AMD once again violating power specifications? (AMD RX-480)

[patrickjp93]

2 minutes ago, Trixanity said:

Interesting. That was a quick withdrawal. Maybe AMD threatened with lawyers or an editor thought it was too early to aim their guns at AMD.

AMD can afford lawyers?

 

[Sintezza]

16 minutes ago, patrickjp93 said:

Apologies, it's a bit early in the morning (or actually stupid late for me right now) and I'm tripping. Everyone is finding overdraw problems on the PCIe slot. Some haven't been willing to push it past 120W over the slot. At least AMD is listening on reddit for now. For now, Tom's is the only one claiming to have borked a test bench.

 

Ah okay so only Toms has killed a board.

But i´m interested in how they all measured the exact powerdraw from the pci-e slot.

I´m going to revoke my bullshit claim on this, however to me its unlikely that it realy is that bad as getting claimed.

Unfortunatly i cannot get my hands on a card to test this myself.

Otherwise i would have had solderd some wires directly to the motherboard, and connected my fluke to it.

To see how much it actualy draws from the slot, and if that comes anywhere near the danger zone.

[Misanthrope]

4 minutes ago, Bouzoo said:

I like to the give everyone the opportunity to make his/her claim, everyone can be right or wrong. No matter the history... most of the time.

Yeah I know it's just an on-going forum thing: it happens so often Patrick might as well change his title to *Citation needed

[TheRandomness]

1 hour ago, Sintezza said:

Ah AMD has launched a decent performing budget oriented card.

On which Nvidia has no aswer asof yet.

And we allready see tons of BullShit like this showing up.

 

 

When we finally get something good for the budget, someone will come along and shoot it down because we can't have nice things in this world c:

[mariushm]

1 hour ago, KeltonDSMer said:

Can you tell me how the 12V traces from 24pin to PCIe socket are wired in a mobo? I was under the impression that they are direct from the 24pin to the PCIe socket...no components to burn out except the traces themselves. If this is the case, why is this an issue? Drawing power via the slot would be no risk to the board unless the PCB itself can't dissipate a watt or two from the voltage drop over the length of the traces.

The motherboard is made out of several layers one on top of the other with VIAs making the connections between layers. At least one layer is mostly just a ground plane (big large copper surface) and a layer is mostly voltages, so you don't really have thin electric traces going to the slots, you have wide large areas carrying voltages around the motherboard and just near the pci express slot you have a trace coming out of that large copper area and connects to the pins in the pci express slot. 

 

Most motherboards have at least 4 layers, high end motherboards usually have 6 layers or more. Some video cards can be made out of up to 8-10 layers (just imagine each GDDR5 chip having a few hundred thin wires going to the GPU, you just can't make all the connections in just two layers of pcb.

 

Problem is the ATX 24 pin connector is kinda old, and kept the same for compatibility reasons, and there's only two 12v wires in the connector, so in total the motherboard has only something like 15 A of current on 12v, excluding the CPU power connectors.

 

A video card should try to not pull more current than recommended from the slots themselves because what if there's 2 or more video cards installed? The currents add up... Still, the RX 480 issue is blown out of proportions.
 

[KeltonDSMer]

@zMeul @Majestic @wolfsbane3083

 

Thanks for the replies guys. I completely agree that the "gauge" (lets just say continuous current rating) of a mobo's traces are less than that of a even a cheap 20AWG wired PSU, but does this matter in this situation? How many 12V traces go to each slot? 1, 2, 3 (I have no idea)? I mean, shouldn't we be asking each other if any of this is actually dangerous instead of bickering back and forth about whether or not the cards are drawing over spec from the slot (they obviously are)?

 

Without knowing how much current the traces can safely support, I can't make the argument that this is putting the mobo at risk.

 

I honestly hate care analogies because no one ever uses a good one, but I'll try . A car's starter motor draws hundreds of amps, yet the wire going to the motor's solenoid is WAY undersized to dissipate the amount of generated heat under a constant load. This is OK because the starter isn't running 100% duty cycle, it only runs for a split second when you start your car. I think of the rx 480's PCIe slot over draw in a similar way. The question isn't if the card is drawing over spec from the slot, but if this is actually a problem.

 

While I appreciate the fan header comparison, these instantaneous spikes are a much different scenario than a constant load. However, if the average current draw through these traces does end up creating more thermal energy than what the mobo can passively dissipate, there definitely is a problem....maybe they should bundle this with the rx 480 : http://www.evga.com/Products/Product.aspx?pn=100-MB-PB01-BR

 

[KeltonDSMer]

3 minutes ago, mariushm said:

Snip
 

Wow that is incredibly useful and insightful information to me, thank you for taking the time to explain that!

[zMeul]

7 minutes ago, Majestic said:

What? No, he's still not making the distrinction between PEG and Mainboard draw. 

Not even mentioning the fact that the 1080 is enthusiast, and the 480 is mainstream.

 

It's about as apples to oranges as you can get.

PEG is the mainboard .. what are you talking about!?!?!?

 

this is the PEG:

PEG - PCI Express Graphics

[Majestic]

2 minutes ago, zMeul said:

PEG is the mainboard .. what are you talking about!?!?!?

 

this is the PEG:

 

PEG - PCI Express Graphics

Between this

 

And this

 

OK?

 

 

 

[spartaman64]

5 minutes ago, zMeul said:

PEG is the mainboard .. what are you talking about!?!?!?

 

this is the PEG:

PEG - PCI Express Graphics

i think that guy was trying to support your arguments

[zMeul]

10 minutes ago, KeltonDSMer said:

I mean, shouldn't we be asking each other if any of this is actually dangerous

dangerous, yes it is

 

you need to keep in mind that OEMs tend to not over-engineer cheap mobos to meat a price bracket

if they design the board to the extend they meet the PCI-SIG standard, 75W, it's absolutely possible that straining those traces will burn them

it's the same thing happening when people put way to many fans on 1A FAN header - it melts the traces on the board

 

what happens when people start OCing them? simple - power draw over PEG will increase too

[zMeul]

4 minutes ago, Majestic said:

Between this

 

And this

 

OK?

 

 

 

what!?

1. PEG
2. PCIe 6-8 PSU power connector

[Majestic]

2 minutes ago, zMeul said:

what!?

1. PEG
2. PCIe 6-8 PSU power connector

They are sometimes also referred to as PEG-cables. Glad you're not debating retarded, pointless semantics. 

Make a point please, this just seems like a red herring.

[zMeul]

2 minutes ago, Majestic said:

They are sometimes also referred to as PEG-cables

wait, what!?

I never heard PCIe 6-8pin referred to a PEG cable

[Dackzy]

How did this go so far? I mean come on they can fix this with a BIOS update, it really is that simple.

[laminutederire]

14 minutes ago, Prysin said:

took nearly a full page before someone got it... wow.

 

The fanboy are strong with this thread.

yeah it's hard to discuss the real matter, or to discuss about what is the real matter.

I mean I root for AMD but they obviously made something wrong, but the discussion wasn't about that so much.

13 minutes ago, Majestic said:

What? No, he's still not making the distrinction between PEG and Mainboard draw. 

Not even mentioning the fact that the 1080 is enthusiast, and the 480 is mainstream.

 

It's about as apples to oranges as you can get.

(An enthusiast card must be more perfect than the other)

I know the distinction, but I haven't used it since I never used that exemple to say nvidia is faulty or whatever.

I repeat myself: I said that the total draw of the 1080 during spikes is above what it should with the pci e slot and the 8pin (i.e. 300W during spikes with 225W with the normal restrictions). That means that it's logically straining either the main board or the 8pin (apparently it strains the 8pin, but that makes no difference for the argumentation). It's certified for everything and it works well, and doesn't break mobos or psus. We can thus infer that the fact the spikes are above doesn't matter for certifications and safety, it's the average that counts. AMD is then faulty of drawing too much power on average from te motherboard, and therefore they should probably try to make that power to be drawn from the 6pin and not the motherboard, but it is above from just 10 to 15 W and not 100W as some maybe have said.

that's how I understand the issue, and if someone thinks I'm wrong, then I'd be happy to discuss it, but only with some rationalization of what's going on.

[raphidy]

2 minutes ago, Dackzy said:

How did this go so far? I mean come on they can fix this with a BIOS update, it really is that simple.

How do you push a bios update to the mainstream?

[Dackzy]

1 minute ago, raphidy said:

How do you push a bios update to the mainstream?

through driver updates is one way

[laminutederire]

2 minutes ago, raphidy said:

How do you push a bios update to the mainstream?

you make a campaign of recall where reseller flash the new bios?

[IGJoe2192]

bunch of whiners on here. 

[Prysin]

27 minutes ago, patrickjp93 said:

No, you're deflecting. AMD is pulling 100-110 ff the PCIe slot ALONE at stock settings in some games, most notably The Witcher 3. That is endangering the motherboard, period.

its spikes, not constant load.

[Majestic]

9 minutes ago, laminutederire said:

(An enthusiast card must be more perfect than the other)

 

I said that the total draw of the 1080 during spikes is above what it should with the pci e slot and the 8pin (i.e. 300W during spikes with 225W with the normal restrictions). That means that it's logically straining either the main board or the 8pin (apparently it strains the 8pin, but that makes no difference for the argumentation). 

The distinction, I repeat myself, is this:

 

GTX 1080

 

RX-480

 

 

You see the GTX1080 exceeding the 75W enveloppe anywhere? No. So you're not making the distinction. And you're missing the entire point of the post by not doing so. The problem is the 480 exceeding the PCI-SIG enveloppe, not the ATX standards.

 

9 minutes ago, laminutederire said:

It's certified for everything and it works well, and doesn't break mobos or psus. We can thus infer that the fact the spikes are above doesn't matter for certifications and safety, it's the average that counts. AMD is then faulty of drawing too much power on average from te motherboard, and therefore they should probably try to make that power to be drawn from the 6pin and not the motherboard, but it is above from just 10 to 15 W and not 100W as some maybe have said.

that's how I understand the issue, and if someone thinks I'm wrong, then I'd be happy to discuss it, but only with some rationalization of what's going on.

No you can't infer, because it's a complete non-sequitur. 

[79wjd]

8 minutes ago, laminutederire said:

We can thus infer that the fact the spikes are above doesn't matter for certifications and safety, it's the average that counts. AMD is then faulty of drawing too much power on average from te motherboard, and therefore they should probably try to make that power to be drawn from the 6pin and not the motherboard, but it is above from just 10 to 15 W and not 100W as some maybe have said.

that's how I understand the issue, and if someone thinks I'm wrong, then I'd be happy to discuss it, but only with some rationalization of what's going on.

No, you can't infer that at all. First of all, the 1080 is still relatively new to the market and the people who have one more than likely also have a quality PSU. Secondly, you can't infer a single thing about it damaging the motherboard because the 1080 doesn't overdraw from the motherboard but rather the PSU.

The distinction is important because traces are a much lower gauge then the wires of a PSU. Continuous is 10-20w above while peak is up to 155w above -- 155w, even for an instant, may be enough to burn out the very small traces that are on a motherboard -- No one here knows enough about the size of the traces and the duration of the peak load to determine one way or another whether or not it's more than just a technicality of breaking spec or if it's actually going to risk damage to the motherboard. 

[SteveGrabowski0]

8 hours ago, alamox said:

i hoped that AMD wouldn't be stupid enough to push the clock  way to far looking for performance, so that they end up out of the card's power envelope or limited OC headroom.

and they went and did just that, to a ridiculous extent, the 480 probably was probably designed for something like 1Ghz base clock, and performance wasn't good enough.

I wouldn't put it past AMD to put out a BIOS and driver that limits power consumption to 150W after they have already gotten all the reviews out of the way at 165W. No one is ever going to go back and review the reference cards again anyways.

[Trixanity]

7 minutes ago, patrickjp93 said:

AMD can afford lawyers?

 

That remains to be seen. They might have to convince someone to do it pro-bono or put some of that leftover Intel money toward lawyers.

6 minutes ago, Misanthrope said:

Yeah I know it's just an on-going forum thing: it happens so often Patrick might as well change his title to *Citation needed

Can't we petition the admins to give him that title? That would be great.