Intel Encourages Adoption of ATX12VO Standard on Alder Lake-S Motherboards

[Zubasa]

Motherboards already have dedicated VRMs for tighter regulation and ripple suppression than what PSU can do.

This might increase requirements for their designs (and hence cost to some extent), but that's about it.

The CPU VRMs do, and those are the most expensive power components as a result.
Most of the auxillary voltages on mobos are regulated by cheap buck converts with little regard to efficiency / ripple suppression etc.

 

[Frank_100]

for a saving of few watts on idle &low power load i don't know if is worth it

the mobo will be hotter as it will convert high amp 12v to high amp 5v and all other voltages.....cooling it will be more complex and what you save with psu (if will be cheaper?) you will spend on mobo and connectors..

is more like re-inventing the wheel by making it more complex but in the end with the same functionality and costs...

This ^^^^^^^

 

[ExcuseMeWtf]

Bulk converters? I presume a typo and you meant buck?

That changes nothing actually in the point. Otherwise, what exactly is stopping them from redesigning converters to work at 12V input voltage instead of currently used ones? It's not like components specc'd to work @12V don't exist, is it?

 

[Zubasa]

That changes nothing actually in the point. Otherwise, what exactly is stopping them from redesigning converters to work at 12V input voltage instead of currently used ones? It's not like components specc'd to work @12V don't exist, is it?

So if it is doing nothing, why reinvent the wheel?

 

[Bomby569]

More e-waste for the sake of a small power savings is not smart in terms of environment. This industry should try and reduce e-waste not make more.

 

[Caring1]

I don't get it, increased heat on the Motherboard will see a short return of passive boards as they return to active cooling to cope.

 

[ExcuseMeWtf]

So if it is doing nothing, why reinvent the wheel?

It clearly doesn't "do nothing", that's the whole point, at the very least it's simplifying PSU design and cutting away a lot of ancient junk from the market (cheapo group regulated nonsense won't exist with single output by definition).

It's not changing that parts of PSU are not being moved onto the mobo, as, again, like VRMs mentioned before, those buck converters were already there to begin with.

 

[Zubasa]

It clearly doesn't "do nothing", that's the whole point, at the very least it's simplifying PSU design and cutting away a lot of ancient junk from the market (cheapo group regulated nonsense won't exist with single output by definition).

It's not changing that parts of PSU are not being moved onto the mobo, as, again, like VRMs mentioned before, those buck converters were already there to begin with.

So whats your point? Either it changes nothing or it changes everything.
As for ancient junk, most motherboards out there reuse VRM / trace layout even across generations to cut costs.
On cheaper motherboard some of those layout / componenets have been used for over a decade.
So what exactly stops them from keep doing what they have been doing?

How is it not moving parts of the PSU on to the motherboard? You either do the conversion from 12V to 5V and 3.3V in the PSU, or you do it on the motherboard as is shown.
Those SATA power OUT connectors on the Mobo are there to provide the SATA drives with the voltages they need. You just end up powering them from the motherboard instead of the PSU.
Even if you toss out all the "ancient junk" SATA drives like I did, USB Type A and M.2 still require 5V and 3.3V.

 

[ZoneDymo]

Then you are not understanding the technical aspect of these changes. They reduce the complexity of PSU construction and thus cost, increase PSU reliability and reduce power consumption.


You have that backwards. PSU's will be LESS expensive and motherboards might increase in cost to a minor degree.

This is an all win situation for everyone. Intel has no stake in the design, it's completely an open standard.


I hope that didn't seem like an attack. I only wanted to help with the understanding that this is a very positive change for the industry. It's a change this is both logical from a design perspective, but also a cost perspective. More efficient and less expensive equals a win for everyone.

1. what has my believe that this is not a bad time to release a new standard have to do with an understanding of the technical aspect of the changes?

2. Yes I did have that backwards, but how did you get to this being an all win situation for everyone? as was said, more expensive motherboards with every upgrade (and in intel's case, that is about every time you want to upgrade) just so save a few cents on power consumption?

 

[R0H1T]

What about existing components? I definitely am not buying a 12vo PSU if it doesn't work with other (older) systems I have, even if you need a spare PSU for a day to two why the eff would I spend on something that'll not work with 90-95% of existing hardware! This is of course assuming they will not work. Same goes for mobos as well ~ they have to work with existing PSU's otherwise I'm not buying them just because Intel has a new (useless) bullet point against AMD.

 

[ExcuseMeWtf]

So whats your point? Either it changes nothing or it changes everything.

Do you really not understand or are you just acting in bad faith now? I thought the point was pretty clearly stated...

As for ancient junk, most motherboards out there reuse VRM / trace layout even across generations to cut costs.

I was talking about ancient group regulated PSUs which by definition won't be compatible with ATX12VO, so they will go the way of Dodo. Better late than never.

How is it not moving parts of the PSU on to the motherboard? You either do the conversion from 12V to 5V and 3.3V in the PSU, or you do it on the motherboard as is shown.

Does, say, your RAM run at 3.3V or 5V? It doesn't on even rather old PCs by now.

PSU is not providing that actual voltage, whichever it currently is for you. How do you think that voltage is derived?

So if those regulators are not fed with 12V currently, why are you so dogmatically opposed to them being fed with 12V instead? A few cents per mobo extra for higher spec'd components? As opposed to significantly higher savings in PSU?

Those SATA power OUT connectors on the Mobo are there to provide the SATA drives with the voltages they need. You just end up powering them from the motherboard instead of the PSU.
Even if you toss out all the "ancient junk" SATA drives like I did. USB and M.2 still require 5V and 3.3V.

Which likely have additional VRMs anyways, as vreg/ripple of PSU itself may not be tight enough to feed components directly with PSU output.

 

[Zubasa]

Do you really not understand or are you just acting in bad faith now? I thought the point was pretty clearly stated...

Does, say, your RAM run at 3.3V or 5V? It doesn't on even rather old PCs by now.

PSU is not providing that actual voltage, whichever it currently is for you. How do you think that voltage is derived?

So if those regulators are not fed with 12V currently, why are you so dogmatically opposed to them being fed with 12V instead? A few cents per mobo extra for higher spec'd components? As opposed to significantly higher savings in PSU?

USB 5V does run on 5V, and M.2 does just run on 3.3V or 5V supply depending on the keying.
These do not have their dedicated VRMs and is on very modern motherboard out there. So you will need to add extra components for them.
So shifting the goal post to RAM doesn't make sense.

The whole point of this 12VO standard is supposedly for efficiency. 12V in theory means it runs less current for a given power.
To minimized power lost on the power plane you want to do the 5V / 3.3V voltage conversion as close to what you are powering as possible.
Ideally you want have somekind of VRM for each of those M.2 slots etc.

All that is skipped is the sight power lost from the PSU cable, and that is insignificant compare to the motherboard traces' resistance.
What you end up doing is the extra "PSU" is just on the edge of the motherboard running through the higher resistance traces to the components you are powering.

Which likely have additional VRMs anyways, as vreg/ripple of PSU itself may not be tight enough to feed components directly with PSU output.

NVME and SATA drives are designed to be fed directly form the PSU and have quite a lenient range for voltage input.
Every PSU worth a damn have tighter voltage regulation than required.

 

[The red spirit]

For motherboard, especially on Intel platform where people have to change board every 2 gen, increasing the cost of motherboard is just moronic.
And the difference in power saving at idle is only a few watts even if efficiency improve by 20%

Who really replaces boards every 2 gens? Most people don't replace boards for 5+ gens, for enthusiast there's hardly any benefit either. If you bought i7 26000k with board, then upgraded to coffee lake, there's almost no benefit to upgrade it to Comet Lake or Zen 3, unless personal workloads scale with cores very well and you absolutely need all that performance, but then again if you are this kind of user, getting HEDT platform makes far more sense.

And the difference during idle can be far more than 20%. That's because on lower power processors like i5 10400F, idle power consumption is really low and it inevitably translates into massive PSU loss of efficiency. PSU may be only 50% efficient at such load. Here's a video about those power supplies and power savings:

Those gains to me seem like no brainer to switch the standard, it's just unfortunate that lots of hardware will be rendered obsolete this way (unless you can connect power cable to motherboard).

 

[Bomby569]

I'm sorry but there are a lot of false arguments around here. This will do nothing to improve power delivery. Unless you are talking about the cheapest of cheapest PSU's, any PSU can safely deliver power to your PC today. In the other hand this will probably create a big problem with motherboards, all but the most expensives will probably be dealing with a lot of problems especially heat. That or prices will increase across the board.

 

[ZoneDymo]

I'm sorry but there are a lot of false arguments around here. This will do nothing to improve power delivery. Unless you are talking about the cheapest of cheapest PSU's, any PSU can safely deliver power to your PC today. In the other hand this will probably create a big problem with motherboards, all but the most expensives will probably be dealing with a lot of problems especially heat. That or prices will increase across the board.

can we finally get back to these majestic beings:

 

[Zubasa]

And the difference during idle can be far more than 20%. That's because on lower power processors like i5 10400F, idle power consumption is really low and it inevitably translates into massive PSU loss of efficiency. PSU may be only 50% efficient at such load. Here's a video about those power supplies and power savings:

Those gains to me seem like no brainer to switch the standard, it's just unfortunate that lots of hardware will be rendered obsolete this way (unless you can connect power cable to motherboard).

A quick look at reviews will tell you that is not the case in 2021.
https://www.techpowerup.com/review/be-quiet-pure-power-11-fm-650-w/5.html

As for dodgy PSUs, there are plenty of dodgy off-brand mobos out there as well.

 

[TheLostSwede]

So far this seems like a solution to a problem that doesn't really exist.
All this does is moving parts of the PSU on to the motherboard.
Now instead of PSU makers having to achieve higher efficiency, the ball is just passed down to the Motherboard.

5 and 3.3V conversion would be more efficient in the motherboard, as it would happen closer to the parts that need those voltages. The only downside would be for SATA drives.
However, it would increase the cost of motherboards more than it's likely to reduce the cost of PSUs, as neither side is going to be willing to lower their earnings percentage.

 

[Zubasa]

5 and 3.3V conversion would be more efficient in the motherboard, as it would happen closer to the parts that need those voltages. The only downside would be for SATA drives.
However, it would increase the cost of motherboards more than it's likely to reduce the cost of PSUs, as neither side is going to be willing to lower their earnings percentage.

I am aware of that as stated in my later post.
The real problem here, is the difference even easily measurable?
How much power lost is happening in the 18-guage PSU cable on the low power rails?

Also as you said, why would motherboard makers care to make their power conversion as efficient as in modern PSUs?
Those rails needs to be there but isn't easily marketable to incentize making them half decent.
PSU on the other hand are always tested in reviews as it is the PSU's only job.

 

[TheLostSwede]

I am aware of that as stated in my later post.
The real problem here, is the difference even easily measurable?
How much power lost is happening in the 18-guage PSU cable on the low power rails?

Also as you said, why would motherboard makers care to make their power conversion as efficient as in modern PSUs?
Those rails needs to be there but isn't easily marketable to incentize making them half decent.
PSU on the other hand are always tested in reviews as it is the PSU's only job.

Well, they'd obviously have to meet some kind of standard, but as you point out, how will that be tested and it's unlikely to be a race to offer the best conversion, as it's simply not a feature that most people are going to give a rat's ass about.

 

[The red spirit]

A quick look at reviews will tell you that is not the case in 2021.
https://www.techpowerup.com/review/be-quiet-pure-power-11-fm-650-w/5.html
View attachment 200627View attachment 200630

Well, my system spends quite a bit of time with 5-6% of PSU load and my power meter, shows really low power factor then. Closer to 60s. Anyway there's one thing, Linus wouldn't have seen any big gains if new standard wasn't more efficient and yet at idle he saw 50% improvement. He certainly doesn't use some cheapo 80+ unit which is worth 40 dollar and power supply industry is honestly one of the slowest in terms of actually changing anything. So I'm not sure why that you posted here seemingly contradicts other sources and I doubt that current power supplies really all that good, but if they are then your linked test still shows that such efficiency will be seen only with very expensive power supplies, that most people won't buy and wouldn't really know about any of their efficiency improvements on very low loads. If new standard can make that much cheaper, then it makes sense to implement it. Motherboards already have VRMs, so for them increased cost will be much smaller than adding more stuff to PSU. In terms of affordability of power supplies, I would dare to say that even 80+ gold models are somewhat unaffordable for lots of computer builders and most of them will end up with 80+ bronze units. Anything higher than gold is straight up too expensive or availability of such units is poor. And I remember that some site did some calculation and found out that there's no financial incentive to get anything more than 80+ bronze rated power supply, because what you spend on it will never pay off in lower electricity expenses. If new standard could mean way more efficient PSUs for 50 dollars than today's 150 dollar units, new standard does make sense.

 

[lexluthermiester]

it's just unfortunate that lots of hardware will be rendered obsolete this way (unless you can connect power cable to motherboard).

There will of course be adapters.

can we finally get back to these majestic beings:

That would be nice. Bizarre looking but undeniably beautiful works of art.

 

[Rais]

So we have confirmed we are basically moving components from PSU to motherboard.
Still no one can explain the technical benefit of this, accounting all the waste this process can make and the reengineering

 

[lexluthermiester]

So we have confirmed we are basically moving components from PSU to motherboard.

Effectively. But in doing so, there is a cost savings on the PSU side and better power management on the motherboard side.

 

[TheinsanegamerN]

So this has a new 10 pin connector, then a 6 pin connector, then two 4 pin connectors of a non standard size.

That adds up to 24 pins. How is this any better then the old 24 pin standard?

 

[lexluthermiester]

So this has a new 10 pin connector, then a 6 pin connector, then two 4 pin connectors of a non standard size.

That adds up to 24 pins. How is this any better then the old 24 pin standard?

10pin id for the mobo and what is connected to it, the 6pin is for the CPU and the 4pin is for components.. IIRC..