ASUS Unveils ROG Zenith II Extreme Alpha Motherboard: Improved CPU VRM

[btarunr]

ASUS updated its AMD socket sTRX4 motherboard series with the new ROG Zenith II Extreme Alpha, a slight step-up from the original ROG Zenith II Extreme that debuted with AMD's 3rd gen Ryzen Threadripper family. Although ASUS' entire sTRX4 motherboard lineup will support the upcoming 64-core Threadripper 3990X, the new Extreme Alpha is better designed for overclocking it. The new board looks visually identical to the original ROG Zenith II Extreme, and has an almost-identical feature-set, with the only difference being the CPU VRM solution. The new ROG Zenith II Extreme Alpha implements a 16-phase CPU VRM with Infineon TDA21490 power-stages replacing the TDA21472 power-stages on the original ROG Zenith II Extreme (possibly increase output current or I-out from 70 A to 90 A). This could marginally increase the product price. The rest of the feature-set is identical.



View at TechPowerUp Main Site

 

[bug]

In other words, on the original board they botched something that couldn't be fix unless physically altering the board.

 

[dj-electric]

In other words, on the original board they botched something that couldn't be fix unless physically altering the board.

Pretty much it. Why make a new board so close to first one just to say that they got right something they should have in the first place

 

[btarunr]

At least it's better than Gigabyte's revision clusterf**kery. Remember how the VRM setup would completely change from, say, 24-phase to 16-phase between Rev 1.0 and Rev 2.0 back in the X58 days? It was a nightmare ordering the revision you wanted online.

 

[bug]

At least it's better than Gigabyte's revision clusterf**kery. Remember how the VRM setup would completely change from, say, 24-phase to 16-phase between Rev 1.0 and Rev 2.0 back in the X58 days? It was a nightmare ordering the revision you wanted online.

Imho, if you're changing the implementation without adding new features, having a new revision makes perfect sense.

Vendors writing rubbish on their website is an entirely different problem ("Nvidia graphics 4GB VRAM" anyone?)

 

[kinjx11]

ASUS loves making Mobo so much they remaster their Mobos with better cooling every 6 months

 

[beedoo]

Primarily, had I have bought the RZIIE (again) and then discovered a RZIIE Alpha so close to the release of the original, I would have been really pissed.

Similar thing happened when I bought my ROG Zenith Extreme for my 2950X.

 

[Chloe Price]

At least it's better than Gigabyte's revision clusterf**kery. Remember how the VRM setup would completely change from, say, 24-phase to 16-phase between Rev 1.0 and Rev 2.0 back in the X58 days? It was a nightmare ordering the revision you wanted online.

Yep, one would think that a newer revision is an improved one, but Gigabyte did exactly the opposite.

 

[TheinsanegamerN]

Imho, if you're changing the implementation without adding new features, having a new revision makes perfect sense.

Vendors writing rubbish on their website is an entirely different problem ("Nvidia graphics 4GB VRAM" anyone?)

Except if you want a "gigabyte x58 walrus tooth" board, there were different revisions, all with different quality of circuitry, and NO way to tell. They were ALL sold as "gigabyte x58 walrus tooth" boards, not "gigabyte x58 walrus tooth 1.0", "gigabyte x58 walrus tooth 1.5 with worse VRM", and "gigabyte x58 walrus tooth 2.0 with shit VRM". That's why they should have been seperate products, like what asus is doing here.

 

[bug]

Except if you want a "gigabyte x58 walrus tooth" board, there were different revisions, all with different quality of circuitry, and NO way to tell. They were ALL sold as "gigabyte x58 walrus tooth" boards, not "gigabyte x58 walrus tooth 1.0", "gigabyte x58 walrus tooth 1.5 with worse VRM", and "gigabyte x58 walrus tooth 2.0 with shit VRM". That's why they should have been seperate products, like what asus is doing here.

No, no, no.
Gigabyte doesn't tell you anything about the VRM in the name of revision 1.0, why would they do that in subsequent revisions? They tell you it's a different revision, it's up to you to look up the differences.

 

[gamefoo21]

LoL Asus finding out their gimpy VRMs are garbage for the price. So they are pushing a revision with a different model, so if you warranty the old one you won't get the new one.

Asus being Asus.

I feel bad for any fools who bought that X399 board that had a VRM that overheated if you actually tried overclocking a 2990...

Their marketing full of lies...

The Asus fanboys having to go all the way back to X58 to go what about Gigabyte!

Heh

 

[bug]

LoL Asus finding out their gimpy VRMs are garbage for the price. So they are pushing a revision with a different model, so if you warranty the old one you won't get the new one.

More likely, everybody else was doing better in the overclocking department and they were forced to up the ante.
Doesn't make the original boards bad or anything.

 

[X4K4]

whats with the new more dimm sloots less PCIEXpress slots x16 approach to motherboards?

 

[kapone32]

LoL Asus finding out their gimpy VRMs are garbage for the price. So they are pushing a revision with a different model, so if you warranty the old one you won't get the new one.

Asus being Asus.

I feel bad for any fools who bought that X399 board that had a VRM that overheated if you actually tried overclocking a 2990...

Their marketing full of lies...

The Asus fanboys having to go all the way back to X58 to go what about Gigabyte!

Heh

The have the same old VRM on the X399 boards and (due to the 40mm fan) it had the coolest VRM temps. In fact Hardware unboxed showed that this was one of only 2 boards that would accept a 4.1 OC across all cores and actually boot. I believe this board is exactly what "btarunr" said for OC the 3990X which is a 64 core monster.

 

[AnarchoPrimitiv]

Does anyone else find it really annoying and logically false when someone tries to "defend" a company, or any entity, by ironically NOT defending them at all and instead resorts to just pointing their finger at some other entity they believe to be guilty of the same offense (or sometimes a completely different offense)?

For example, whatever gigabyte did or may have done, it is absolutely irrelevant and not a defense or justification whatsoever for whatever Asus has done. We've never, nor will we ever here the following defense in a court of law:

Defendant: "Your Honor, Yes I murdered that person. But I believe I should be found 'Not Guilty' due to the fact that there are numerous other individuals committing murder as well."

Judge: "What you've just said is one of the most insanely idiotic things I have ever heard. At no point in your rambling, incoherent response were you even close to anything that could be considered a rational thought. Everyone in this room is now dumber for having listened to it. I find you guilty, and may God have mercy on your soul."

Humor aside, whenever someone tries to defend the actions of someone or something by pointing at the actions of someone or something else, they're basically announcing that they're incapable of a legitimate defense or justification and I feel as though it's the duty of everyone in favor of logic observing it to give this ridiculous "defense" absolutely no quarter and to attack it whenever it manifests.... Which is mainly, for example, on sites like Wccftech where if an article discusses some alleged or proven underhanded behavior by Nvidia, its self-appointed defenders will immediately chime in with:

"... But... But... But 12 years ago Intel did [insert some completely irrelevant and unrelated example]..."
-or-
"... But Intel did [insert perceived offense] this too..."

P.S. I'm not a fanboy of any brand or company, I'm a fanboy of logic and legitimate debate, so whenever someone is tryng to utilize any "argument" that's either irrelevant or logically fallacious, I feel as though I have a compulsory duty to bring it to the attention of those involved

 

[Rage Set]

So now I can see why I had to get a replacement board from Newegg.com for my Zenith II Extreme and why I gave it the review that I did on Newegg. If Asus is willing to rerelease this board so soon, that means the VRM's are not adequate for overclocking (of which they do market this board for). The VRM's on my original Zenith II Extreme got toasted by a 3960X, I can only imagine what is going to happen when I get the 3990X on the replacement board. I think I'm done with Asus for good. I love their BIOS but that won't do it for me anymore.

 

[mouacyk]

gaming... motherboard

 

[kapone32]

gaming... motherboard

For $1100 (the cost of a decent system)

 

[cdawall]

Means the 3990X pulls more power than they originally expected. We saw the same issue with TR4 boards.

 

[Live OR Die]

In other words, on the original board they botched something that couldn't be fix unless physically altering the board.

Pretty much it. Why make a new board so close to first one just to say that they got right something they should have in the first place

Money grab thats all it is they have done the same thing on many X299 boards, Rampage VI Exteme > Rampage VI Omega > Rampage VI encore with each one they upgraded the power delivery lol.

 

[ypsylon]

Totally lame from Asus.

Gigabyte did it properly at first outing. Went with XL-ATX, stuffed Aorus Extreme with monstrously overkill VRM and it worked amazingly.

Asus thought, oh... we can cheap-out a bit on our flagship board... and then realized that VRM may not be up to snuff with 3990X/$. So now they basically manufacturing two identical boards with two different VRMs. Whoever is responsible for such wasteful practices should be fired at once. If they would redesign ZE2A with 3 x16 PCI express slots I would congratulate them. Just enough for 2 rendering VGAs and 1 quad NVMe card or 3 rendering VGAs. As things stands except Gigabyte (XL-ATX Aorus Extreme and Designare) and AsRock Taichi all other boards cannot plug 3 or 4 2-slot VGAs. I know it's rare, but these boards cost fortune and primarily are not designed as gaming platforms. Designers should think about this stuff and not tinker how to stuff RGB everywhere while compromising motherboard functionality.

 

[cdawall]

Totally lame from Asus.

Gigabyte did it properly at first outing. Went with XL-ATX, stuffed Aorus Extreme with monstrously overkill VRM and it worked amazingly.

Asus thought, oh... we can cheap-out a bit on our flagship board... and then realized that VRM may not be up to snuff with 3990X/$. So now they basically manufacturing two identical boards with two different VRMs. Whoever is responsible for such wasteful practices should be fired at once. If they would redesign ZE2A with 3 x16 PCI express slots I would congratulate them. Just enough for 2 rendering VGAs and 1 quad NVMe card or 3 rendering VGAs. As things stands except Gigabyte (XL-ATX Aorus Extreme and Designare) and AsRock Taichi all other boards cannot plug 3 or 4 2-slot VGAs. I know it's rare, but these boards cost fortune and primarily are not designed as gaming platforms. Designers should think about this stuff and not tinker how to stuff RGB everywhere while compromising motherboard functionality.

Many companies do this or ignore the issue and let the end user suffer. I would rather they release the updated board with the upgraded VRM, then just ignoring the issue, which is what basically everyone did except MSI with TR4 and the 2990WX.

 

[Totally]

Many companies do this or ignore the issue and let the end user suffer. I would rather they release the updated board with the upgraded VRM, then just ignoring the issue, which is what basically everyone did except MSI with TR4 and the 2990WX.

Going by the comments here one would think the correct course of action for ASUS would have been to ignore the issue or silently update the boards.

 

[player-x]

A lot of people are shouting, but have no clue what they are talking about!

In other words, on the original board they botched something that couldn't be fix unless physically altering the board.

Nope they changed the VRMs from previous 70A best ones to now 90A best ones, a drop-in replacement.

We talking about a $750+ mobo, with VRM that theoretical could deliver 1344W (TDA21472) and that now can do 1728W (TDA21490).
(16 x 90A = 1440 ampere X 1.2V = 1728W)

And yes 1728W is is pure theoretical.

As when you run Prime95 @ 4GHz this would be what the old VRMs dissipate on heat, these new ones would be marginal better, but also properly more stable and/or faster, and better is better, so why not change.

• 200A = 16W
• 300A = 24W = 3960X
• 400A = 32W = 3970X
• 600A = 59W = 3980X
• 800A = 96W = 3990X (*)

(*)With 800A it runs out of spec, the motherboard power connections of 24 (180W) + 2x 8 pin (2x 235W), but the old and new VRMs could handle it easily.

Instead of taking what ASUS is doing as positive, everyone here thinks they screwed up and are fixing it now.