Motherboard Asrock B550 Phantom Gaming ITX/AX

[orekart]

Did you locate the "Post Status Checker"? I can't find it even though it is bandied about on the ASRock website product page for the B550 Phantom Gaming ITX/ax

As an update it appears ASRock's product page has been updated and references to Post Status Checker LEDs is removed.

For posterity here is the archived version from time of product release:
https://web.archive.org/web/20200606131820/https://www.asrock.com/mb/AMD/B550%20Phantom%20Gaming-ITXax/index.asp

 

[gobenji]

KitGuru has a review out. Mostly good, but there seem to be some erroneous sensor readouts (power and voltage) causing the CPU to run at higher voltages than necessary and thus draw more power than it should. Hopefully there'll be a BIOS update to fix this. Other than that, good marks all around.

ASRock B550 Phantom Gaming-ITX/AX Review - KitGuru

B550 and mini-ITX can go hand-in-hand for AM4 buyers, so does the ASRock B550 Phantom Gaming-ITX/AX

www.kitguru.net

I was also concerned about that erroneous sensor readout. I believe this is the issue described in detail in this GN video:

Around 13:41, they explain that ASRock has used the same trick on another motherboard before. IIUC, it causes the reported power usage to be lower than the actual power usage. This means the automatic CPU frequency adjustment algorithm (PB2) may run the CPU beyond the spec TDP (which is just a guideline, not an absolute limit) because it "thinks" that the power usage limit hasn't been reached when in fact it has. There are other limits (like temperature) that will prevent the CPU from actually burning out.

Is there a setting in the bios/Ryzen Master/whatever to lower the allowed power limit and avoid the higher heat and electricity bills that this false power reporting leads to?

 

[spiral]

Is this a ASRock exclusive ? According to HWinfo i a "Power reporting Deviaton" of 102% on a Gigabyte X570.

 

[gobenji]

Is this a ASRock exclusive ? According to HWinfo i a "Power reporting Deviaton" of 102% on a Gigabyte X570.

100% is the target if reported power == expected power at full load, so a value around 100% is good. According to the original HWinfo post, a deviation of ±5% around that is acceptable (so 95-105%); beyond that it becomes suspicious.
According to the kitguru review, on the ASRock B550 Phantom Gaming ITX/ax they saw a value of 45%!

These inaccuracies are also highlighted by a dreadful power reporting percentage of around 45% in HWiNFO when under stress test load

In the GN video, they report a value of 32% on an ASRock X570 Taichi and 95% on a Gigabyte X570 Master.
Going back to the HWinfo thread, post 1 says:

Since at least two of the largest motherboard manufacturers, still insist on using this exploit to gain an advantage over their competitors despite being constantly asked and told not to, we thought it would be only fair to allow the consumers to see if their boards are doing something they're not supposed to do.

The author doesn't say which are the two manufacturers but it's safe to assume one of them is ASRock.

This followup (and the link) is really interesting. Basically, it looks like ASRock decided to hide the data with a bios update after that HWinfo feature was added and people noticed the problem. That's for a X570 Steel Legend mobo but we can see that there was a bios update for the B550 mini itx with a similar description:
https://www.asrock.com/mb/AMD/B550 Phantom Gaming-ITXax/index.asp#BIOS
There was also another update after that. At this point, we need someone to do actual tests with the updated bioses.

The annoying thing about that fake power reporting data is that it only provides a marginal performance gain (~1-2%) at the cost of a substantially increased power consumption (~20-30% total system power). The GN video arrives at a similar conclusion. The other annoying thing is that it's hidden from the user - ie. you didn't enable any OC.

The difference compared to overclocking or using AMD PBO, is that this is done completely clandestine and that in the past, there has been no way for most of the end-users to detect it, or react to it.

 

[spiral]

This deviation peaked my interest, so i've tested a bit more. After 12 hours under high load it says an average of 81% (low 48%)

 

[ellroy80]

This power issue is extremely disappointing. I'm planning to build a quiet system in a T1 with a 65W CPU, and was looking to use this mobo. That might not be possible if the power readouts are not accurate, as more power = more heat = more noise.

 

[Mackan]

This power issue is extremely disappointing. I'm planning to build a quiet system in a T1 with a 65W CPU, and was looking to use this mobo. That might not be possible if the power readouts are not accurate, as more power = more heat = more noise.

I've also stayed away from AsRock due to this. But I guess they don't care.

 

[BeerNsoup]

I watched the optimum tech b550 round-up and he said if you plan to set your own clocks/voltages it probably wouldn't be an issue. That may have been before the bios update though...

If they are hiding the readings does that mean that even if you customize your settings it may still run hotter? My upcoming ryzen 4000 (or 5000 apparently?) build will be my first time messing around with this stuff.

I was orignally leaning towards this board because it's fair bit cheaper than the asus boards in Canada. I've ruled out the gigabyte/msi models because of the higher socket position. If this board's sketchy power draw rules it out that leaves me with just the asus boards which are really quite expensive. The asus b550 is only like 20-30 less than the x570 and seriously lacks in usb A support (mouse, keyboard, headset, webcam leaves no free slots).

Other possibly not valid concerns I have with the asus boards that maybe someone could clear up are regarding the raised sound card/m.2 section. I'll be using an ak120 mini cooler which uses 120mm fans that are pretty close to the board (likely wouldn't clear the top of the m.2 section). Is the m.2/sound card bit at least 60cm from the centre of the socket? Also some of the new GPUs have really thick back plates, which is pretty cool, but the proximity of the m.2/sound card to the pci-e slot maybe would be incompatible with thicker backplates?

Thanks for any help.

 

[ellroy80]

I watched the optimum tech b550 round-up and he said if you plan to set your own clocks/voltages it probably wouldn't be an issue. That may have been before the bios update though...

If they are hiding the readings does that mean that even if you customize your settings it may still run hotter? My upcoming ryzen 4000 (or 5000 apparently?) build will be my first time messing around with this stuff.

I was orignally leaning towards this board because it's fair bit cheaper than the asus boards in Canada. I've ruled out the gigabyte/msi models because of the higher socket position. If this board's sketchy power draw rules it out that leaves me with just the asus boards which are really quite expensive. The asus b550 is only like 20-30 less than the x570 and seriously lacks in usb A support (mouse, keyboard, headset, webcam leaves no free slots).

Other possibly not valid concerns I have with the asus boards that maybe someone could clear up are regarding the raised sound card/m.2 section. I'll be using an ak120 mini cooler which uses 120mm fans that are pretty close to the board (likely wouldn't clear the top of the m.2 section). Is the m.2/sound card bit at least 60cm from the centre of the socket? Also some of the new GPUs have really thick back plates, which is pretty cool, but the proximity of the m.2/sound card to the pci-e slot maybe would be incompatible with thicker backplates?

Thanks for any help.

I watched the Gamers Nexus video after my comment and they also said it wouldn't be an issue if you set things manually. EDIT: I haven't read how this is affected by the latest BIOS update.

 

[gobenji]

I was orignally leaning towards this board because it's fair bit cheaper than the asus boards in Canada. I've ruled out the gigabyte/msi models because of the higher socket position. If this board's sketchy power draw rules it out that leaves me with just the asus boards which are really quite expensive. The asus b550 is only like 20-30 less than the x570 and seriously lacks in usb A support (mouse, keyboard, headset, webcam leaves no free slots).

After investigating this power issue with the ASRock B550, I'm now leaning towards the Gigabyte X570 I AORUS PRO WIFI instead. Although it's an X570 board, it's cheaper than the ASRock (in my market at least). I had initially discarded the Gigabyte board because of the chipset fan, however it turns out that it can be set to "silent" mode in the bios. This enterprising chap also replaced the stock fan with a quieter one. In an AnandTech review, the Gigabyte x570 board did very well in terms of power consumption and boot times and ranked medium to high in terms of cpu performance. I didn't check the finer points regarding cooler clearance though.

 

[gobenji]

This deviation peaked my interest, so i've tested a bit more. After 12 hours under high load it says an average of 81% (low 48%)

Thanks a lot for providing those numbers. Which CPU are you using and which bios version are you running?

 

[spiral]

I use a 3600, bios version is F20b. I was planning on upgrading when the new one's come out, but as it stands now it's plenty for my needs.

 

[gobenji]

I use a 3600, bios version is F20b. I was planning on upgrading when the new one's come out, but as it stands now it's plenty for my needs.

F20b... Wait, what motherboard do you have?
Also, as a matter or reproducibility, what did you use for load? Sorry I didn't ask earlier.

 

[spiral]

Oh, i have the Gigabyte X570 itx. Did mostly gaming, Division 2/RDR2 @3440*1440. And like 10 runs of CB.
Looking at Hwinfo now after 6 hours of mixed load the average is 79%

 

[gobenji]

Ah! At first I tought you had the ASRock B550. Interesting, the Gigabyte X570 is #1 on my list ATM. I want to get a Zen 3 processor so I'm waiting a bit to order the mobo.

Quoting from the original HWinfo post I linked earlier:

As stated before, this metric is only valid during a relatively stable near-full-load condition. That is due to the typical measurement accuracy of the VRM controller telemetry, and also due to the highly advanced and fast power management on Ryzen CPUs, that not only result in extremely low idle, but also in extremely rapidly changing power consumption. A suggested workload to get a stable and reproducable deviation metric is Cinebench R20 NT, with the HWiNFO sample rate set to less or equal to 1000ms.

My guess is that the formula HWinfo uses for Power Reporting Deviation is something like (PPT reported by motherboard)/(expected PPT for processor model when computational load generates max power usage). PPT (Package Power Tracking) is the "allowed socket power consumption", ie. total CPU power consumption. FYI, the 3600 is a 65W TDP part, which has "at least 88W" PPT limit. I don't know if HWinfo has a table of more specific PPT per CPU SKU. Anyways, all that to say that if you consider the average Power Reporting Deviation while you have variable load like gaming, some of which is most likely under the max possible load, the value will be < 100% even if the motherboard correctly reports power consumption. So if you have the time for more measurements, do a run of Cinebench like the quote above suggests and look at the Power Reporting Deviation during that time only.

 

[spiral]

I did a our of cb (46 passes) and it says 88% average. I see when idling it fluctuates a lot, from 30-80.
When it comes to gaming the gpu is the limiting factor at this resolution since i'm just running a 2070s.

 

[smitty2k1]

Hey all - I may have finally scored a 5600x, so I'm looking for a B550 motherboard to pair it with. Ive got an NCASE so I'm not thrilled about the sideways SATA ports but this Asrock board seemed to check all the boxes and is frequently on sale in the USA for $170.

Has the power "issue" been fixed with a BIOS update or are there any workarounds?

Would you guys consider any other B550 or 570 boards for a 5600x in the NCASE? Two M.2 slots is a must, as is good heatsink compatibility and a good audio chip. I have zero concern about VRMs as I think they are all fine and everyone blows VRM importance out of the water.

The Gigabyte see great but I'd really like to use the front panel USB C on my NCASE and the Asus looks great but is a lot more money for essentially the same features.

Thoughts?

 

[Mackan]

Hey all - I may have finally scored a 5600x, so I'm looking for a B550 motherboard to pair it with. Ive got an NCASE so I'm not thrilled about the sideways SATA ports but this Asrock board seemed to check all the boxes and is frequently on sale in the USA for $170.

Has the power "issue" been fixed with a BIOS update or are there any workarounds?

Would you guys consider any other B550 or 570 boards for a 5600x in the NCASE? Two M.2 slots is a must, as is good heatsink compatibility and a good audio chip. I have zero concern about VRMs as I think they are all fine and everyone blows VRM importance out of the water.

The Gigabyte see great but I'd really like to use the front panel USB C on my NCASE and the Asus looks great but is a lot more money for essentially the same features.

Thoughts?

I don't think the sensor issues have been fixed on this board. The latest BIOS is 1.80 though, and there might have been fixes along the way that they don't mention. But I wouldn't count on it. Also, be aware that this board has no BIOS flash back, so depending on what BIOS it ships with, you might have to use an older CPU to flash the BIOS to support your 5600x.

There is no perfect B550 ITX board. Stay away from the MSI board, since it uses a custom backplate and will not work well with some CPU coolers.

I guess the Asus would be the one to get, if you want internal USB-C port, and you don't care about VRM, as you said. Or take your chances with AsRock.

 

[Ricky Zeng]

Does anyone know if the Asrock B550 PG ITX/AX can set the PCI-E mode manually to 3.0 when using one of the newer PCI-E 4.0 graphics card? I went through the whole bios sections and also user manual but didn't find this function. There's only an option in AMD PBS screen to select x16/x8/x4 mode for the PCIE slot. I wonder if anyone's using a PCI-E 3.0 raiser in a A4 sandwich case with this board and a newer PCI-E 4.0 GPU, and if it poses any issues? Mine's RX580 (PCIE 3.0) so far, but don't know if there'll be any problems when I upgrade the GPU...

 

[smitty2k1]

Does anyone know if the Asrock B550 PG ITX/AX can set the PCI-E mode manually to 3.0 when using one of the newer PCI-E 4.0 graphics card? I went through the whole bios sections and also user manual but didn't find this function. There's only an option in AMD PBS screen to select x16/x8/x4 mode for the PCIE slot. I wonder if anyone's using a PCI-E 3.0 raiser in a A4 sandwich case with this board and a newer PCI-E 4.0 GPU, and if it poses any issues? Mine's RX580 (PCIE 3.0) so far, but don't know if there'll be any problems when I upgrade the GPU...

I am almost positive it's in there somewhere. I remember seeing it when I was going through my bios setup