on one line and the other goes out of regulation and potentially kills the mobo. I can't wait for 12VO to become mainstream.
I've never been a fan of 12VO, conceptually, Motherboard longevity (or rather, the lack thereof) is one of the more frustrating things about the PC industry. CPU's outlast their motherboards significantly. Really, you can't expect a component with that many separate parts to be as robust as a CPU which has been binned for its unitary integrity already. many CPU's in history last so long that they outnumber their host computers, then having to be scrapped purely because there are no matching components to make them useful. It's rather frustrating, then, to see that standards are moving towards increasing motherboard complexity and responsibility, when, really, I'd much rather see more modularity. I think the E-waste saving potential of having a standard with more numerous, smaller, and less responsible components, is vastly more important than the power saving potential from voltage conversion efficiencies.
Well said. I have multiple still good & working CPUs (Intel Core 4790K, Intel Core2Quad Q9650) where the motherboard died and I couldn't find a replacement anymore because they were only manufacturing/selling newer socket generations.
What kind of modularity do you mean by (presumably) separating motherboard components and making them replaceable? It sounds like being able to DIY build your own motherboard, if I'm getting it right. I personally would not like to see increased motherboard costs, but ideally if that weren't to happen and there could be other efficiency improvements by combining multiple power connectors and re-working expansion slots (and combining SATA data and power cables especially) I would quite welcome it.
@@jamesbuckwas6575 Ideally, I'd wish for the ATX standard to evolve into a less monolithic design. DIY motherboard is one way to think about it. I think, Ideally, the motherboard itself would be mostly the chipset and CPU socket. The goal would be mostly to reduce the size of the board significantly, by removing extraneous features. The rest of the components, such as power management, I/O, sound, USB, could be put onto add-on boards, which either socket onto the smaller motherboard, or cable into it. What that would look like, or how it would affect latency, I have very little idea. Though I do know it is antithetical to the current momentum of motherboards, which are now beginning to expect you to install the storage on the board itself. Perhaps instead of that, there could be a dedicated PCIe x8 slot that connects to storage board, which can be installed behind the motherboard, perhaps with the memory controller on the board itself? Who knows. Perhaps the ultimately solution is to replace most I/O with PCIe. Making CPU's more variable. Return to soundcards, and the likes. Perhaps with separate PCIe-slot boards, linked to the main motherboard with a special, denser riser.
The worst part is all these e-waste generating measures are sold as "green". Many accept these claims at face value, and the community fights itself over the often non-existent environmental merits. In this case, you'd need to save an awful lot of watt-hours to make up for disposing of motherboard sooner and replacing it with a newly manufactured model, and we're talking about a watt or two on standby with these changes. It's possible that the breakeven point is hundreds of years with something like this.
I think the endgame would be something like removing the 5v and 3.3v rails completely, basically giving components like 2.5/3.5in drives a 12v input which they convert down internally. The current layout with 12v to the motherboard and lower voltage to those off-board components means the voltage converters on the motherboard still need to be overprovisioned since they need to be capable of providing the maximum power when all the components are connected, when in reality it's likely that only a handful of them are hooked up. 12VO can be huge for SFF builds, since the PSU can be made far smaller due to simplifying the DC-DC circuitry and cutting out 3 rails of filtering.
I feel like that last thing we need is more things the motherboard manufactures can use to "differentiate" their product lines. I run 8 SATA drives, would I need to get some sort of special add on card to support that many?
you could use a 6 or 8 pin add in card. 12V to 5V is very easy to do and there's already supply chains for it. you (or a manufacturer) just have to add a 6 pin connector on one side and sata connectors on the other. there's probably gonna be single active cables that do 12Vo sata to 12V+5V sata too.
@@satibel Add in card? Like PCIe SATA expansion cards that will take power off the slot and give you 5v? If that is the though a 1x card it would be limited in power output.
It's a good thing that framework is starting to offer their products in standalone packages. If current motherboard pricing gets too far out of control then for my next build I will probably just get a framework mainboard and put it in the cooler master enclosure and use an external GPU. Also especially since SFF looks like it will get screwed by this change.
My concern is real estate on the motherboard, especially on ITX cases where space is already at a premium. Then again, on a very compact ITX build, does a 5V line coming from the motherboard even make sense, if it eschews even the SATA data ports, for m.2 NVME only storage. Otherwise, 12VO would definitely make life easier not having to manage a 24-pin cable, though that only matters 1-2x a year when I would be doing anything with a PC internally.
Yeah, looking at this I immediately jump to builds where every component is 12V, right? If the standard started to gain traction you can imagine a scenario where some of the other components start shipping with 12V out of the gate and unlocking more builds where you just outright don't have the other rails at all. Or maybe some of the remaining components do the conversion themselves, like a breakout LED box with the 12v to 5v in the hub. Presumably you eventually have a simpler, tighter build with way fewer unpopulated modular outputs and a smaller PSU. But I feel that's like when USB was supposed to be "universal". The complexity is still there for the legacy stuff and just gets spread out across the standard eventually.
+1 for itx! Love my small, custom SFF cases and with how much space is a premium on ITX boards I fear what they will have to remove to handle voltage step down on the mobo itself.
@@mademedothis424 USB C basically did away with the simplicity that USB added. I see lots of reddit posts where novices are confused with what their USB C port on their laptops are capable of doing, particularly DP-alt mode.
Definitely very problematic on ITX motherboards. If you want to do conversion from 12V to 5V and 3.3V on the motherboard, you need to allocate space for capacitors, high side and low side mosfets, inductors for each rail of 5V and 3.3V. Then, you need to put SATA power headers. ITX boards are already so cramped as it is now, I don't see how they are going to do it unless sacrifice DIMM slots for SODIMM .
@@fleurdewin7958 Don't bother with the sata headers, provide a breakout board with sata connectors. Between that and going away from the 24pin, I don't think it's impossible and the reduced complexity on the PSU could make for extra tiny GaN units.
18:58 One interesting thing about that Motherboard is that it has the mounting holes for the VRM heatsinks. This means that they planed to put heatsinks on that board or they intended to release a higher price variation with heatsinks. The fact that it has the mounting holes also makes adding custom heatsinks much easier. Its rarely done by moders but the adding of heatsinks to cheap Motherboards that doesn't have it can increase its values by something between 20$~80$. The biggest difference between low end and mid end motherboards many times boils down to the addition of the VRM heatsinks. The mod is relatively easy to do. A hand saw, a drill, old scrap heatsink, some thermal pads, some small springs and screws with nuts is all you need.
Gotta chime in and say that 12v only (sometimes even 24v) PSUs have been used in data centers in the US for several years now. I have been working on them for at least 3 years now. They are more stable in every sense than traditional PSUs. Power distribution boards then do all the "magic." This has also been great for cable management and thermals. I would love to see this adopted more mainstream so it becomes more available, aka more reasonable prices. Good day ya'll.
I do some 12V-intensive things at home and my go-to for those tasks are used 1U server power supplies. 12V 50A+ single output and they're generally free. Some I have are more than 15 years old and still working just fine.
ATX12VO is a standard that is there to make the SI look good because they can print a good PSU efficiency spec on the case, but don't have to worry about the losses "downstream" because that's on the motherboard manufacturer, and not relevant to the spec.
Two important points to 12V only is: 1. Easy battery backup: Just connect at 12 V battery. No 110V sine inverter and other lossy and expensive system. 2. Reduced cost: The motherboard only converts the voltages it needs, and as much as it needs. Usually -12V and 5V is dropped. This simplifies the system overall.
I don't know how significant it is in the case of computer power supplies, but transmitting a given amount power through a wire at a higher voltage is also more efficient because the lower current leads to less resistance loss in the wire. So by moving the point of conversion to 3.3v and 5v closer to the components, there is less overall loss due to resistance in wires when those components are under load.
while true, this is only insteresting for power delivery infrastructure via high voltage lines that are in the ballpark of 380,000V (germany) that can be up to 450km long for 230V households. 3,3V, 5V and 12V isnt much of a difference and that not even accounting for the short cable lenght of half a meter.
@@fmskaisergsaw5567 ohms law is useful everywhere, not just high voltage lines. The telecoms world standardised on 48V supplies decades ago and for most cars they moved from 6V batteries to 12V in the 1950s. Most diesel trucks use 24V batteries rather than 12V, at least for starting the engine. In many data centers in the US, they distribute power at 240V, 480V or higher, rather than the domestic 110/120V that is common for home appliances. Whenever you have high power loads it is more efficient to do your power distribution at a higher voltage.
as an SFF PC enthusiast, I really can't wait for atx12vo. Cabling is way simpler, and without the 5v and 3.3v circuits in the power supply, manufacturers can fit more power into smaller PSUs. I personally am curious about how complicated Z#90 mITX motherboards will become, but considering manufactures like ASUS is taking inspiration from lego with how many stacked carrier boards they use, I believe it can figure it out.
i like the power supply handling everything instead of a motherboard, since you might need more connectors down the road or hardware configs might change and need different plugs so you could be screwed since the motherboard has so many, and can physically can fit so much already
And that's why you want to handle power conversion on the board. One pair of 12V can carry more power than two pairs of 5V or three pairs of 3.3 V, and the DCDC is much smaller than the additional space taken up by the wider plug.
@10:45 That variable resistor is likely a trim resistor for doing fine-adjustment of the output voltage, you can easily validate by measuring the output voltage while adjusting it. It's fairly common to find on open-terminal power supplies for industrial purposes where you just screw some wires on to use it. The primary use of a trim resistor in these kind of power supplies is so you can compensate for voltage drops during high power draw (depending on connected equipment) by essentially adding a voltage offset to ensure it's always above the required threshold regardless of load.
It's tough enough picking a good motherboard that has the features I want, at a price point I can stand, and looking good (not a point for some but definitely a point for me) based on how things are already. I don't want to worry about even more variation being added to motherboards. I'm half asleep so may not have fully heard and understood the reasoning, but if the reason is for lower power draw in standby or low power mode, couldn't something similar be accomplished with a data connection and a little bit added to the PSU? Or is it simply just that much more efficient still to stick it to the motherboard instead? I wouldn't mind a changing standard if it's not making as drastic a change to motherboards. Would be interesting to have a sit down with an engineer that worked on the newer spec to talk about it and why that was what they went with rather than other options. Although maybe there's already good info out there for why not that I just need to actually read/watch.
I think that main reason is, that ATX is ancient. It was developed in times when switching DC-DC power sources weren't as cheap and it was more efficient to make -12V 5V and 3.3V on a smaller transformer in the PSU. Now the PSU uses DC-DC converters anyway and -12V is not used, it makes sense to reduce wiring to only 12V. The converters can be optimised for onboard devices and managed by MB's logic.
@@Papinak2 A better solution I would still think is segregating responsibility of tasks. A power supply should do what it says it does. A motherboard should only serve as a board for other items. The standard could have simply asked for some pmbus standard with specific addresses and communication between motherboard and PSU. Although complex to implement, I would expect a lot less issues and less components means lower cost and higher reliability. This seems like a "standard is outdated we will simply do everything because working to help change the other parts of the industry we don't control is hard".
@@epeglab1o1 This is impossible. CPU's and such use very low voltages these days, so that buck converters ("VRM's") have been necessary on every mobo for decades now. It will be a nightmare to even imagine reliably supplying dozens of amps at 1.2V over any length of wire. And totally inefficient.
@rexsceleratorum1632 okay you took my comment too literally. Yes the mobo does have multiphase buck and other bucks to provide various voltage rails for the cpu.
I'm not in favor of shifting the cost to motherboards. I've invested into a 12 year warranty evga model, and with intel switching motherboards every two years, and even amd probably not having am5 lifespan as long as am4 - I'd rather buy psu once, and then use it for a long time, rather than paying extra each time I change platforms.
@@dakoderii4221 even if it adds just an extra 20-30$ it's still going to be recurring cost. There's no benefit to changing your psu every time you rebuild your platform, and I'd like to stay that way. We're getting here into dangerous territory of pricing people out of pc gaming.
7:25 After years of cutting/trimming stuff on electronic assemblies, I found the best cutting tool is just an actual scalpel blade. I use Swann Morton non-sterile carbon steel blades, they come in different shapes but mostly #10 for slicing and #11 for scoring
Another advantage of ATX12VO is better dimensioning of the +5V and +3.3V rails as regular ATX psus have grossly overdimensioned rails to account for the possibility of them ever being connected to an ancient computer that may pull power from them. The board manufacturer will have a good idea of how much power could be pulled from those so it should end up being cheaper. The negative rails could also be omited, most modern computers don't use them at all. The +24% cost to matx boards and needing extra layers sounds like bs to me (I have designed 4 layer pcbs for fpgas and microcontrollers). You could have the dc-dc converters required near the connector, if the power is low enough the area freed up by the smaller connector may be enough for the converters and that way you just route the output much like you always needed to. I only foresee an increase in cost if the board packs a crazy amount of IO like usb or sata which could potentially draw a lot from +5V and +3.3V needing beefy multiphase DC-DC converters in an already cramped pcb but that shouldn't be common.
That just makes DIY PCs expensive for no reason, because the market for PSUs will be small and full of enthusiasts who are susceptible to advertising and willing to overpay for form-over-function nonsense like sleeved cables in cloth bags. It's better to be a small fish in a big pond. Multi-rail ATX belongs in a museum, not your computer.
Would be good to see a bit more on this for the ITX space... Seems like the PSU's themselves are smaller as-is, but if even mATX boards are hurt by moving all the voltage conversion onto the mobo (increasing size/layers/cost), that could be even worse for ITX. Seems like the best-case would ultimately be for end-components (PCI / Graphics cards, Hard Drives and SSDs, etc) to do their voltage conversion internally so the motherboards can be passive 12v as well...
The cost changes and the cable management are both absolute No-Go. The biggest problem that I see is, that PSUs very often outlive Motherboards. My current Thermaltake 650W PSU is now nearly 14 years old and will power my 3rd PC Hardware Generation now (Core2Quad Q9650, i7-7700K and now i9-13900K). So the same regulators reused all the time and still working perfectly fine. With 12VO the additional hardware on the motherboard would have to be paid and then gone to waste every time, raising the cost and amount of E-Waste. At the end we WILL pay for the extra components and MoBo prices are already insanely high with all the crap put on them. This might be beneficial overall for commercially used all in one systems but it will hurt ANY DIY build. Not even speaking of additional coil whine risk with open regs on the board compared to enclosed and dampened regs inside a PSU case.
It takes a great deal of faith to entrust your daily driver PC to a power supply from mid-2009, complete with group regulation and an efficiency level of most likely either 80Plus White (W0104) or Bronze (W0163). Yes, one would rather hope that a quality unit makes it that long with the occasional dusting, but it should rather be enjoying its well-deserved retirement in a more fitting retro build by now.
I'm glad to see GN get back on the PSU topic. I've really been hoping that you would review a few PSUs a month because the market desperately needs reviews.
This is why Steve is one of the best! I totally agree with your perspective on this and don't get the lack of basic debugging hardware indicators on newer motherboards.
There is a SATA connector (SFF-8482) that includes the power connector within the same cable and connector which really should be used for these 12VO motherboards. This would make cable management for SATA drives a whole lot easier. Then motherboard manufacturers can just provide another one or two of the 4-pin motherboard headers for users to connect up non-SATA peripherals that require a SATA or Molex power connector. Or, you know, they could just make a whole new combo connector for the motherboard side that provides both power and signaling which has a SATA data/power combo connector (SFF-8482) on the other end of the cable - they could even double up on this to provide SATA/Molex power connectors from that same header using a different cable if you don't need all of your SATA ports. Then you could have 4-8 of these ports on your motherboard and all the motherboard manufacturer needs to do is provide the cables.
Get more hard drives that need more power than the motherboard can supply, replace the motherboard (or fry it). Hard pass. 12VO PSUs need a 5V rail for drives, period. Let the MB take care of -12V, 3.3, fine. But the sole source of 5V shouldn't be the MB.
@@TiagoJoaoSilva Considering that SATA ports only support one harddrive per channel I highly doubt that you could overload the motherboard's 5V regulator if the mother board vendor actually speced the VRM properly. That said, it is the 12V rail that gets hit hard when you have a lot of harddrives as that is what the harddrives use for their motors and the peak power used is when all of the drives are spinning up. The actual electronics which use 5V don't draw that much power. What you should be more worried about is having a whole lot of aRGB LEDs running. They use around 150mA per RGB LED at full brightness and if you add in a bunch of strips then the power draw can really start to add up. This could be easily solved by having a 12v -> 5v control box for the aRGB strips that can take in the data from the motherboard and supply the 5v required via a 12v->5v buck convertor. That said, I love the fact that you think that this is a problem that is unique to 12VO. Staggered startup has been a option on ATA controllers designed to take a bunch of harddrives for a long time as spinning up a bunch of harddrives at the same time has the risk of overloading the 12v rail and triggering OCP.
I'm getting flashbacks from my project of converting an ATX power supply to work with a Lenovo OEM motherboard. For anyone interested: It's almost like a 12VO connection but not quite. You connect the 12V wires and ground directly, but then you need to hotwire the 5V standby to ground. My PSU also required a load on the 5V rail, so I put a 10ohm resistor there. It works fine now, but holy smokes what a headache.
I have a bit of a bone to pick with 12vo. I don't like moving all the voltage regulation stuff from a 150 dollar power supply to a 500 dollar mainboard. I'd rather replace the former if something goes wrong instead of pissing around pulling a much more expensive and harder to get motherboard. I could be crazy but this seems more work than is necessary. And yes, they will absolutely charge me more for both the new 12vo power supply, and absolutely will for a new 12vo motherboard. Because reasons.
If it impacts board failure rate, then that's definitely a good point. Even if it's not a $500 board but a cheaper one, it's still a lot more work to replace.
Another reason for the transition to ATX12VO is that 3.3V and 5V are difficult to efficiently deliver over long cable lengths. Those long cables act like antennas, picking up unwanted EMI. Moving all of the conversion to the motherboard is much more efficient.
Always looking forward to a more simplified connection setup and more power efficient setups. I just hope we can adapt existing power supplies to the new connections when they become more mainstream.
Very interesting to see! Considering the points about Mainboard size towards the end of the video, I'm a little left to wonder about ITX size boards with 12VO, as it would be especially interesting for small builds, having the reduced footprint of the 12VO connectors and a simplified PSU. But I'd think it would further increase cost of ITX boards, needing more layers or probably even stacked PCBs to fit everything else into their size... I love building in SFF, and 12VO seems like a step in the right direction, I'm especially thrilled to see the I_PSU% feature coming one day, I just hope pricing won't be insane and that complexity doesn't actually increase from trying to decrease it at another level...
The footprint is only smaller with the connector, Steve was ignoring all the circuits needed to get the lower voltages now taking up space on the board making the actual footprint larger even though the connector is smaller.
@@JETWTF yup, that's what I figured. It's just moving stuff from one place to another. Everything that had its place well within the PSU itself now needs to move over and require even more space on the board.... While PSU probably won't even change their form factor
@@encryptedunicorn7659 I wonder about the improved efficiency since it is just moving PSU functions over to the motherboard. Could that efficiency be achieved without moving the PSU onto the motherboard? I am not an electrical engineer but I cannot see how a few wires makes enough difference to matter unless the wire used is wrong and thats super easy to rectify.
I've been shopping for PSUs lately and have been confused by the varying pin-outs. This video perfectly explains to me WHY I'm confused (but I'm still a bit confused).
Hello Steve, because of moving RAM VRM from motherboard onto DDR5 RAM Module itself, now I am experiencing Coil Whine directly from DDR5 RAM when RAM intensive loads are in operation. That kind of coil whine was NONE-Existent on DDR4 motherboard. That actually mean DDR5 RAM makers either don't know how to build a good VRM or they use cheap components. But Motherboard manufactures knew/learned from many years of experience what to use as quality VRM for RAM. Now if motherboard makers, which do not have the experience of PSU makers, start adding part of the PSU (5V+3.3V circuit) onto the motherboard, Sure we will have more possibility of coil whine + increased probability of failure for a motherboard. The motherboard already has so many different circuits that if any of them fail, the manufacturer should take back the whole unit and change it with another (at least within warranty time), now adding another new circuity with high probability of failure (due to manufacturer's not having enough experience like a PSU maker) it will ultimately increase chances of failure - equally increase motherboard manufacturers cost for after sales services. Add the fact that more and more users are turning their PC case into a Christmas tree by overusing their 5V rail and adding multiple ARGB devices. Last but not least, the Motherboards, specially mainstream and high end classes, are already filled with so many connectors that make cable management a difficult task, now adding more cables out of motherboards, will just worsen the cable management more than ever.
I never thought fully about the office PC situation regarding 12VO. With the tiny SFF computers, sometimes smaller than ITX Windows machine's with external PSUs (the laptop style ones I'm blanking on the name) something 12VO seems even harder to implement, and that's a huge chunk of business I'm sure as I see them everywhere. One more reason Gallium Nitride/Galllium-related tech is so important and promising lol. And as far as usability of other things, I agree there will still be plenty of adapters or retrofitting supplies etc.
I think doing the voltage conversion is better off done inside the PSU. Not sure if moving the down conversion to the MB causes more noise issues but it certainly adds more heat where you don't want it.
Not necessarily though - if you can move the power to where it needs to be at a higher voltage, the current is lower. Which means less losses in the traces. Of course the conversion also has losses, they're at efficiencies of 95-98%... in the end it's probably a negligible difference though, as CPU and GPUs currently use the 12V rails and convert down already. Noise is probably better too, as the motherboard manufacturer can add filters without having long cables that pick up noise from the environment. But price is a different topic, and reliability.. you can just replace the PSU and I don't think that a PSU with 12V only will actually be cheaper..
@@marcelleuenberger4648 I don't think your going to see much difference in line loss. 98% of which gets turned into heat anyhow. An I9 alone can soak-up 300 watts when throttled. They filter power supplies as well which can house much larger components like caps and inductors for filtering. My Corsair AX1600 has a very clean output and is 94% efficient. The PSU is one of the largest sources of noise in the system anyhow. A poorly designed power supply is going to introduce noise in the system no matter how many rails it has. Hardware engineers add filtering all along the circuit even on the front-end of the chip I/O so it's usually not a problem. I'm sure both set-up's have the pluses and minuses. Should be interesting to see if this catches on.
@@donuthole7236 Most power hungry devices are powered from 12V being converted locally anyway (CPU, GPU) because it is not practical to do it any other way (450W being pushed to GPU at ~1.1V is over 400A - and numbers are similiar for the CPU (which needs few different voltages that are generated on the motherboard)). CPU, ram, chipset alone need 7 or 8 different voltages (DDR5 needs 3 different voltages (usually two of those are identical, but that is not guaranteed by the spec)). Adding 2 extra votlages to all of this is going to do very little to BoM costs, but are the manufacturers going to use this as opportunity to increase price?
@@donuthole7236 Of what use is that filtering in the PSU, when you send that current down to the MB, over cable and two connectors? Yes PSU filtering isolates the noise from the main, but does not help with noise and voltage oscillations down the line...
We've moved well beyond the times when changes in technology were made to improve said technology. Any change now is simply an opportunity to charge more. This change would somehow increase the cost of power supplies and motherboards, and likely lower the lifespan of each to boot - exactly what manufacturers want. I'll wait for your testing on these which I'm confident will show little/no increase efficiency, with significantly higher cost for both PSU and MB.
Technology already improved, which is why ATX wiring is redundant. You don't really need to convert 12V to lower voltage in the PSU, so why not move the converters to MB? It can be optimised and managed by MB logic - for example, shutting off 5V to SATA, if it's not needed, but keeping it for USB.
I have custom built NAS, based on Advantech board based on Q170 chipset and Pentium G4400 CPU. It's powered by 12V power brick and it's crazy efficient compared to my previous Atom setup.
There's something sinister about a knife with the phrase Be Quiet! on it. Instead of telling your victim to be quiet because you just pulled a knife out, they'll get the hint real quick.
I hope it never does on home PCs. It’s very suspect to have the voltage regulation in the MB, if it does break it can cost 700 bucks if the MB did. cost as much. But if something breaks, the PSU can be replaced for cheap, easily and it also has voltage protection on it so it can’t break other components. Think about that!
Technically, all the existing PSUs based on modern DC-DC design are already ATX12VO considering the fact that the PSU itself only converts mains voltage to 12V and then this 12V gets further converted into 5V & 3.3V for minor rails (all inside the PSU). Now the major problem is with the 5Vsb that eats up tiny bit of power even when on standby mode and instead of figuring out a better solution to this standby power consumption problem intel stupidly (imo) came up with a solution that seemingly disrupts motherboards as well as PSU designs.
Edit: The original was incorrect but is kept below for context. USB PD support various voltages including 12v. Original message: USB really fits weirdly with this. Just about the only voltage that the PD spec doesn't provide for is 12V.
Yes it does. It's complete bs that they don't. I've seen so many articles talking about how very few pd chargers actually have 12v. I've checked every single one in my house and car and they all have it. I checked Amazon listing specs and same thing. Gotta be careful tho because some charger support 12v only on usb a like anker chargers. Got me wondering if I should test mine to make sure they support pd on usb c since mine have usb a too
@@andrewk8636 Oops, turns out you're entirely correct and I have got it wrong. The reason I made the mistake is that I looked at a table of USB PD rev 2.0 voltages and currents that didn't list 12v, missing that the table only listed values that weren't already present in USB PD rev 1.0, which already accounted for 3 different current levels at 12 volts. I've put a correction in the original post.
@@andrewk8636 Apparently the wires matter too. To provide more than 60 watts of power your cable has to be able to tell the other devices it can handle that. Then again, that's probably more relevant to laptop charging than desktop pcs and phone chargers.
@@martineyles wires shouldn't matter at all for the voltage. Amperage is another matter. If the wire can do 9v and 20v there's no reason it can't do anything in between
@@andrewk8636 I was talking specifically about higher power specs above 60 watts in the last reply, available at 20 volts and above. Eg. 28V at 3.57A for 100 watts. That's why I stated it was more relevant to laptop charging. It's possible that it's because of the > 3A current though, as I don't see the option of 3A, 28V, 84W.
People are talking about motherboards dying but even without that, you still usually need to upgrade the motherboard more often than the PSU. I can buy a really good power supply if I only have to do it every 10 years, but I'm already keeping all of my hardware for a minimum of 5 years in the best case, so it might get even worse.
I know there are a lot of possible issues with this, like more points of failure and the injustified markup that's bound to happen, but let me dream for a second: just imagine how a build with this could look if you just used M.2-SSDs. You would just have two big cables, one going to the MoBo and one going to the GPU. In addition to that you could daisychain the casefans and now cablemanagement would be a dream. Especially for SFF it would be great to have the option, then you could choose a cleaner build for more money and otherwise you would just take a standard 24-Pin MoBo.
You'll still have EPS connectors. Also itx boards are already being stressed to their size limitations and now you want more DC-DC conversations to happen on board than before? Good job.
@@NotThatGuyJD Well I said I am dreaming a bit but I will kind of try to address your concerns. The boards wouldn't necessarily have to be bigger. For example there is more I/O than most people ever need on some itx boards. Some people would be willing to trade that for better cable management. As for the EPS-Connector: most CPUs could theoretically work on the power budget that is provided by the 10 Pin. A 13900K? Probably not, but that one consumes way too much power to begin with and if you want to use it you could just poupulate the EPS connector or use another board.
I like it. It may bring a motherboard plug that would be easyer to remove (aaaugh my hand always hurts doing that), And the motherboard could have small lvo for a handfull of connections, like a 3.3V one for each ssd, one heavy duty for usb ports, another for the rgb lighting, etc...
The way I see this is power supplies will become cheaper to manufactur... but the manufacturers won't actually drop the prices. This will only increase motherboard prices making it a bad deal for the consumer. Also this will add more points for failure on the motherboard itself.
Difference there is that the Connectors are completely undocumented in most cases, and not everything has an adapter to make it work. Dell/HP also don't just use one odd layout of connectors, they use different ones for different products. Here it's a standard, documented, and adapters exist to reliably convert from one to the other.
Dell and HP are selling their business lineup to accountants who make their purchase decision because one vendor provided the same specification 50 cents cheaper than the other.
I have an additional 6pin PCI-e power connector on my X670e motherboard. Didn't realize what it was until I looked it up and it's just like the one you point out around 20:00.
That'd be cool! Seems like it'd rapidly enter "we didn't like the standard so we invented a new standard" territory for the pre-built manufacturers, though!
@@GamersNexus Good point. I don't like the lack of backward compatibility either, although adapter cables should be feasible. Make it an open standard agreed by a consortium?
@@odjsjaks The PSU could be lower cost due to reduced components and complexity as it only needs +12v and GND, like the PSU's discussed in this video. Also cheaper due to fewer bundled cables. The motherboard would increase in cost, offset by PSU cost decrease, as it now has the 5V and other circuitry. The data cables would increase in cost because they now must also include power, which adds more complexity and material, but this should be offset by PSU price decrease. When buying a new PC this should all balance out to no change in overall cost. Adapter cables can provide converters to allow interoperability between old and new hardware. So you can keep your old PSU, motherboard, or other hardware, if you like. PSU manufacturers may make less money off this, whilst mobo make more, maybe? It's hard to tell how it would shake out. Regardless, this is a quality of life improvement, not a money making scheme.
If it's just 10 pin replacing the 24 pin I'm all up for it. I'm due for an upgrade anyway. But this one seems like not only it would increase the cost of the motherboard, because the extra component needed to power the remaining stuff, it didn't reduce cable clutter either. Look at that janky ass power cable from the mobo to sata.
I really love these indepth stuff you do did not even know that there was a new version of PSU lol. I got mine in 2017 and it's still going strong. Ty Steve
Yeah, while I understand the gain in efficiency of 12vo, I don't think integrating more features into the motherboard is a good idea. What I think would be a good compromise would something like a 12VO PSU, mostly as as an AC to DC converter and voltage transformer, connected to something similar to a DC to DC picoPSU in a breakout board which could be located near the board as to reduce losses.
So my Seasonic Syncro has a 650 W PSU that is 12vo and goes into a distro board that converts everything into normal stuff. Not quite the same but I love it as a case. (Would love to see you guys review it someday)
Moving this type of change into 1 form factor would drive this, like into ITX- dtx, in the mean time support an interface between them, simple 12vo ps feeds small adaptor for power division for use on older boards just like that Corsair cable. Nice old silly scope on top shelf left. Ipsu looks like current monitoring nice. I will be looking into this for next build as spec matures
This could be very interesting in combination with the last couple of motherboards with reverse mounted headers and stuff because I can see all of your sata connectors and a lot of your lower voltage adapted headers coming off the back of the board making cable management easier
HP does something similar in some of their small forms office PCs. We had one. We needed to add a video card that wouldn't fit. So, lol, we cut the motherboard tray out of it and a Dell optiplex case and soldered the HP tray into the Optiplex. We installed a 600w psu and found a problem because HP used a 12V standby voltage. We had to purchase a small voltage converter circuit and glue it on the side of the psu, ran the 5v sb to the converter and the new 12v sb to the HP main board. I had to build a converter cable from 24 pin to the HP connector. Everything worked except we kept getting a fan error. Turns out the HP power supply sent a tac signal for the PSU fan through a second connector to the motherboard. We were going to tap into the CPU fan tac to correct this but since we could bypass the error by pressing a key during boot we didn't bother. Frankenstein worked great for two or three years before we replaced it. I am curious to see if this new standard works out or if people just say "meh" and keep purchasing boards and PSUs using the current standard. I know I have no plans to replace my 2-year-old Seasonic. I hate moving the SATA connectors to the mobo. That's such an hp thing to do.
This will result in motherboard manufacturers cheaping out on components to create the lower voltage power rails, then troubleshooting power issues will be 2-pronged. At least with the current state we know that if there's a low voltage issue on the USB, it's a power supply (or load) issue. I like the idea that all power rails come from a power supply that's been tested, certified, etc.
Looks like your run of the mill boost converter circuit with a 12V clamp, looks like a quick and dirty fix for the voltage conversion. Likely it will be totally fine for the conversion as standby voltage is only drawing power when the board is off. That definitely beats tossing out the PSU you had before.
So not only are motherboards beginning to be kind of unaffordable but they now also get new things that could fail (power conversion parts which were previously handeled by the PSU) I'd rather replace a faulty PSU which costs ~90€ than replacing a motherboard which starts at ~120€ because some power conversion parts are broken or "smoked". I'm not sure if I like where the PC development is going regarding the changes mentioned in your video. Not to mention the effort that changing the motherboard brings along. Nevertheless: Great video, keep up the good work at GN ^^
The extra lines are like used to reduce Vdroop on the 12V rail under CPU load, this drop in voltage is caused by the inductance to the PSU under fast switching operations (what you might call transients).
Would it be possible to transition to a single cable for both SATA (data) and SATA power ? With an adapter on the device side to accept a single cable instead of two. I think this could reduce the "worst cable management" aspect of 12VO boards.
If the board will handle the power supply, wouldn't it be better to combine both power and data cables into one. Eg: for sata drives you have one cable that carries both power and data?
My main issue with 12vo is your adding more points of failure to your board. A PSU is usually a lot cheaper to replace than your motherboard although not always. Oh and we would get much more expensive boards
12v has been standard in datacentres for years now. We haven't seen any issues because the 5v, 3.3v and -12V lines do so little these days that the regulators on the motherboard that supply those voltages never fail. They're just doing so little that they don't fail. The same goes for laptops. You don't see failures in those regulators because they do next to nothing. Even in desktops by Dell, HP, Lenovo etc have been 12v for years and again, it's a non-problem. Not that any of that matters. ATX12VO is an update of the ATX12V spec. It's been around for decades and it has had zero effect on motherboard and PSU manufacturers, and there's no reason to believe that this is likely to change at any point in the near to mid future.
Taking into account the overall equipment roadmap, where most of the features are transferred from the industrial grade to DIY and than to retail products, no surprise that it's just a question of time when the common powerbus is getting its place in home applications. Most of the complex electric/electronic machines have a number of PCBs powered by a single voltage transformed by internal means into required range.
I can guarantee you 100% the price on power supplies will not drop despite the cost being moved to the board. If anything we've seen in the past 3 years of pc hardware industry scheming, this is a highly improbable future. The only thing going down in price these days is nvme storage which seems to have a mind of its own.
So I imagine that cases will need to change to handle all of the new cable management that will be needed for cables starting from the Mobo instead of the PSU.
![BOWKYLION - วิงวอน (ex-change) [Official MV]](http://i.ytimg.com/vi/cLdnZWDaO-w/mqdefault.jpg)
Watch our power supply playlist, including exploding power supplies, over here! th-cam.com/video/SMn2F2wx1LM/w-d-xo.html
Grab a Modmat as shown in this video! Large is here: store.gamersnexus.net/products/modmat-volt-large
Or Medium here: store.gamersnexus.net/products/medium-modmat-gpu
love you steve
ATX12VO is quite interesting. My PC is already powered on 12v lead acid batteries and I am using buck converters to step down the voltage to 5 and 3.3v with. The batteries are being charged by an MPPT solar power charger. I am also using a 12v power supply connected to unreliable grid that goes on and off unexpectedly. There are relays and filtering capacitors that works as a DC UPS. All of this would have been a lot easier if my PC is an ATX12VO.
Can you share the part number from that corsair rm850 12vo?
Just to be clear, are all the spicy capacitors still going to remain within the PSU or is there going to be an increased risk of deadifying yourself by handling the MB?
I would prefer that they stay the same. Given: I don't know about actual/real data, but I'd be concerned about the 5V system dying on the MB therefore rendering the MB virtually useless. And since the MB a primary component of any system you'd have to RMA the MB or replace it at a fair cost. Whereas it currently is - you can go out and purchase a PS; or swap it out as some people probably have a PS that they could install, whereas most if not all would not have a MB they could swap out.
I wonder if they realised how intimidating a knife with "be quiet!" written on it would be before they made that
hahahaha
To the people who put the logo on the knife, it probably didn‘t even register. English as a second language, semantic saturation from seeing one‘s own company name everywhere and all that
be quiet! **STAB**
(of course this is comedy)
@@Lebon19 I know, if you didn't know what the brand was you'd say why the heck does that knife say BE QUIET!
@@volvo09 because it makes people be quiet. Duh!
I have a fear that this move will end up being problematic for a while. Motherboard manufacture tend to cheap out on components far more than a beefy PSU maker would. I can see a lot of RMAs in this conversion era happening.
This will play out for sure. Mobo manufacturers are famous for skimping on capacitors.
@@prashanthb6521 in this case, since the power conversion happens closer to the load, and the input currents are smaller (due to higher voltage), there's actually less need for expensive tantalum or polymer or (for cheap boards) electrolytic caps. Given that those are basically the only common modern electronic components that need a "lifetime" spec (due to being chemistry, not solid state), having conversion on the board is actually a win.
Caps are needed to supply power during the time it takes for the PSU to realise voltage is falling and to increase current to compensate. If the regulation happens on the board, the inductivity from the load to the regulator is much lower, meaning the regulator can react quicker, meaning there's less need for the capacitive buffer.
Same... Not to mention there's piles on top of piles of broken motherboards every time, same for PSU's and GPU's, but I'd rather have a 10 year warranty PSU going bad than having even more failure points on like a expensive 2 year warranty motherboard that will most likely only start to go bad after warranty ends...
And then I'd rather buy another well chose good PSU to replace a bad one than having to buy a very outdated motherboard to maintain a old PC cause it takes me MANY years to be able to buy a whole new one. The PSU even if bought to fix a old PC can then be used on a new PC at any time.
ps: F* this...😡
@@Validole You didn't thought far enough. Or better, you believe/expect to much from the Mobo Manufactures 😂
1. There need to be many capacitors, each sized for the specific need, instead of an big one in the PSU, usually high quality Japanese ones. (Which leads to lower efficiency)
2. There will be everywhere buck converters, cheap regulated buck converters, like MP1584 or even a cheaper texas instruments ones. Which leads to less efficiency either.
3. The mobo will run a lot hotter, because of the onboard conversation, which takes a lot more effort to cool, as in the PSU.
So in conclusion, Manufacturers will surely save out and only convert all voltages on one place, not on multiple places for every component, like people expect.
This is because, they will have only one place to put the converters under an heatsink.
The capacitors and converters will be the cheapest crap.
There only 2 benefits:
- First, that you eliminate almost any voltage drop, because the converters are still nearer to the components, means less cable length etc...
- Second, you will need an a lot smaller cable and even less cables coming from the PSU.
But i don't see any efficiency benefits, it's just some marketing bullshit to me which will simply lead to no efficiency benefit. I think that efficiency will get worse tbh, because of just shifting the converters to the mobo.
Only one converter for 12 -> 5v and another to 12 -> 3.3v (or even worse 5 -> 3.3v). So only 2 regulated buck converters to save costs, instead of multiple converters like people think here.
To make it really efficient we need to get rid of 5v and 3.3v entirely, but this is impossible, because usb power is 5v and almost all IC's use 3.3v.
I love the idea of smaller and less Cables from the PSU and i love the idea of the possibility of new SATA connectors, only with an single Cable directly from the Mainboard tho, but this 12vo spec won't be like people are expecting.
It's all just marketing and reasons to make everything a lot more expensive.
And Mainboards are already 3x more expensive as some years ago.
Cheers
@@RamaOlama Except motherboard is already doing a lot of conversion (300W+ cpu, ram, chipset supplies), the relatively low power 3.3 and 5V conversions are going to do very little to change thermal situation.
What this most certainly will mean is that motherboards will become significantly more expensive while power supplies will largely retain their price.
Definitely agreed on PSUs retaining their price! They won't want to give up the current pricing, even if cost goes down, unless there is sudden and significant competitive pressure from a newcomer. Meanwhile, motherboards seem like they're just infinitely skyrocketing...
@@GamersNexus If this standard is pushing more power into the MB for conversion and step down doesn't that mean more heat there?
So its more expensive overall. Takes up more space inside a build. Moves heat generation closer to sensitive components. For a benefit that won't benefit the end user and could be accomplished by other means.
@@GamersNexus its widely commonly known that enduser stick to a Mainboard for decadeds but switch the PSU on a yearly routine
Motherboards already have very high margins and are suffering from low sales due to the stupid prices that came with LGA1700/AM5 sockets.
If they push it harder into the customer, they will buy even less. This also means less CPU sales. At some point Intel/AMD will get grumpy that Mobo AIBs are sabotaging their sales because of the stupid mobo prices.
I'm guessing that' true even though I don't understand it. If the cost of the PS is reduced, wouldn't PS makers try to gain market share by lowering their prices?
From experience, this type of set up introduces a whole load of different motherboard issues. HP has done something similar to this for YEARS with their desktop models and motherboards are more common failures than platter drives, and that quite frankly is impressive.
Nonsense. I have a Hp sff with 240w original power supply 12 years old, I also added a Quadro k620 discrete video card from a workstation. I haven't had any problems, no damage. I only clean the dust thermal paste I have not changed and it works perfectly with the fourth generation operating system
@@stonehardy580 Well you must be the only one then lol, HP is hot garbage.
@@stonehardy580 You have to be the exception... I've seen this across multiple models. I'll bet a lot of the problems is the mobo manufacturer, last I checked it was Foxconn. Introducing more parts into a single point of failure inherently is going to cause more problems.
@@titan1983x The exact motherboard model is hp 1850. My model is hp 6305 sff with amd fm2 socket. The only fan is Foxconn, I replaced it with Artic 92pwm pst, not because it doesn't work, but because it's very noisy. The CPU heatsink is passive without a fan. I haven't changed the power supply fan - it's the original one, I just clean the dust periodically. Model is from 2012 or 13 I think, 100w processor a10 5800b and discrete video card from quadro k620 workstation work perfectly with 240w original hp power supply. I have had no problems, no defects or repairs so far.
@@titan1983x This is the internet, one anecdote obviously negates any and all claims to the contrary.
My concern with the 12VO standard is that by shifting more things onto the mobo there's more points of failure on a more expensive piece of hardware. Right now if something goes wrong in power delivery then the PSU can be changed. If the power delivery system breaks in a 12VO system you need a new mobo and all the headache and cost associated with that.
That makes complete sense and saw another user post similarly! Agreed on that. I guess the question is whether or not those parts increase failure potential in a meaningful way.
they could probably make the power conversion part of the mobo modular
One issue: Did you have a failed power delivery quite recently?
Even if it fails, you are likely to throw the board out.
@@cyko5950 Cost of a new power converter add in card-50$. Cost of a new motherboard-600$. I can tell you which one I would choose, and which one the motherboard makers will choose.
@@AlfaPro1337 decent PSUs just stop working when failure happens. If you were using a 20$ 850w PSU then yes when it blows the whole system is on fire
I'm surprised Intel went the 12v route for power efficiency savings, rather than extending the existing ATX standard with smarts (via sense/data pins like the NVIDIA 12-pin) to shut off rails when not in use. 12vo seems squarely aimed at OEM usage, rather than DIY. I can definitely see these standards existing side-by-side, potentially leading to fragmentation.
Yup. Imagine all the listings for components segregated into ATX & 12VO pages. What a headache.
That add pins and costs even more than obsolete multi-voltage ATX.
It's already fragmented and has long been so with OEMs having their own proprietary power delivery schemes. At least this might standardize things on the OEM side. As it is not only do you need the manufacturer's replacement or an adapter cable/voltage converter when attempting to replace an OEM power supply with an off-the-shelf part, you might need a model-specific version. But, yeah, this seems a net negative for DIY.
12VO being more energy efficient is good. But I don't like the notion of it being on the motherboard too. Board prices will go up, PSUs will NOT go down and now there is one extra reason for your motherboard to die, like they aren't the most common part to fail already.
Board prices are already going up - but many points won't change. I expect to see mobos with less pin out options dropping.
And another reason To make certificate thing to motherboard and make it even more pricey
sure, lets talk about power efficiency in a world where graphics card draws 300watts
@@Hsiss Only 300?
GPUs need a serious power efficiency change, that's for sure.
@@Hsiss meanwhile Intel 13900k reaches 350 watts.
The 5V and 3.3V rails are low current utilization now. There's already a VRM for memory (and on each DDR5 stick) , CPU; any inductor you see on a motherboard is part of a VRM. Most new PSUs convert the 5 and 3.3 from the main 12V rail (DC DC converter) anyway. I don't know why people freak out about adding two new voltage converters to the mobo.
Servers and maybe cars are going to 48V, voltage conversion is very easy and compact now.
Speaking of power supplies, I think I speak for a lot of viewers when I beg of you for more power supply reviews please!! Especially the lower end stuff! Aris is one person that can only review so many units, leaving a bunch of the budget segment a complete mystery!
Thanks! We are actually in the process of repairing our test equipment! One of the PSUs blew up a board....
@@GamersNexus Please post more on noise/spikes/surges tests.....power quality essentially. How well dirty mains power gets filtered should be know before we buy a PSU.
@@GamersNexus was it a Gigabyte? 🤔
@@GamersNexus i hope it wasn't the new 500 dollar board with the 7 segment display 😅
48VDC seems like the next step once 12VO is widespread. That would only require a 4-pin connection to the motherboard for most setups 😅
Yes it would but the cost of power converters that can work on 48V input probably start to cost heaps more. The capacitors and indicators and MOSFETs all go up in spec and cost as well on anything that is 48V. Not sure if this also translates to increased track clearances on the PCB. Don't go putting 48V into lower voltage pins. I've seen someone plug their LAN into an old 48V phone system. The smoke and charcoal was amazing. So was the user's expectation that it would continue working after they let the smoke out.
@@vk3fbab I regularly put ethernet cables carrying 48VDC in devices, no issue there
@@vk3fbab perhaps that old 48V phone system also had an AC ringing voltage? In any case, 48v is everywhere in the telecoms world and it's great. Also for a given load in watts, your capacitors and mosfets might need higher voltage, but lower current, so capacitors can have less microfarads to do the same job. ST electronics call sub-30v mosfets "low-voltage" and the 30-350v mostfets are called "medium-voltage", so no one would need any special components.
PoE+ and newer use 48-56V to power devices over ethernet cables. USB PD3.1 goes up to 48V. I think the limit before major protective changes are required is when you’re needing 60V and above.
How about 24V?
My problem with putting more and more on the Motherboard is, that you get more and more possible points of failure. And at the Motherboard prices often easily pass a few hundret € it gets expensive if anything on the Motherboard dies.
yeah, this.
it's a bad solution, for a regulatory problem....
Agreed. Overloading the motherboard with extra power, is asking for trouble. Motherboard will need significant heat cooling methods, even so, motherboards is the most sensitive component. Its better to leave the Power sockets and power stuff on the PSU as its own standalone handling.
Also, there are plenty of quality 24-pin PSUs for reasonable prices on the new and used market. Hard to convince consumers to choose 12VO when enconomies of scale favor ATX 24 pin PSUs.
Motherboards are a different situation, since you are more likely to be swapping it out. It may be better for DIY consumers to save on motherboard costs if they already have a 24 pin PSU.
That being said, OEMs have pushing their own proprietary versions of 12VO before the standard was created to save on costs to hit power efficiency targets. However, OEMs largely haven't moved to ATX 12VO, and having a standard for all OEMs to use will improve consumer choice and reduce e-waste IMHO.
OEMs also need to stop using proprietary board form factors. If they somehow save costs by doing something like soldering front USB connectors on the board, make it a ratified standard! Otherwise, they are just making more e-waste!
That's the sole purpose that intel promoted this
@@jay_ta708 motherboards won't need more cooling, you're running at most like 50 watts through the 5v and 3v3, and vrm are 95%+ efficient, so you have to dissipate like 2.5W in like 4 components at worse. dc-dc is also very reliable, it won't be a major increase in failure rate.
will motherboard manufacturers add 10-50 bucks to the mobo for a $1 increase in cost? most likely, but that's capitalism for you.
A few extra things to note:
- Backwards compatibility to older motherboards is going to be a nightmare. Having to convert down to the other voltages.
- Nowadays, a good PSU can last over 10 years, but people usually change the rest of the components earlier than that, keeping their current PSU. That means that you are going to end up spending more money due to the incresed cost of the motherboards and the more regular rate of swaps.
I don't see why it would be a nightmare, would be trivial to design a 24pin adapter that takes 10pin + PCI-e or 8pin power and uses step down regulators to make 5v, 3.3v, I'm not 100% how the -12v pin would work but is that even used anymore?
@@KriLL325783 even a negative voltage can be generated. No problem. Every current PSU does this already. They are already 12V only designs for years now, except the dedicated 5V stand-by supply.
@@ronakparikh Watch the video and actually pay attention, and maybe you'll figure it out.
But if you're too lazy for that, the short of it is: If it got completely adopted in just a single city, it could remove not just kilowatts, but MEGAWATTS off of the total power consumption. Linus tested these ages ago, and found a solid 30 watt reduction in idle power use.
@@ronakparikh not sure where you live but I live in a city of 4.5M people, the regional centres are larger than your 100k number.
It adds up very quickly, also Steve mentioned it more than once but it went over your head.
Backwards compatibility won't matter, you'll just use an older PSU. It's not like a change is made and you can't by old stuff any more, this will take years to get into all home PC's. If they can get it into all government and big businesses in the next 3-5 years that would be great, most of those things are left on so IT departments can run upgrades after hours, so if they aren't using as much standby power it's a big win for the company/grid/planet. 😁
One thing I always wondered the future of...connecting cases to motherboards, I always felt something much better would have come along by now
You mean like the front panel connectors? It is definitely one of the few things that has remained the same for over a decade. Still frustrating to plug them in when there's poor light or space!
@@GamersNexus yeah that was it.....was thinking at this point would have like one snap in connecters that links all the front panel ports and switches, I know somethings have slightly changed like the block u can connect the cables to and then connect the block to motherboard
Spring loaded clamps for each of the screw holes and maybe one button on the case to unlock
@@GamersNexus Don't most cases have all of the front panel connectors together now? Rather than having a bunch of 1 and 2 pin connectors, my case all of them are put together into a single connector for all of them together. It's a lot easier to plug that in.
@@MistyKathrine sadly not as common as you'd think
I really appreciate these deeper dive videos. Like I had heard about the 12VO stuff a little bit, but haven't really heard anything recently. So it was nice to get the run down on it in more detail.
I feel 12VO would be most useful as an OEM standard. From what I've gathered, the biggest excuse for oems using proprietary psu's is exactly because of the lower power draw. Lenovo has a 14 pin connector they started using in like 2014 for this reason. It would be nice for oems to use the 12vo standard so you could actually buy an off the shelf replacement for it.
Like BTX it should go to Dell to die.
Opiplex 3020s 9020s both is 4th gen intel use these connectors and they are a pain.
Exactly this, if 12vo is going to be a standard, OEM's need to pull their thumbs out of their ass.
@@AP_Sim They are non standard ATX12VO, all connectors are a mess.
They do it for a reason. Big companies like Dell, HP and Lenovo now, IBM, Fujitsu and Compaq back then, use proprietary stuff is because they sell to other companies. Those companies don't necessarily care about upfront cost as much as speed of servicing. So, of course, they made it so the only people who could service a unit would be that company and that only. No one could undercut them then.
I do like the efficiency aspect of 12VO, but I was also hoping that mATX/ ITX boards would benefit from reduced cabling rather than more expensive and bigger boards.
A lot of the voltage conversion stuff can be placed on the back side of an ITX board. Those are not gonna be as cheap as they used to be because they're no longer a budget option. SFF is seen as an enthusiast space alas. But I do think 12VO will be a net positive for ITX because that 24-pin takes A LOT of real estate and I think most could probably do without that extra 6-pin connector too.
I'm not gonna lie, the "be quiet!" knife at 7:16 is *mildly* threatening! What an interesting thing to print on a knife.
You don't actually need +3.3V. You can remove the +3.3V pins from the cables of your regular PSU, and the PC will probably work just fine.
10:52 The additional pins on the PSU side are sense pins, hence the doubled cables on the motherboard side.
As for +5V, it doesn't need to be created on the mobo, you can use an external buck converter, that you can stash somewhere else inside the case. Corsair already makes one ("+5V Load Balancer").
The big advantage of ATX12VO is that the PSU is much smaller (SFF, anyone?), more efficient, and quieter. Such PSUs will also provide much better performance at equivalent price, since there's only one rail that needs to perform well. Switching to ATX12VO would clear the market from all the garbage group-regulated PSUs.
Finally a sane comment here. Here in India I've only ever seen the garbage group-regulated ones. Pull too many amps on one line and the other goes out of regulation and potentially kills the mobo. I can't wait for 12VO to become mainstream.
I've never been a fan of 12VO, conceptually,
Motherboard longevity (or rather, the lack thereof) is one of the more frustrating things about the PC industry. CPU's outlast their motherboards significantly. Really, you can't expect a component with that many separate parts to be as robust as a CPU which has been binned for its unitary integrity already.
many CPU's in history last so long that they outnumber their host computers, then having to be scrapped purely because there are no matching components to make them useful.
It's rather frustrating, then, to see that standards are moving towards increasing motherboard complexity and responsibility, when, really, I'd much rather see more modularity. I think the E-waste saving potential of having a standard with more numerous, smaller, and less responsible components, is vastly more important than the power saving potential from voltage conversion efficiencies.
Well said. I have multiple still good & working CPUs (Intel Core 4790K, Intel Core2Quad Q9650) where the motherboard died and I couldn't find a replacement anymore because they were only manufacturing/selling newer socket generations.
What kind of modularity do you mean by (presumably) separating motherboard components and making them replaceable? It sounds like being able to DIY build your own motherboard, if I'm getting it right.
I personally would not like to see increased motherboard costs, but ideally if that weren't to happen and there could be other efficiency improvements by combining multiple power connectors and re-working expansion slots (and combining SATA data and power cables especially) I would quite welcome it.
@@jamesbuckwas6575 Ideally, I'd wish for the ATX standard to evolve into a less monolithic design. DIY motherboard is one way to think about it.
I think, Ideally, the motherboard itself would be mostly the chipset and CPU socket.
The goal would be mostly to reduce the size of the board significantly, by removing extraneous features.
The rest of the components, such as power management, I/O, sound, USB, could be put onto add-on boards, which either socket onto the smaller motherboard, or cable into it.
What that would look like, or how it would affect latency, I have very little idea.
Though I do know it is antithetical to the current momentum of motherboards, which are now beginning to expect you to install the storage on the board itself.
Perhaps instead of that, there could be a dedicated PCIe x8 slot that connects to storage board, which can be installed behind the motherboard, perhaps with the memory controller on the board itself? Who knows.
Perhaps the ultimately solution is to replace most I/O with PCIe. Making CPU's more variable. Return to soundcards, and the likes.
Perhaps with separate PCIe-slot boards, linked to the main motherboard with a special, denser riser.
The worst part is all these e-waste generating measures are sold as "green". Many accept these claims at face value, and the community fights itself over the often non-existent environmental merits. In this case, you'd need to save an awful lot of watt-hours to make up for disposing of motherboard sooner and replacing it with a newly manufactured model, and we're talking about a watt or two on standby with these changes. It's possible that the breakeven point is hundreds of years with something like this.
@@mattrogers6646time to break out the aliexpress salvage boards
I think the endgame would be something like removing the 5v and 3.3v rails completely, basically giving components like 2.5/3.5in drives a 12v input which they convert down internally. The current layout with 12v to the motherboard and lower voltage to those off-board components means the voltage converters on the motherboard still need to be overprovisioned since they need to be capable of providing the maximum power when all the components are connected, when in reality it's likely that only a handful of them are hooked up.
12VO can be huge for SFF builds, since the PSU can be made far smaller due to simplifying the DC-DC circuitry and cutting out 3 rails of filtering.
I feel like that last thing we need is more things the motherboard manufactures can use to "differentiate" their product lines.
I run 8 SATA drives, would I need to get some sort of special add on card to support that many?
you could use a 6 or 8 pin add in card. 12V to 5V is very easy to do and there's already supply chains for it. you (or a manufacturer) just have to add a 6 pin connector on one side and sata connectors on the other. there's probably gonna be single active cables that do 12Vo sata to 12V+5V sata too.
@@satibel Add in card? Like PCIe SATA expansion cards that will take power off the slot and give you 5v? If that is the though a 1x card it would be limited in power output.
@@IrocZIV I'd say it's more likely that you'd take one GPU plug and plug it into a dongle that would put out SATA power connectors.
@@Validole Ah, yeah that would make more sense.
@@Validole Better idea: How about we have the power supply do that job and not have extra add-in cards?
It's a good thing that framework is starting to offer their products in standalone packages. If current motherboard pricing gets too far out of control then for my next build I will probably just get a framework mainboard and put it in the cooler master enclosure and use an external GPU. Also especially since SFF looks like it will get screwed by this change.
I want to see DTRL get a standard MB formfactor. Some of those laptops have a larger build volume than some SFF desktops.
If you have a 12VO arrangement does that then mean you could make a case that’s powered straight from a laptop style power brick?
My concern is real estate on the motherboard, especially on ITX cases where space is already at a premium. Then again, on a very compact ITX build, does a 5V line coming from the motherboard even make sense, if it eschews even the SATA data ports, for m.2 NVME only storage.
Otherwise, 12VO would definitely make life easier not having to manage a 24-pin cable, though that only matters 1-2x a year when I would be doing anything with a PC internally.
Yeah, looking at this I immediately jump to builds where every component is 12V, right? If the standard started to gain traction you can imagine a scenario where some of the other components start shipping with 12V out of the gate and unlocking more builds where you just outright don't have the other rails at all. Or maybe some of the remaining components do the conversion themselves, like a breakout LED box with the 12v to 5v in the hub. Presumably you eventually have a simpler, tighter build with way fewer unpopulated modular outputs and a smaller PSU.
But I feel that's like when USB was supposed to be "universal". The complexity is still there for the legacy stuff and just gets spread out across the standard eventually.
+1 for itx! Love my small, custom SFF cases and with how much space is a premium on ITX boards I fear what they will have to remove to handle voltage step down on the mobo itself.
@@mademedothis424 USB C basically did away with the simplicity that USB added. I see lots of reddit posts where novices are confused with what their USB C port on their laptops are capable of doing, particularly DP-alt mode.
Definitely very problematic on ITX motherboards. If you want to do conversion from 12V to 5V and 3.3V on the motherboard, you need to allocate space for capacitors, high side and low side mosfets, inductors for each rail of 5V and 3.3V. Then, you need to put SATA power headers. ITX boards are already so cramped as it is now, I don't see how they are going to do it unless sacrifice DIMM slots for SODIMM .
@@fleurdewin7958 Don't bother with the sata headers, provide a breakout board with sata connectors. Between that and going away from the 24pin, I don't think it's impossible and the reduced complexity on the PSU could make for extra tiny GaN units.
18:58 One interesting thing about that Motherboard is that it has the mounting holes for the VRM heatsinks. This means that they planed to put heatsinks on that board or they intended to release a higher price variation with heatsinks.
The fact that it has the mounting holes also makes adding custom heatsinks much easier. Its rarely done by moders but the adding of heatsinks to cheap Motherboards that doesn't have it can increase its values by something between 20$~80$.
The biggest difference between low end and mid end motherboards many times boils down to the addition of the VRM heatsinks.
The mod is relatively easy to do. A hand saw, a drill, old scrap heatsink, some thermal pads, some small springs and screws with nuts is all you need.
Can’t wait for budget motherboards to cost $300+ with atx 12VO and PSUs to still cost the same
Gotta chime in and say that 12v only (sometimes even 24v) PSUs have been used in data centers in the US for several years now. I have been working on them for at least 3 years now. They are more stable in every sense than traditional PSUs. Power distribution boards then do all the "magic." This has also been great for cable management and thermals. I would love to see this adopted more mainstream so it becomes more available, aka more reasonable prices. Good day ya'll.
I do some 12V-intensive things at home and my go-to for those tasks are used 1U server power supplies. 12V 50A+ single output and they're generally free. Some I have are more than 15 years old and still working just fine.
ATX12VO is a standard that is there to make the SI look good because they can print a good PSU efficiency spec on the case, but don't have to worry about the losses "downstream" because that's on the motherboard manufacturer, and not relevant to the spec.
Two important points to 12V only is: 1. Easy battery backup: Just connect at 12 V battery. No 110V sine inverter and other lossy and expensive system. 2. Reduced cost: The motherboard only converts the voltages it needs, and as much as it needs. Usually -12V and 5V is dropped. This simplifies the system overall.
I don't know how significant it is in the case of computer power supplies, but transmitting a given amount power through a wire at a higher voltage is also more efficient because the lower current leads to less resistance loss in the wire. So by moving the point of conversion to 3.3v and 5v closer to the components, there is less overall loss due to resistance in wires when those components are under load.
while true, this is only insteresting for power delivery infrastructure via high voltage lines that are in the ballpark of 380,000V (germany) that can be up to 450km long for 230V households.
3,3V, 5V and 12V isnt much of a difference and that not even accounting for the short cable lenght of half a meter.
@@fmskaisergsaw5567 ohms law is useful everywhere, not just high voltage lines. The telecoms world standardised on 48V supplies decades ago and for most cars they moved from 6V batteries to 12V in the 1950s. Most diesel trucks use 24V batteries rather than 12V, at least for starting the engine. In many data centers in the US, they distribute power at 240V, 480V or higher, rather than the domestic 110/120V that is common for home appliances. Whenever you have high power loads it is more efficient to do your power distribution at a higher voltage.
as an SFF PC enthusiast, I really can't wait for atx12vo. Cabling is way simpler, and without the 5v and 3.3v circuits in the power supply, manufacturers can fit more power into smaller PSUs. I personally am curious about how complicated Z#90 mITX motherboards will become, but considering manufactures like ASUS is taking inspiration from lego with how many stacked carrier boards they use, I believe it can figure it out.
i like the power supply handling everything instead of a motherboard, since you might need more connectors down the road or hardware configs might change and need different plugs so you could be screwed since the motherboard has so many, and can physically can fit so much already
And that's why you want to handle power conversion on the board. One pair of 12V can carry more power than two pairs of 5V or three pairs of 3.3 V, and the DCDC is much smaller than the additional space taken up by the wider plug.
@10:45 That variable resistor is likely a trim resistor for doing fine-adjustment of the output voltage, you can easily validate by measuring the output voltage while adjusting it. It's fairly common to find on open-terminal power supplies for industrial purposes where you just screw some wires on to use it. The primary use of a trim resistor in these kind of power supplies is so you can compensate for voltage drops during high power draw (depending on connected equipment) by essentially adding a voltage offset to ensure it's always above the required threshold regardless of load.
It's tough enough picking a good motherboard that has the features I want, at a price point I can stand, and looking good (not a point for some but definitely a point for me) based on how things are already. I don't want to worry about even more variation being added to motherboards.
I'm half asleep so may not have fully heard and understood the reasoning, but if the reason is for lower power draw in standby or low power mode, couldn't something similar be accomplished with a data connection and a little bit added to the PSU? Or is it simply just that much more efficient still to stick it to the motherboard instead? I wouldn't mind a changing standard if it's not making as drastic a change to motherboards.
Would be interesting to have a sit down with an engineer that worked on the newer spec to talk about it and why that was what they went with rather than other options. Although maybe there's already good info out there for why not that I just need to actually read/watch.
I think that main reason is, that ATX is ancient. It was developed in times when switching DC-DC power sources weren't as cheap and it was more efficient to make -12V 5V and 3.3V on a smaller transformer in the PSU. Now the PSU uses DC-DC converters anyway and -12V is not used, it makes sense to reduce wiring to only 12V. The converters can be optimised for onboard devices and managed by MB's logic.
@@Papinak2 A better solution I would still think is segregating responsibility of tasks. A power supply should do what it says it does. A motherboard should only serve as a board for other items. The standard could have simply asked for some pmbus standard with specific addresses and communication between motherboard and PSU. Although complex to implement, I would expect a lot less issues and less components means lower cost and higher reliability. This seems like a "standard is outdated we will simply do everything because working to help change the other parts of the industry we don't control is hard".
@@epeglab1o1 This is impossible. CPU's and such use very low voltages these days, so that buck converters ("VRM's") have been necessary on every mobo for decades now. It will be a nightmare to even imagine reliably supplying dozens of amps at 1.2V over any length of wire. And totally inefficient.
@rexsceleratorum1632 okay you took my comment too literally. Yes the mobo does have multiphase buck and other bucks to provide various voltage rails for the cpu.
गहन शोध , स्पष्ट अवलोकन , प्रज्ञावान विश्लेषण , सुन्दर छायांकन कुल मिलाकर बहुत सुन्दर प्रस्तुति।
I'm not in favor of shifting the cost to motherboards. I've invested into a 12 year warranty evga model, and with intel switching motherboards every two years, and even amd probably not having am5 lifespan as long as am4 - I'd rather buy psu once, and then use it for a long time, rather than paying extra each time I change platforms.
Yeah, motherboards have already almost tripled in price. What's this going to cost? Another tripling of the tripled price? 🤔
@@dakoderii4221 even if it adds just an extra 20-30$ it's still going to be recurring cost. There's no benefit to changing your psu every time you rebuild your platform, and I'd like to stay that way. We're getting here into dangerous territory of pricing people out of pc gaming.
7:25 After years of cutting/trimming stuff on electronic assemblies, I found the best cutting tool is just an actual scalpel blade. I use Swann Morton non-sterile carbon steel blades, they come in different shapes but mostly #10 for slicing and #11 for scoring
Another advantage of ATX12VO is better dimensioning of the +5V and +3.3V rails as regular ATX psus have grossly overdimensioned rails to account for the possibility of them ever being connected to an ancient computer that may pull power from them. The board manufacturer will have a good idea of how much power could be pulled from those so it should end up being cheaper. The negative rails could also be omited, most modern computers don't use them at all.
The +24% cost to matx boards and needing extra layers sounds like bs to me (I have designed 4 layer pcbs for fpgas and microcontrollers). You could have the dc-dc converters required near the connector, if the power is low enough the area freed up by the smaller connector may be enough for the converters and that way you just route the output much like you always needed to. I only foresee an increase in cost if the board packs a crazy amount of IO like usb or sata which could potentially draw a lot from +5V and +3.3V needing beefy multiphase DC-DC converters in an already cramped pcb but that shouldn't be common.
Thanks SOOOO much for this. I've been waiting for a good 12VO overview and this delivered.
in my opinion ATX 12VO could be quite usefull in office pc´s but it think the normal atx PSU´s should still be there for the DIY pc builder
Yes, it's basically up to the OEM's, system builders don't want it.
That just makes DIY PCs expensive for no reason, because the market for PSUs will be small and full of enthusiasts who are susceptible to advertising and willing to overpay for form-over-function nonsense like sleeved cables in cloth bags. It's better to be a small fish in a big pond.
Multi-rail ATX belongs in a museum, not your computer.
Would be good to see a bit more on this for the ITX space... Seems like the PSU's themselves are smaller as-is, but if even mATX boards are hurt by moving all the voltage conversion onto the mobo (increasing size/layers/cost), that could be even worse for ITX. Seems like the best-case would ultimately be for end-components (PCI / Graphics cards, Hard Drives and SSDs, etc) to do their voltage conversion internally so the motherboards can be passive 12v as well...
The cost changes and the cable management are both absolute No-Go.
The biggest problem that I see is, that PSUs very often outlive Motherboards. My current Thermaltake 650W PSU is now nearly 14 years old and will power my 3rd PC Hardware Generation now (Core2Quad Q9650, i7-7700K and now i9-13900K). So the same regulators reused all the time and still working perfectly fine. With 12VO the additional hardware on the motherboard would have to be paid and then gone to waste every time, raising the cost and amount of E-Waste. At the end we WILL pay for the extra components and MoBo prices are already insanely high with all the crap put on them. This might be beneficial overall for commercially used all in one systems but it will hurt ANY DIY build.
Not even speaking of additional coil whine risk with open regs on the board compared to enclosed and dampened regs inside a PSU case.
It takes a great deal of faith to entrust your daily driver PC to a power supply from mid-2009, complete with group regulation and an efficiency level of most likely either 80Plus White (W0104) or Bronze (W0163). Yes, one would rather hope that a quality unit makes it that long with the occasional dusting, but it should rather be enjoying its well-deserved retirement in a more fitting retro build by now.
I have things to do but hearing Steve talk is more powerful than family guy with subway surfers
I'm glad to see GN get back on the PSU topic. I've really been hoping that you would review a few PSUs a month because the market desperately needs reviews.
This is why Steve is one of the best!
I totally agree with your perspective on this and don't get the lack of basic debugging hardware indicators on newer motherboards.
There is a SATA connector (SFF-8482) that includes the power connector within the same cable and connector which really should be used for these 12VO motherboards. This would make cable management for SATA drives a whole lot easier. Then motherboard manufacturers can just provide another one or two of the 4-pin motherboard headers for users to connect up non-SATA peripherals that require a SATA or Molex power connector.
Or, you know, they could just make a whole new combo connector for the motherboard side that provides both power and signaling which has a SATA data/power combo connector (SFF-8482) on the other end of the cable - they could even double up on this to provide SATA/Molex power connectors from that same header using a different cable if you don't need all of your SATA ports. Then you could have 4-8 of these ports on your motherboard and all the motherboard manufacturer needs to do is provide the cables.
Get more hard drives that need more power than the motherboard can supply, replace the motherboard (or fry it). Hard pass. 12VO PSUs need a 5V rail for drives, period. Let the MB take care of -12V, 3.3, fine. But the sole source of 5V shouldn't be the MB.
@@TiagoJoaoSilva Considering that SATA ports only support one harddrive per channel I highly doubt that you could overload the motherboard's 5V regulator if the mother board vendor actually speced the VRM properly. That said, it is the 12V rail that gets hit hard when you have a lot of harddrives as that is what the harddrives use for their motors and the peak power used is when all of the drives are spinning up. The actual electronics which use 5V don't draw that much power.
What you should be more worried about is having a whole lot of aRGB LEDs running. They use around 150mA per RGB LED at full brightness and if you add in a bunch of strips then the power draw can really start to add up. This could be easily solved by having a 12v -> 5v control box for the aRGB strips that can take in the data from the motherboard and supply the 5v required via a 12v->5v buck convertor.
That said, I love the fact that you think that this is a problem that is unique to 12VO. Staggered startup has been a option on ATA controllers designed to take a bunch of harddrives for a long time as spinning up a bunch of harddrives at the same time has the risk of overloading the 12v rail and triggering OCP.
I'm getting flashbacks from my project of converting an ATX power supply to work with a Lenovo OEM motherboard.
For anyone interested:
It's almost like a 12VO connection but not quite. You connect the 12V wires and ground directly, but then you need to hotwire the 5V standby to ground. My PSU also required a load on the 5V rail, so I put a 10ohm resistor there. It works fine now, but holy smokes what a headache.
I have a bit of a bone to pick with 12vo. I don't like moving all the voltage regulation stuff from a 150 dollar power supply to a 500 dollar mainboard. I'd rather replace the former if something goes wrong instead of pissing around pulling a much more expensive and harder to get motherboard. I could be crazy but this seems more work than is necessary.
And yes, they will absolutely charge me more for both the new 12vo power supply, and absolutely will for a new 12vo motherboard. Because reasons.
If it impacts board failure rate, then that's definitely a good point. Even if it's not a $500 board but a cheaper one, it's still a lot more work to replace.
@@eye776 Tell me about it. It will stay the same or probably go up because 'new spec'
@@eye776 PSU prices will stay the same but the motherboard skyrocket
I would bet they would raise the price of PSUs along with the price of motherboards and then tell us what a great deal it is.
Exactly, except now it's a $700 motherboard with lousy power conversion parts. Caps don't last forever.
Another reason for the transition to ATX12VO is that 3.3V and 5V are difficult to efficiently deliver over long cable lengths. Those long cables act like antennas, picking up unwanted EMI. Moving all of the conversion to the motherboard is much more efficient.
Changing the motherboard connector is probably overdue, but will be a hassle for people to track between the two.
Always looking forward to a more simplified connection setup and more power efficient setups. I just hope we can adapt existing power supplies to the new connections when they become more mainstream.
And pay twice for a mb.
Very interesting to see! Considering the points about Mainboard size towards the end of the video, I'm a little left to wonder about ITX size boards with 12VO, as it would be especially interesting for small builds, having the reduced footprint of the 12VO connectors and a simplified PSU. But I'd think it would further increase cost of ITX boards, needing more layers or probably even stacked PCBs to fit everything else into their size... I love building in SFF, and 12VO seems like a step in the right direction, I'm especially thrilled to see the I_PSU% feature coming one day, I just hope pricing won't be insane and that complexity doesn't actually increase from trying to decrease it at another level...
The footprint is only smaller with the connector, Steve was ignoring all the circuits needed to get the lower voltages now taking up space on the board making the actual footprint larger even though the connector is smaller.
There are SBC's for ITX lovers.
@@JETWTF yup, that's what I figured. It's just moving stuff from one place to another. Everything that had its place well within the PSU itself now needs to move over and require even more space on the board.... While PSU probably won't even change their form factor
@@encryptedunicorn7659 I wonder about the improved efficiency since it is just moving PSU functions over to the motherboard. Could that efficiency be achieved without moving the PSU onto the motherboard? I am not an electrical engineer but I cannot see how a few wires makes enough difference to matter unless the wire used is wrong and thats super easy to rectify.
I've been shopping for PSUs lately and have been confused by the varying pin-outs. This video perfectly explains to me WHY I'm confused (but I'm still a bit confused).
My main complaint is this is only going to make motherboards more expensive and 12vo PSU's will be more expensive than current atx PSU's
Hello Steve, because of moving RAM VRM from motherboard onto DDR5 RAM Module itself, now I am experiencing Coil Whine directly from DDR5 RAM when RAM intensive loads are in operation. That kind of coil whine was NONE-Existent on DDR4 motherboard. That actually mean DDR5 RAM makers either don't know how to build a good VRM or they use cheap components. But Motherboard manufactures knew/learned from many years of experience what to use as quality VRM for RAM.
Now if motherboard makers, which do not have the experience of PSU makers, start adding part of the PSU (5V+3.3V circuit) onto the motherboard, Sure we will have more possibility of coil whine + increased probability of failure for a motherboard.
The motherboard already has so many different circuits that if any of them fail, the manufacturer should take back the whole unit and change it with another (at least within warranty time), now adding another new circuity with high probability of failure (due to manufacturer's not having enough experience like a PSU maker) it will ultimately increase chances of failure - equally increase motherboard manufacturers cost for after sales services.
Add the fact that more and more users are turning their PC case into a Christmas tree by overusing their 5V rail and adding multiple ARGB devices.
Last but not least, the Motherboards, specially mainstream and high end classes, are already filled with so many connectors that make cable management a difficult task, now adding more cables out of motherboards, will just worsen the cable management more than ever.
Rather than making the 5-12VSB inline on the cable, they should make it a module that you plug into the PSU, and then plug the cable into that.
I never thought fully about the office PC situation regarding 12VO. With the tiny SFF computers, sometimes smaller than ITX Windows machine's with external PSUs (the laptop style ones I'm blanking on the name) something 12VO seems even harder to implement, and that's a huge chunk of business I'm sure as I see them everywhere. One more reason Gallium Nitride/Galllium-related tech is so important and promising lol.
And as far as usability of other things, I agree there will still be plenty of adapters or retrofitting supplies etc.
I think doing the voltage conversion is better off done inside the PSU. Not sure if moving the down conversion to the MB causes more noise issues but it certainly adds more heat where you don't want it.
Not necessarily though - if you can move the power to where it needs to be at a higher voltage, the current is lower. Which means less losses in the traces. Of course the conversion also has losses, they're at efficiencies of 95-98%... in the end it's probably a negligible difference though, as CPU and GPUs currently use the 12V rails and convert down already.
Noise is probably better too, as the motherboard manufacturer can add filters without having long cables that pick up noise from the environment.
But price is a different topic, and reliability.. you can just replace the PSU and I don't think that a PSU with 12V only will actually be cheaper..
@@marcelleuenberger4648 I don't think your going to see much difference in line loss. 98% of which gets turned into heat anyhow. An I9 alone can soak-up 300 watts when throttled.
They filter power supplies as well which can house much larger components like caps and inductors for filtering. My Corsair AX1600 has a very clean output and is 94% efficient. The PSU is one of the largest sources of noise in the system anyhow. A poorly designed power supply is going to introduce noise in the system no matter how many rails it has. Hardware engineers add filtering all along the circuit even on the front-end of the chip I/O so it's usually not a problem.
I'm sure both set-up's have the pluses and minuses. Should be interesting to see if this catches on.
@@donuthole7236 Most power hungry devices are powered from 12V being converted locally anyway (CPU, GPU) because it is not practical to do it any other way (450W being pushed to GPU at ~1.1V is over 400A - and numbers are similiar for the CPU (which needs few different voltages that are generated on the motherboard)). CPU, ram, chipset alone need 7 or 8 different voltages (DDR5 needs 3 different voltages (usually two of those are identical, but that is not guaranteed by the spec)).
Adding 2 extra votlages to all of this is going to do very little to BoM costs, but are the manufacturers going to use this as opportunity to increase price?
@@donuthole7236 Of what use is that filtering in the PSU, when you send that current down to the MB, over cable and two connectors? Yes PSU filtering isolates the noise from the main, but does not help with noise and voltage oscillations down the line...
Love that old scope you have on the shelf behind you Steve.
We've moved well beyond the times when changes in technology were made to improve said technology. Any change now is simply an opportunity to charge more. This change would somehow increase the cost of power supplies and motherboards, and likely lower the lifespan of each to boot - exactly what manufacturers want. I'll wait for your testing on these which I'm confident will show little/no increase efficiency, with significantly higher cost for both PSU and MB.
Technology already improved, which is why ATX wiring is redundant. You don't really need to convert 12V to lower voltage in the PSU, so why not move the converters to MB? It can be optimised and managed by MB logic - for example, shutting off 5V to SATA, if it's not needed, but keeping it for USB.
Sounds about right!
I have custom built NAS, based on Advantech board based on Q170 chipset and Pentium G4400 CPU. It's powered by 12V power brick and it's crazy efficient compared to my previous Atom setup.
How do you power the drives? The SATA connector has 3.3V, 5V and 12V pins.
There's something sinister about a knife with the phrase Be Quiet! on it. Instead of telling your victim to be quiet because you just pulled a knife out, they'll get the hint real quick.
I hope it never does on home PCs. It’s very suspect to have the voltage regulation in the MB, if it does break it can cost 700 bucks if the MB did. cost as much. But if something breaks, the PSU can be replaced for cheap, easily and it also has voltage protection on it so it can’t break other components.
Think about that!
Technically, all the existing PSUs based on modern DC-DC design are already ATX12VO considering the fact that the PSU itself only converts mains voltage to 12V and then this 12V gets further converted into 5V & 3.3V for minor rails (all inside the PSU). Now the major problem is with the 5Vsb that eats up tiny bit of power even when on standby mode and instead of figuring out a better solution to this standby power consumption problem intel stupidly (imo) came up with a solution that seemingly disrupts motherboards as well as PSU designs.
Edit: The original was incorrect but is kept below for context. USB PD support various voltages including 12v.
Original message: USB really fits weirdly with this. Just about the only voltage that the PD spec doesn't provide for is 12V.
Yes it does. It's complete bs that they don't. I've seen so many articles talking about how very few pd chargers actually have 12v. I've checked every single one in my house and car and they all have it. I checked Amazon listing specs and same thing. Gotta be careful tho because some charger support 12v only on usb a like anker chargers. Got me wondering if I should test mine to make sure they support pd on usb c since mine have usb a too
@@andrewk8636 Oops, turns out you're entirely correct and I have got it wrong. The reason I made the mistake is that I looked at a table of USB PD rev 2.0 voltages and currents that didn't list 12v, missing that the table only listed values that weren't already present in USB PD rev 1.0, which already accounted for 3 different current levels at 12 volts. I've put a correction in the original post.
@@andrewk8636 Apparently the wires matter too. To provide more than 60 watts of power your cable has to be able to tell the other devices it can handle that. Then again, that's probably more relevant to laptop charging than desktop pcs and phone chargers.
@@martineyles wires shouldn't matter at all for the voltage. Amperage is another matter. If the wire can do 9v and 20v there's no reason it can't do anything in between
@@andrewk8636 I was talking specifically about higher power specs above 60 watts in the last reply, available at 20 volts and above. Eg. 28V at 3.57A for 100 watts. That's why I stated it was more relevant to laptop charging. It's possible that it's because of the > 3A current though, as I don't see the option of 3A, 28V, 84W.
People are talking about motherboards dying but even without that, you still usually need to upgrade the motherboard more often than the PSU. I can buy a really good power supply if I only have to do it every 10 years, but I'm already keeping all of my hardware for a minimum of 5 years in the best case, so it might get even worse.
I know there are a lot of possible issues with this, like more points of failure and the injustified markup that's bound to happen, but let me dream for a second: just imagine how a build with this could look if you just used M.2-SSDs. You would just have two big cables, one going to the MoBo and one going to the GPU. In addition to that you could daisychain the casefans and now cablemanagement would be a dream. Especially for SFF it would be great to have the option, then you could choose a cleaner build for more money and otherwise you would just take a standard 24-Pin MoBo.
You'll still have EPS connectors.
Also itx boards are already being stressed to their size limitations and now you want more DC-DC conversations to happen on board than before? Good job.
@@NotThatGuyJD Well I said I am dreaming a bit but I will kind of try to address your concerns. The boards wouldn't necessarily have to be bigger. For example there is more I/O than most people ever need on some itx boards. Some people would be willing to trade that for better cable management. As for the EPS-Connector: most CPUs could theoretically work on the power budget that is provided by the 10 Pin. A 13900K? Probably not, but that one consumes way too much power to begin with and if you want to use it you could just poupulate the EPS connector or use another board.
I like it. It may bring a motherboard plug that would be easyer to remove (aaaugh my hand always hurts doing that), And the motherboard could have small lvo for a handfull of connections, like a 3.3V one for each ssd, one heavy duty for usb ports, another for the rgb lighting, etc...
The way I see this is power supplies will become cheaper to manufactur... but the manufacturers won't actually drop the prices.
This will only increase motherboard prices making it a bad deal for the consumer.
Also this will add more points for failure on the motherboard itself.
Yay!!!! Not like motherboard prices were low already, right?!
If you want to know how this can play out, just go look at current Dell/HP motherboards on their business lineup to see how bad it can get
Difference there is that the Connectors are completely undocumented in most cases, and not everything has an adapter to make it work.
Dell/HP also don't just use one odd layout of connectors, they use different ones for different products.
Here it's a standard, documented, and adapters exist to reliably convert from one to the other.
Dell and HP are selling their business lineup to accountants who make their purchase decision because one vendor provided the same specification 50 cents cheaper than the other.
I have an additional 6pin PCI-e power connector on my X670e motherboard.
Didn't realize what it was until I looked it up and it's just like the one you point out around 20:00.
I really hope that there will always be a traditional ATX variant of motherboards, provided that 12VO takes off.
µatx & mini-ITX is all we need 😏
With a 7-10 years warranty PSU you know you're getting quality, it's very doubtful the mobo manufacturer will be putting in this much effort.
If we could combine power and data into one motherboard connector it could reduce cables, material usage, and clean up builds. I don't hate it.
That'd be cool! Seems like it'd rapidly enter "we didn't like the standard so we invented a new standard" territory for the pre-built manufacturers, though!
@@GamersNexus Good point. I don't like the lack of backward compatibility either, although adapter cables should be feasible. Make it an open standard agreed by a consortium?
@@odjsjaks The PSU could be lower cost due to reduced components and complexity as it only needs +12v and GND, like the PSU's discussed in this video. Also cheaper due to fewer bundled cables.
The motherboard would increase in cost, offset by PSU cost decrease, as it now has the 5V and other circuitry.
The data cables would increase in cost because they now must also include power, which adds more complexity and material, but this should be offset by PSU price decrease.
When buying a new PC this should all balance out to no change in overall cost.
Adapter cables can provide converters to allow interoperability between old and new hardware. So you can keep your old PSU, motherboard, or other hardware, if you like.
PSU manufacturers may make less money off this, whilst mobo make more, maybe? It's hard to tell how it would shake out. Regardless, this is a quality of life improvement, not a money making scheme.
If it's just 10 pin replacing the 24 pin I'm all up for it. I'm due for an upgrade anyway. But this one seems like not only it would increase the cost of the motherboard, because the extra component needed to power the remaining stuff, it didn't reduce cable clutter either. Look at that janky ass power cable from the mobo to sata.
@@Leo99929 Have you ever met a Manufacturer of anything who reduces cost on a new generation of a thing? I wouldn't expect them to start now.
I really love these indepth stuff you do did not even know that there was a new version of PSU lol. I got mine in 2017 and it's still going strong.
Ty Steve
Yeah, while I understand the gain in efficiency of 12vo, I don't think integrating more features into the motherboard is a good idea.
What I think would be a good compromise would something like a 12VO PSU, mostly as as an AC to DC converter and voltage transformer, connected to something similar to a DC to DC picoPSU in a breakout board which could be located near the board as to reduce losses.
In this i prefer the seasonic syncro, you have the power supply providing 12VO to a external unit that do the conversion to the mothrboard.
So my Seasonic Syncro has a 650 W PSU that is 12vo and goes into a distro board that converts everything into normal stuff. Not quite the same but I love it as a case. (Would love to see you guys review it someday)
Iirc, they already did.
Moving this type of change into 1 form factor would drive this, like into ITX- dtx, in the mean time support an interface between them, simple 12vo ps feeds small adaptor for power division for use on older boards just like that Corsair cable. Nice old silly scope on top shelf left. Ipsu looks like current monitoring nice. I will be looking into this for next build as spec matures
This could be very interesting in combination with the last couple of motherboards with reverse mounted headers and stuff because I can see all of your sata connectors and a lot of your lower voltage adapted headers coming off the back of the board making cable management easier
I don't see how that's any easier than cable management now. Plus it requires you to change your PSU, motherboard, and case for no good reason.
HP does something similar in some of their small forms office PCs.
We had one. We needed to add a video card that wouldn't fit. So, lol, we cut the motherboard tray out of it and a Dell optiplex case and soldered the HP tray into the Optiplex. We installed a 600w psu and found a problem because HP used a 12V standby voltage. We had to purchase a small voltage converter circuit and glue it on the side of the psu, ran the 5v sb to the converter and the new 12v sb to the HP main board. I had to build a converter cable from 24 pin to the HP connector.
Everything worked except we kept getting a fan error. Turns out the HP power supply sent a tac signal for the PSU fan through a second connector to the motherboard. We were going to tap into the CPU fan tac to correct this but since we could bypass the error by pressing a key during boot we didn't bother.
Frankenstein worked great for two or three years before we replaced it.
I am curious to see if this new standard works out or if people just say "meh" and keep purchasing boards and PSUs using the current standard. I know I have no plans to replace my 2-year-old Seasonic.
I hate moving the SATA connectors to the mobo. That's such an hp thing to do.
Maybe with the smaller footprint of the connector, board manufacturers actually have the space to put debug code to every board. :D
I mean high end mbs are usually the one with debug code displays and more connectors. It has nothing to do with space.
@@erebuxy I obviously know that. This was just a funny reflection on GN's video from yesterday or the day before.
This will result in motherboard manufacturers cheaping out on components to create the lower voltage power rails, then troubleshooting power issues will be 2-pronged. At least with the current state we know that if there's a low voltage issue on the USB, it's a power supply (or load) issue. I like the idea that all power rails come from a power supply that's been tested, certified, etc.
I'd be worried about the mini-ITX form factor.
Is there enough space on those boards to handle this 12V to 5/3.3V conversion?
they will find a way to do it and charge us 400 for a basic itx board
Looks like your run of the mill boost converter circuit with a 12V clamp, looks like a quick and dirty fix for the voltage conversion. Likely it will be totally fine for the conversion as standby voltage is only drawing power when the board is off. That definitely beats tossing out the PSU you had before.
Steve Burke is just Jason Momoa's long lost twin brother who got into computers instead of acting.
You got to love the
Popp
Permanent
Over
Power
Protection
So not only are motherboards beginning to be kind of unaffordable but they now also get new things that could fail (power conversion parts which were previously handeled by the PSU) I'd rather replace a faulty PSU which costs ~90€ than replacing a motherboard which starts at ~120€ because some power conversion parts are broken or "smoked". I'm not sure if I like where the PC development is going regarding the changes mentioned in your video. Not to mention the effort that changing the motherboard brings along.
Nevertheless: Great video, keep up the good work at GN ^^
The extra lines are like used to reduce Vdroop on the 12V rail under CPU load, this drop in voltage is caused by the inductance to the PSU under fast switching operations (what you might call transients).
Would it be possible to transition to a single cable for both SATA (data) and SATA power ? With an adapter on the device side to accept a single cable instead of two. I think this could reduce the "worst cable management" aspect of 12VO boards.
It is possible, an cables like that already exist.
BTW, one can make an adapter cable to power off mainboards with 12V only PSU. It would also have a PCB that makes the other needed voltages.
If the board will handle the power supply, wouldn't it be better to combine both power and data cables into one. Eg: for sata drives you have one cable that carries both power and data?
I just bought a 750W platinum rated PSU. It's fully modular so luckily 12VO compatible.
My main issue with 12vo is your adding more points of failure to your board. A PSU is usually a lot cheaper to replace than your motherboard although not always. Oh and we would get much more expensive boards
12v has been standard in datacentres for years now. We haven't seen any issues because the 5v, 3.3v and -12V lines do so little these days that the regulators on the motherboard that supply those voltages never fail. They're just doing so little that they don't fail.
The same goes for laptops. You don't see failures in those regulators because they do next to nothing.
Even in desktops by Dell, HP, Lenovo etc have been 12v for years and again, it's a non-problem.
Not that any of that matters. ATX12VO is an update of the ATX12V spec. It's been around for decades and it has had zero effect on motherboard and PSU manufacturers, and there's no reason to believe that this is likely to change at any point in the near to mid future.
Taking into account the overall equipment roadmap, where most of the features are transferred from the industrial grade to DIY and than to retail products, no surprise that it's just a question of time when the common powerbus is getting its place in home applications. Most of the complex electric/electronic machines have a number of PCBs powered by a single voltage transformed by internal means into required range.
Ahhh yes.. motherboards don't cost enough right now.. let's add more power delivery complexity.
I can guarantee you 100% the price on power supplies will not drop despite the cost being moved to the board. If anything we've seen in the past 3 years of pc hardware industry scheming, this is a highly improbable future. The only thing going down in price these days is nvme storage which seems to have a mind of its own.
So I imagine that cases will need to change to handle all of the new cable management that will be needed for cables starting from the Mobo instead of the PSU.
Genuine journalism going on here here
Needs more algorithmic coverage