Overclock.net › Forums › Graphics Cards › AMD/ATI › Reports of RX 480 damaging PCI-E slots
New Posts  All Forums:Forum Nav:

Reports of RX 480 damaging PCI-E slots - Page 4  

post #31 of 139
Quote:
Originally Posted by EightDee8D View Post

750ti ( no 6pin) has sold probably in millions, people also overclocked that gpu and used with cheap i3+mobo. haven't seen any complaints yet. but this must be the end of universe.

Yup and that gpu pulls 120W from the PCI-E slot alone. There was also the 960 which pulled 100W from the slot. No one complained about those.
post #32 of 139
Quote:
Originally Posted by Redeemer View Post

Haven't heard of R9 295x2 owners having issues regarding the PCIe slot, and I am sure that card especially when overclock pulls 500w easy.

The R9 295x2 is powered from two 8-pin PCIe power connectors and the PCIe slot. The maximum specified power draw from these adds up to 375 watts. According to AMD, the R9 295x2 can draw up to 500 watts on some typical scenarios and can even peak to about 600 watts

In other words the whole RX 480 PCIe violation is all FUD

That's because 295X2 draws most of its power through the 2x8 pins and not the PCIe slot.



PEG = PCIe x16 slot, and at its peak the card is only drawing 40W.
post #33 of 139
So, are you telling me that AMD didn't bother to perform a test run on these cards first? Are you telling me it basically went straight from a blueprint to the store?

The entire quality control assurance department should be fired, and I'm being completely honest. No mistake like this should ever be made.
Edited by aweir - 7/1/16 at 4:56pm
post #34 of 139
If power distribution is bios configurable and not hardwired as some have speculated, then whoever programmed the bios dun goofed big time. Could be as simple as flipping the values for PEG vs auxiliary power input.

If 480 had gone over spec on the 6 pin this would've almost been a non-issue.
post #35 of 139
Quote:
Originally Posted by looniam View Post

how many times do you need told that is wrong?

Are you just ignorant of the actual mechanical facts about the PCIe connectors? PCIe power connectors are Molex type connectors.

#1 Molex pins - these pins are rated for 6A each under normal load. This is why older motherboards used multiple pins - too much draw on a single one would overload it and burn the pin. This engineering standard HAS NOT CHANGED IN THIRTY YEARS. While "capable" of 8 amps, it's not advisable to run them higher than 6. The spec for PCIE allows for about 3.35A per pin but they'll carry 6A each all day long without getting hot at 12V.

#2


You see how it has the note in there "Official spec lists 'no connection' about pin 2 on the 6pin connector? Where it is commonly supplied with 12v the same as the other two? Like 93% of supplies power this pin. The way to tell if your supply DOESN'T power that pin is if there's no wire going into it's connection.

On a 6+2 connector the extra 2 pins are a duplicate ground and a duplicate of the first sense line... or more commonly they're just both grounds

This piece of garbage joke 8( with a jumper across the +2 component.

So... while you may have heard something different: 6pin PCIe cable is 3-hot 3-not and an 8pin PCIe is 3-hot 6-not providing extra grounding to deal with sudden shifts in power on the card causing a need to dissipate the charged state of the components and ensuring the non-energized surface contact of the grounding exceeds the needs of the power draw.

When you're dealing with slip connectors you want the grounding connection to have MORE surface area than the power feed connection - its why the ground connection on a 3 wire plug almost always has a larger surface area in contact in much of the world. Ground fault equipment reduces this need somewhat...

So basically, I know the mechanical facts about the PCIe auxiliary molex 6 pin, 8 pin and 6+2 pin connectors. And the twin-end 6+2 cables as well.

It is entirely plausible that the RX480 is trying to draw power from the #2 pin but the failures are occurring on supplies that don't have it connected.
Edited by prjindigo - 7/1/16 at 5:19pm
post #36 of 139
Quote:
Originally Posted by magnek View Post

If power distribution is bios configurable and not hardwired as some have speculated, then whoever programmed the bios dun goofed big time. Could be as simple as flipping the values for PEG vs auxiliary power input.

If 480 had gone over spec on the 6 pin this would've almost been a non-issue.

Yeah, what gets me is that the draw from both sides seems equal, like someone traced a circuit path wrong or somehow bound the two supply lines together.
post #37 of 139
This isn't something you can't fix with a BIOS update other than forcing the card to only use 150w, regardless of Wattman settings. The issue has to do with how the power phases are connected to their power sources.
First Build
(10 items)
 
  
CPUMotherboardGraphicsRAM
i3-4370 ASROCK H97M PR04 380x Nitro Corsair 8GB 
Hard DriveHard DriveHard DriveOS
Crucial M500 120GB Seagage 320GB Caviar Blue 1TB Windows 8.1 x86-64 
PowerCase
XFX 550W PS07B 
  hide details  
First Build
(10 items)
 
  
CPUMotherboardGraphicsRAM
i3-4370 ASROCK H97M PR04 380x Nitro Corsair 8GB 
Hard DriveHard DriveHard DriveOS
Crucial M500 120GB Seagage 320GB Caviar Blue 1TB Windows 8.1 x86-64 
PowerCase
XFX 550W PS07B 
  hide details  
post #38 of 139
Quote:
Originally Posted by prjindigo View Post

Quote:
Originally Posted by looniam View Post

how many times do you need told that is wrong?
Warning: Spoiler! (Click to show)
Are you just ignorant of the actual mechanical facts about the PCIe connectors? PCIe power connectors are Molex type connectors.

#1 Molex pins - these pins are rated for 6A each under normal load. This is why older motherboards used multiple pins - too much draw on a single one would overload it and burn the pin. This engineering standard HAS NOT CHANGED IN THIRTY YEARS. While "capable" of 8 amps, it's not advisable to run them higher than 6. The spec for PCIE allows for about 3.35A per pin but they'll carry 6A each all day long without getting hot at 12V.

#2


You see how it has the note in there "Official spec lists 'no connection' about pin 2 on the 6pin connector? Where it is commonly supplied with 12v the same as the other two? Like 93% of supplies power this pin. The way to tell if your supply DOESN'T power that pin is if there's no wire going into it's connection.

On a 6+2 connector the extra 2 pins are a duplicate ground and a duplicate of the first sense line... or more commonly they're just both grounds

This piece of garbage joke 8( with a jumper across the +2 component.

So... while you may have heard something different: 6pin PCIe cable is 3-hot 3-not and an 8pin PCIe is 3-hot 6-not providing extra grounding to deal with sudden shifts in power on the card causing a need to dissipate the charged state of the components and ensuring the non-energized surface contact of the grounding exceeds the needs of the power draw.

When you're dealing with slip connectors you want the grounding connection to have MORE surface area than the power feed connection - its why the ground connection on a 3 wire plug almost always has a larger surface area in contact in much of the world. Ground fault equipment reduces this need somewhat...

So basically, I know the mechanical facts about the PCIe auxiliary molex 6 pin, 8 pin and 6+2 pin connectors. And the twin-end 6+2 cables as well.

It is entirely plausible that the RX480 is trying to draw power from the #2 pin but the failures are occurring on supplies that don't have it connected.

you missed what i bolded?

the grounds on the power connector has NOTHING to do with the power supplied from the slot; which has its own grounds.

and btw, that piece of garbage is fully pci-sig compliant:

Edited by looniam - 7/1/16 at 5:29pm
loon 3.2
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7-3770K Asus P8Z77-V Pro EVGA 980TI SC+ 16Gb PNY ddr3 1866 
Hard DriveHard DriveHard DriveOptical Drive
PNY 1311 240Gb 1 TB Seagate 3 TB WD Blue DVD DVDRW+/- 
CoolingCoolingOSMonitor
EKWB P280 kit EK-VGA supremacy Win X LG 24MC57HQ-P 
KeyboardPowerCaseMouse
Ducky Zero [blues] EVGA SuperNova 750 G2 Stryker M [hammered and drilled] corsair M65 
AudioAudio
SB Recon3D Klipsch ProMedia 2.1  
  hide details  
loon 3.2
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7-3770K Asus P8Z77-V Pro EVGA 980TI SC+ 16Gb PNY ddr3 1866 
Hard DriveHard DriveHard DriveOptical Drive
PNY 1311 240Gb 1 TB Seagate 3 TB WD Blue DVD DVDRW+/- 
CoolingCoolingOSMonitor
EKWB P280 kit EK-VGA supremacy Win X LG 24MC57HQ-P 
KeyboardPowerCaseMouse
Ducky Zero [blues] EVGA SuperNova 750 G2 Stryker M [hammered and drilled] corsair M65 
AudioAudio
SB Recon3D Klipsch ProMedia 2.1  
  hide details  
post #39 of 139
Quote:
Originally Posted by NuclearPeace View Post

This isn't something you can't fix with a BIOS update other than forcing the card to only use 150w, regardless of Wattman settings. The issue has to do with how the power phases are connected to their power sources.

And that's just pure speculation at the moment.

Actually come to think of it, if the phases are hardwired to their power sources, doesn't it simply limit how much power you can draw from each source while staying within the specs of the mosfets themselves? Plus there's nothing stopping you from drawing LESS power from one source.
post #40 of 139
Quote:
Originally Posted by lolfail9001 View Post

You know, if it was someone trying to get a free mobo out of this, he would not mention that he overclocked both rx 480 and CPU. And he would definitely not mention that he uses a freaking CX500 instead of a power supply.

So either, it's a troll, or it's legitimate problem. Which one remains to be seen.

Exactly. That CPU at 4.5GHZ is pulling quite a bit of power, not to mention having a rx 480 overclocked, but not overvolted.

Anyone who has "fried" a motherboard from the rx 480, should list their PSU and CPU frequency. Seems like in this case it could be the PSU.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
This thread is locked  
Overclock.net › Forums › Graphics Cards › AMD/ATI › Reports of RX 480 damaging PCI-E slots