ASRock B450 Gaming-ITX/ac

[cadaveca]

ASRock has applied the Fatal1ty treatment to AMD's B450 chipset, and in a way only ASRock can; using the Mini-ITX format. It's not often that we get such a truly gaming-oriented mITX product that can also handle some overclocking while supporting 3466 MHz memory speeds.

Show full review

 

[Ferrum Master]

Where are the mATX boards mate?

 

[Aderbas]

thumbs down: "NO RGB", "It's small", "Only 2 DIMM"... WTF?!

 

[cadaveca]

thumbs down: "NO RGB", "It's small", "Only 2 DIMM"... WTF?!

If you have suggestions for another fault this board might have, I'm listening.

No RGB is a fault since nearly everything is RGB these days, making the lack of onboard RGB a lack of a feature, especially given that the board supports addressing 5050 LED strips (and memory sticks) right within its BIOS. The RGB control is great too.. there's no conflict between BIOS settings and software settings.. change it in one, and the other retains the changes for you. That means you don't need the RGB software running all the time to get the lights to light up the way you like, unlike many other board products. With RGB things so good, it sure is a shame that they didn't include a bank or two on the board itself!

 

[wrathchild_67]

Are the wattage figures from this page measured at the wall? If so, 7w at idle for the whole system is incredible. I've only ever managed about 19w idle with a 35w TDP Intel CPU.

 

[cadaveca]

Are the wattage figures from this page measured at the wall? If so, 7w at idle for the whole system is incredible. I've only ever managed about 19w idle with a 35w TDP Intel CPU.

No, this is not at the wall. I'm measuring power through the 8-pin power plug using a clamp-on meter. That's why the cooling is also talked about on the same page, as that board power cooler is cooling that load that is mentioned in the power draw listings.

 

[wrathchild_67]

No, this is not at the wall. I'm measuring power through the 8-pin power plug using a clamp-on meter. That's why the cooling is also talked about on the same page, as that board power cooler is cooling that load that is mentioned in the power draw listings.

OK, so then idle power consumption for the whole platform is probably unchanged from the previous series motherboards then. I'll assume power draw at idle from the wall was at least 20-25w.

 

[HTC]

If you have suggestions for another fault this board might have, I'm listening.

No RGB is a fault since nearly everything is RGB these days, making the lack of onboard RGB a lack of a feature, especially given that the board supports addressing 5050 LED strips (and memory sticks) right within its BIOS. The RGB control is great too.. there's no conflict between BIOS settings and software settings.. change it in one, and the other retains the changes for you. That means you don't need the RGB software running all the time to get the lights to light up the way you like, unlike many other board products. With RGB things so good, it sure is a shame that they didn't include a bank or two on the board itself!

This doesn't seem to be a fault, IMO. It's a plus if it has it but not having it doesn't make it a fault.

My 2 cents.

 

[cadaveca]

OK, so then idle power consumption for the whole platform is probably unchanged from the previous series motherboards then. I'll assume power draw at idle from the wall was at least 20-25w.

For the whole system from the wall, no it is not that much, actually. The resolution of the meter used will give you completely different pictures here, too. I have no care for power use of memory, CPU, or other devices, PSU efficiency, etc. I am also not testing this platform. These are specific metrics looking at the board's CPU VRM design only. I would like to expand this testing, actually, but the way I want to requires a bit more investment in hardware that is not possible at this time. It's a struggle.

In other words, the numbers I have given are specifically NOT looking at what you are asking about here. I'm not reviewing the platform or AMD's design at all.

This doesn't seem to be a fault, IMO. It's a plus if it has it but not having it doesn't make it a fault.

My 2 cents.

1000% agreed. Aesthetics will always be a point of contention. But can I ask, would you mind if the default was to have such abilities "OFF", so you never see them unless you want to enable them, like a hidden feature?

 

[HTC]

For the whole system from the wall, no it is not that much, actually. The resolution of the meter used will give you completely different pictures here.


1000% agreed. Aesthetics will always be a point of contention. But can I ask, would you mind if the default was to have such abilities "OFF", so you never see them unless you want to enable them, like a hidden feature?

If it came with them, my personal preference would be for it to be OFF by default with enabling via BIOS and / or software, but it could be the other way around, with a way in BIOS and / or software to disable it.

If i buy a piece of PC hardware that has RGB, i treat it as an extra, but i didn't buy it specifically because of the RGB. I actually tend to avoid those because they usually cost more VS non-RGB version.

 

[AltCapwn]

Nice review @cadaveca as always and very very nice ITX board.

I imagine ryzen 5 in that thing, in an itx case with watercooling and a r9 580 8gb, I gotta start putting some money aside for my new build!

I'm kind of happy it's now possible to have a HEDT-like computer without having to spend thousands of thousands nowadays. Thanks AMD for that by the way.

 

[Sandbo]

I literally bought the X470 version few weeks ago. Seems like I could have got this.

 

[Fx]

• No onboard RGB lighting

This is a pro! Thanks for the review.

 

[Disparia]

I literally bought the X470 version few weeks ago. Seems like I could have got this.

I had the X470 in my Buy Later cart, it just became the B450 version. It seems like the only difference is the WiFi included.

The B450 also fixes spelling:

 

[Sandbo]

I had the X470 in my Buy Later cart, it just became the B450 version. It seems like the only difference is the WiFi included.

View attachment 104669

The B450 also fixes spelling:

View attachment 104670

Yeah, the "no difference than price" thing is kind of annoying LOL
I am connecting my new PC wired anyway.

Though it is within the return windows, I wouldn't disassemble it now.

 

[GorbazTheDragon]

"CPU Power: 8 phase"
Please do not help manufacturers mislead consumers, this board has a 3 phase vCore VRM with paired components on each phase along with a 2 phase VRM for the SOC. The MSi board is also not a 9 phase VRM.

I think that for the sake of actually testing boards properly you should additionally bench them with the 8 core chips at reasonable overclocks. It would also be nice to know where the temperature measurements came from.

I know plenty of people that would like (or have wanted) to stick the 6 and 8 core ryzens in B350/450 boards and you have to tread your ground really carefully because a lot of the boards don't hold up to even the normal boost clocks of the 2700x

 

[cadaveca]

"CPU Power: 8 phase"
Please do not help manufacturers mislead consumers, this board has a 3 phase vCore VRM with paired components on each phase along with a 2 phase VRM for the SOC. The MSi board is also not a 9 phase VRM.

Sigh. You act like having an input driver push more than a singular output is something new (some are purposefully designed for it, even). I used to do those investigations (look at my reviews from 5+ years ago), and found them useless overall. The power testing we do easily sorts out these types of issues, as we know exactly how much power the CPU VRM is pulling under load. Gamers who would buy this board won't be pushing it enough for such to even matter (My APU pulls a lot of power when using iGPU, and from the problematic SOC region that seems to be a problem with most boards.). There is also the issue of boards on the AMD side of things being a little bit more sensitive to PSUs, too, something I don't see anyone who is mentioning all of this looking at, nor are they looking at power use over the 8-pin in general.

As to using a 95W or 105W CPU being relevant? you bet. You got one I can have? I mentioned in the review I used the one I did because it was cheap. AMD isn't exactly handing out CPUs in multiples to the same site, so I paid for this chip myself so I could do this review. You're talking about spending hundred's of dollars, so please, when you make such suggestions, offer some cash with it too, OK?

You can rest assured, however, that we do look at such problems, and identify them when present. You can see such an analysis done here:

https://www.techpowerup.com/reviews/Biostar/Z370GT6/13.html

Sometimes, a BIOS update can fix these problems.

 

[Tsukiyomi91]

a decent mITX, but the lack of supporting an M.2 PCIe or SATA driven SSD is kinda letdown... =/

 

[Sandbo]

a decent mITX, but the lack of supporting an M.2 PCIe or SATA driven SSD is kinda letdown... =/

It has, from the first pages spec.

 

[Voluman]

Thanks for the review, but actually not a single energy consumption or memory handling test? Or there is no point , is every similar product inside margin?

And for system specs, graphic: Vega 8, software NV Geforce driver, are you sure?

 

[cadaveca]

a decent mITX, but the lack of supporting an M.2 PCIe or SATA driven SSD is kinda letdown... =/

The M.2 port is on the back of the board. It's listed in the specs and shown in the pictures...

Thanks for the review, but actually not a single energy consumption or memory handling test? Or there is no point , is every similar product inside margin?

And for system specs, graphic: Vega 8, software NV Geforce driver, are you sure?

memory OC is shown in the CPU-z screen, and I talked about it. Power usage and how the cooling holds up has its own page. DId you actually read the review, or just skim to the end (which many people do)? I mean, you saw I listed wrong driver, so how come not the other stuff?

NVidia driver lsiting, yep, that's an oversight. I used adrenaline 17.8.2. fixed.

 

[Voluman]

Power usage and how the cooling holds up has its own page. DId you actually read the review, or just skim to the end (which many people do)?

Ah, i saw it that just was too short, im not really memorizes or 'get it in my mind'. You are right, it is there, just need to focus veeeery

 

[GorbazTheDragon]

If the actual details of the phase count and the fets are irrelevant the phase count itself is irrelevant too because there is such a vast difference between the designs used. Z77 and Z87 were not particularly troublesome VRM wise because at the time manufacturers were massively overbuilding their boards.

2700Xs don't really seem to keep to the TDP very much out of the box because of the boost clocking. Guess it's probably related to the motherboards like it was on the intel side with the out of box overclocking stuff? But XFR2 seems to do almost all the overclocking work anyway. The B350/450 boards do support overclocking so it should be an objective to account for that in the review.

Whether you guys are capable of delivering that at any specific time is beyond my control, all I can say is that it is something that should be covered if possible.

As for the cooling/overclocking stuff would it be a stretch to ask for some graphical comparisons? I know there aren't many boards tested yet, but I think it would be a nice addition, the table doesn't exactly draw the readers attention.

For example the power consumption is done very well in the GPU reviews, I think something similar could be done for both power and VRM temps for motherboards.

 

[Sandbo]

If the actual details of the phase count and the fets are irrelevant the phase count itself is irrelevant too because there is such a vast difference between the designs used. Z77 and Z87 were not particularly troublesome VRM wise because at the time manufacturers were massively overbuilding their boards.

2700Xs don't really seem to keep to the TDP very much out of the box because of the boost clocking. Guess it's probably related to the motherboards like it was on the intel side with the out of box overclocking stuff? But XFR2 seems to do almost all the overclocking work anyway. The B350/450 boards do support overclocking so it should be an objective to account for that in the review.

Whether you guys are capable of delivering that at any specific time is beyond my control, all I can say is that it is something that should be covered if possible.

As for the cooling/overclocking stuff would it be a stretch to ask for some graphical comparisons? I know there aren't many boards tested yet, but I think it would be a nice addition, the table doesn't exactly draw the readers attention.

For example the power consumption is done very well in the GPU reviews, I think something similar could be done for both power and VRM temps for motherboards.

Not trying to judge TPU's review as I think it has already addressed most questions of potential buyers, but as a comparison, when I am about to buy a motherboard I often refer to the tests done by Tweaktown: (for example)
https://www.tweaktown.com/reviews/8487/asrock-x370-gaming-itx-ac-amd-motherboard/index3.html

I am not an expert in the electrical hardware, but I am often impressed by their amount of details.
Like thermal imaging and descriptions over the VRMs. Maybe something similar can be done if the cost allows? Just a suggestion.

 

[cadaveca]

If the actual details of the phase count and the fets are irrelevant the phase count itself is irrelevant too because there is such a vast difference between the designs used. Z77 and Z87 were not particularly troublesome VRM wise because at the time manufacturers were massively overbuilding their boards.

Or did vrm phases lose relevance because the vrm moved to the CPU?

I do not see that many strong boards out there to be honest. You notice we don't review every board, but we used to... I think OEMs learned what to not send me. I tend to be sent very good boards with little to no actual issues, and if there are issues, the board is generally replaced.

SO for me, yeah, phase count is rather irrelevant if the voltage supplied is good. You see, loadline calibration and such can fix some issues anyway, so I might look a this stuff differently (as is usual). VRM efficiency is important though, so we look at each board with the same CPU, and compare what it does, since we know what the CPU requires.

To me, phase begins with MOSFETs. There are very few phase controllers that even do 6 phases, so nearly every board has some sort of "Phase doubling". Back in 2010 and 2011 my reviews looked at this a lot. look at the Z77 reviews, for example; you see I mention things like "4 phases per input driver is the minimum number of active phases" because of how this all works.

So, some FETs have input drive inside them (DRMOS)

some are dual channel
some are single channel

PHases to me are high-side and low-side mosfet, choke, and capacitor/filter. so we can have a board with a controller in 3+2 mode, and still have 8 phases, because there are two phases per input channel on the "3" side. the "2" side is single phases without doubling up on input driver.

2700Xs don't really seem to keep to the TDP very much out of the box because of the boost clocking. Guess it's probably related to the motherboards like it was on the intel side with the out of box overclocking stuff? But XFR2 seems to do almost all the overclocking work anyway. The B350/450 boards do support overclocking so it should be an objective to account for that in the review.

Right, but what power limit are users with traditional cooling going to have as a limit to begin with? This is a pretty specific number that is easy to test for.

Whether you guys are capable of delivering that at any specific time is beyond my control, all I can say is that it is something that should be covered if possible.

I did cover it, and now find it irrelevant. Our power testing of the VRM already manages to find issues that might be caused by VRM design anyway, and usually the difference between board in these numbers is small enough that most users will not care. Those that are critical of electrical aspects rather than compute stuff might be, but there are many different users types we get to write for.

As for the cooling/overclocking stuff would it be a stretch to ask for some graphical comparisons? I know there aren't many boards tested yet, but I think it would be a nice addition, the table doesn't exactly draw the readers attention.

For example the power consumption is done very well in the GPU reviews, I think something similar could be done for both power and VRM temps for motherboards.

We will discuss within our motherboard team for sure this one. Like I said, our power testing does expose board VRM efficiency, but maybe if we showed this data ina different way you'd see it easier.

Not trying to judge TPU's review as I think it has already addressed most questions of potential buyers, but as a comparison, when I am about to buy a motherboard I often refer to the tests done by Tweaktown: (for example)
https://www.tweaktown.com/reviews/8487/asrock-x370-gaming-itx-ac-amd-motherboard/index3.html

I am not an expert in the electrical hardware, but I am often impressed by their amount of details.
Like thermal imaging and descriptions over the VRMs. Maybe something similar can be done if the cost allows? Just a suggestion.

StevenB and I used to talk a lot about board VRM designs and such back what is now like 7-8 years ago. he and I covered much of the same stuff, so his reviews covering this stuff did influence my choice of how my own reviews are presented. I know he knows his stuff, and he and I agree on so much when it comes to PC hardware that pretty much anything he's ever said I've repeated at some point. Although I'm sure there some things we don't agree on. After doing board reviews for years we've analyzed which pages people look at and what info they want the most too, so over the years my reviews have evolved a lot into what they are today. I take a non-competitive approach to a lot of things that is quite different from most others.