RX 480 is apparently killing pcie slots

[DukenukemX]

I hate to say this but ,Your Jedi mind tricks are not going to work in this thread

The force is not with AMD

I smell a class action lawsuit building up

The hell are you on about? Anyway, someone needs to do more tests to see how other cards pull power from the pcie slot. And yes, reports said that some 1080 cards are bricking because of clock speeds. Asus denies it, but so will AMD deny this as well.

Tests tests and more tests.

 

[Ehren8879]

At the very least I would think it would void your motherboard warranty if you use a card that violates the PCI-E spec for the slot. But what if you buy an RX 480 from the company that made your motherboard?

Interesting question. If my new Gigabyte RX 480 kills my P67A-UD4-B3 we'll find out.

But then again, I'd just use that as an excuse to finally upgrade.

I will be sure to heavily tax my rig once my 480 is installed

 

[Imhotep]

NO its not, you think a motherboard company doesn't care about a product that is in spec or not? Why the hell are they part of PCI-SIG in the first place then? Do you realize that PCI-SIG is made up of every company that is from manufacturing of computer components to OEM's and system builders?

Motherboards fail for different reasons. You may have been unlucky with a few bad caps failing. If it really ware this bad the mobo lifetime of some of the mining crowd running 4 cards on a single mobo would be about 3 months.
Get your facts straight.
In other not related news. Reddit reports the world will end exactly at 12:01.02 PM.

 

[Daniel_Chang]

Motherboards fail for different reasons. You may have been unlucky with a few bad caps failing. If it really ware this bad the mobo lifetime of some of the mining crowd running 4 cards on a single mobo would be about 3 months.
Get your facts straight.

Still waiting for you to provide ANY facts...

 

[Chimpee]

Still waiting for you to provide ANY facts...

I am looking forward to his charts regarding overclock cards going over specs.

 

[Daniel_Chang]

I am looking forward to his charts regarding overclock cards going over specs.

Probably has to go out and buy a new box of crayons first...

 

[GaryJohnson]

NO its not, you think a motherboard company doesn't care about a product that is in spec or not? Why the hell are they part of PCI-SIG in the first place then? Do you realize that PCI-SIG is made up of every company that is from manufacturing of computer components to OEM's and system builders?

I don't think they care. I mean I've warrantied more than a few boards in my life, they don't ask a lot of questions about what you had plugged into it. You just tell them your board died and they send you a new one.

 

[razor1]

Motherboards fail for different reasons. You may have been unlucky with a few bad caps failing. If it really ware this bad the mobo lifetime of some of the mining crowd running 4 cards on a single mobo would be about 3 months.
Get your facts straight.
In other not related news. Reddit reports the world will end exactly at 12:01.02 PM.

I'm not saying my motherboards failed due to overclocking or what not, I don't really care when why stuff fails, I go out and buy a replacement and then RMA why stuff if its something I didn't do from my end.

What I am saying is any reason to go out of spec can give a motherboard company reason to deny you the warranty that you paid for in the cost of the original purchase of that motherboard.

 

[Imhotep]

power draw of overclocked cards - Google Search
These maybe a little old, google is your friend. have fun.

 

[razor1]

I don't think they care. I mean I've warrantied more than a few boards in my life, they don't ask a lot of questions about what you had plugged into it. You just tell them your board died and they send you a new one.

They do care, sometimes if ya get that one tech that knows his or her stuff they will ask everything. And if its a problem with the PCI-e slot the first thing they will ask what type of card was in there and what not.

 

[razor1]

power draw of overclocked cards - Google Search
These maybe a little old, google is your friend. have fun.

And do those tell you where the power is being drawn from PCI-e slot or the connectors?

 

[Presbytier]

I certainly don't doubt the possibility of this I imagine this would really only effect older mobo. Unfortunately if you are using an AMD cpu you are likely using an older mobo since that line has not been refreshed in ages.

 

[jbc029]

No, but then he just was providing evidence of specs being violated, not where it had to be at the slot draw. No need for google there (or overclocking for that matter). You can leave your card at stock frequencies, set the PL to +50%, fire up furmark and watch the wattage consumption climb. My 390 peaks at 304W pulled when PL is set to default and the highest I saw at +50% PL was 384W before I was too nervous to let it continue. Lol.

 

[Imhotep]

And do those tell you where the power is being drawn from PCI-e slot or the connectors?

Using a AMD Radeon Pro Duo with 2 x 8 pin = 150+150 + 75 watts of PCI-E how is that card even running. ?

 

[Presbytier]

I'm not saying my motherboards failed due to overclocking or what not, I don't really care when why stuff fails, I go out and buy a replacement and then RMA why stuff if its something I didn't do from my end.

What I am saying is any reason to go out of spec can give a motherboard company reason to deny you the warranty that you paid for in the cost of the original purchase of that motherboard.

I'm actually going to say that in most instances they probably just replace the board because for a lot of companies that can be cheaper in the long run. However if this becomes a widespread issue they are certainly not going to replace every board fried by a 480.

 

[lolfail9001]

Using a AMD Radeon Pro Duo with 2 x 8 pin + 75 watts of PCI-E how is that card even running. ?

It uses 3 8 pins to begin with.

 

[Trimlock]

Thread exploded - guess this is a hot topic!

We all know the RX480 pulls the most on average over the PCIe bus than any other card. These motherboard failures could be due to components that were on the brink of failure to finally get taken out due to the new additional stress. If the slot did indeed burn you'd see black traces on the copper connectors and maybe black residue around the plastic connector.

This could be a combo of old motherboards, shitty motherboards, old shitty motherboards or a component that was potentially on the brink of failure anyways. The stress did help push it over to finally die but that doesn't mean you can't look at both components.

BTW I was thinking of putting a RX480 (Non ref) into an old HP computer with an i7 950 to give it some new life, I'll be keeping an eye out for how AIB's handle this now.

edit: Wanted to add its better to draw the extra power from the 6 pin, the dedicated rail (if using one) has less components for failure and generally has really good tolerance for over draw. Plus if it fries on most PSU's (maybe all?) you only lose the rail the 6 pin is on.

 

[GaryJohnson]

They do care, sometimes if ya get that one tech that knows his or her stuff they will ask everything. And if its a problem with the PCI-e slot the first thing they will ask what type of card was in there and what not.

I've never had that happen. It's usually some under-educated person in India, and they just want to get you on the phone, in the system, and off again as quickly as possible. They don't care what happened to the board. Several times they haven't even asked to have the old board shipped back.

 

[razor1]

I've never had that happen. It's usually some under-educated person in India, and they just want to get you on the phone, in the system, and off again as quickly as possible. They don't care what happened to the board. Several times they haven't even asked to have the old board shipped back.

not saying it won't happen the possibility is there, And since this is now a known issue, its easy for them to say it.

 

[Imhotep]

It uses 3 8 pins to begin with.

Even with 3 8 pins + PCI - E it will draw more than 525 watts at times. How >?

 

[Araxie]

Motherboards fail for different reasons. You may have been unlucky with a few bad caps failing. If it really ware this bad the mobo lifetime of some of the mining crowd running 4 cards on a single mobo would be about 3 months.
Get your facts straight.
In other not related news. Reddit reports the world will end exactly at 12:01.02 PM.

LoL another [H]AMD defender with no freaking knowledge about what's talking about trying to defend this card, man you are so funny... you know what's better than a fanboy trying to defend his brand loyalty?. an Ignorant Fanboy trying to defend his brand loyalty with no prof of their blatant ignorance about his statements.

BTW let me enlight you a bit about mining, no serious miner do mining without the utilization of powered raisers to avoid damages to the motherboard from heat and overpower damages to the PCI-E slots, anyone considered a serious miner, those that utilize all available PCI-E slots and several machines to mining, utilize dedicated PSUs for the GPUs connected apart from the motherboard to precisely avoid any current damage to the motherboard, so first rule of mining? always run powered raisers, do you know what's a raiser? it's an extended PCI-E slot, low quality just work as extender, high quality have dedicated capacitors and power to avoid motherboard damages just like the one attached.

Spoiler

and even that isn't one of the best used...

Spoiler

this is the kind of preferable used raisers to be able to do some configurations like this one:

Spoiler

This is the safer option to prevent motherboard damages from mining.. .

that's how yo avoid shit like this:

Spoiler

or this:

Spoiler

...

This is how a serious mining farm looks:

Spoiler

yo see? al the freaking GPUs are isolated from the motherboards with dedicated covered and protected risers.. with dedicated and protected high Quality PSUs.

So next time save yourself the embarrassing statements you are making here, this is not the forum for people like you.

EDIT: Adjusted size and added as spoiler to save some space.

 

[Logan321]

Seems like a whole lot of "I'm right unless you prove me wrong" flying about with no proof either way yet. I doubt this will be solved in a forum either way. Maybe chill out for a few days and see what other information comes to light?

 

[razor1]

Even with 3 8 pins + PCI - E it will draw more than 525 watts at times. How >?

Because its out of spec and drawing from the PCI-e Connectors, that is hell of alot better than pulling over spec from the board slot, at least you are limited to a PSU failure, not a system failure or other component failure, you can limit yourself. The 6990 did the same thing.....

Also are you talking about total system power or just the card, cause I think just the card it doesn't do that much.

 

[razor1]

LoL another [H]AMD defender with no freaking knowledge about what's talking about trying to defend this card, man you are so funny... you know what's better than a fanboy trying to defend his brand loyalty?. an Ignorant Fanboy trying to defend his brand loyalty with no prof of their blatant ignorance about his statements.

BTW let me enlight you a bit about mining, no serious miner do mining without the utilization of powered raisers to avoid damages to the motherboard from heat and overpower damages to the PCI-E slots, anyone considered a serious miner, those that utilize all available PCI-E slots and several machines to mining, utilize dedicated PSUs for the GPUs connected apart from the motherboard to precisely avoid any current damage to the motherboard, so first rule of mining? always run powered raisers, do you know what's a raiser? it's an extended PCI-E slot, low quality just work as extender, high quality have dedicated capacitors and power to avoid motherboard damages just like the one attached.

and even that isn't one of the best used...

this is the kind of preferable used raisers to be able to do some configurations like this one:

This is the safer option to prevent motherboard damages from mining.. .

that's how yo avoid shit like this:

or this:

...

This is how a serious mining farm looks:

yo see? al the freaking GPUs are isolated from the motherboards with dedicated covered and protected risers.. with dedicated and protected high Quality PSUs.

So next time save yourself the embarrassing statements you are making here, this is not the forum for people like you.

Impressive set up!, were you setting up a testing system before you put things into racks?

 

[Imhotep]

LoL another [H]AMD defender with no freaking knowledge about what's talking about trying to defend this card, man you are so funny... you know what's better than a fanboy trying to defend his brand loyalty?. an Ignorant Fanboy trying to defend his brand loyalty with no prof of their blatant ignorance about his statements.

BTW let me enlight you a bit about mining, no serious miner do mining without the utilization of powered raisers to avoid damages to the motherboard from heat and overpower damages to the PCI-E slots, anyone considered a serious miner, those that utilize all available PCI-E slots and several machines to mining, utilize dedicated PSUs for the GPUs connected apart from the motherboard to precisely avoid any current damage to the motherboard, so first rule of mining? always run powered raisers, do you know what's a raiser? it's an extended PCI-E slot, low quality just work as extender, high quality have dedicated capacitors and power to avoid motherboard damages just like the one attached.

and even that isn't one of the best used...

this is the kind of preferable used raisers to be able to do some configurations like this one:

This is the safer option to prevent motherboard damages from mining.. .

that's how yo avoid shit like this:

or this:

...

This is how a serious mining farm looks:

yo see? al the freaking GPUs are isolated from the motherboards with dedicated covered and protected risers.. with dedicated and protected high Quality PSUs.

So next time save yourself the embarrassing statements you are making here, this is not the forum for people like you.

I have no business defending anything here. Its all out there just grasp it son.

 

[Trimlock]

Even with 3 8 pins + PCI - E it will draw more than 525 watts at times. How >?

Each 8 pin supplies "up to" 150 watts

3x 150 = 450

+
PCIe 75 watts

equals 525watts

You can over draw from all 4 possible connections and still stay within spec, the RX480 is already over drawing and some reports have it OUT of spec.

 

[Formula.350]

fair point about the fx8350, but even with a shitty psu it shouldnt kill the board

[Context: leldra was replying to this comment: "(not to mention that this was done on CX500, oh boy)." from lolfail9001 in post #17]
I've been reading a fair bit of stuff you've been posting, and in general you seem to have a good handle on computing.... However, this comment makes me question that. :\ I don't quite understand how you don't think that a PSU can't fry a motherboard. We have no clue as to how old the PSU is, what it's service life has consisted of, the amount of usage it's had; these all matter.
Situation 1: Once you start driving something out of spec, for instance the motherboard's PCIe power rails, which those components being ran outside of their intended (designed) capacities I can only assume are going to start causing voltage and/or current ripple. Now, imagine what's going to happen if a PSU is old or seen a lot of service... It will eventually drift out of spec and its own ripple will only be amplified in this case.
Situation 2: The PSU's age, capacitors start to dry up, ripple increases, fries the motherboard. Don't think that's possible? I hope you do, since that's exactly what happened to me 7 years ago. I had an Antec 430W (or something like that) which I had bought back in the SMP Socket A days for my Dual Athlon MP system. I continued to run it with my S939 Athlon 64 X2 rig for a few more years. We moved, then I unpacked all my stuff, system ran for awhile, but ultimately that board died. I figured just age, so I plugged my stuff into another S939 I had... that board is dead too... Not only that, so is my X2! *sniff*

I dunno, I'm no electrical engineer, so maybe I'm waaayyy off base here. To me, it seems like a very real possibility that this all could possibly start a chain reaction resulting in board death, that a newer or better (quality) PSU may have been able to prevent. This is all just my take on things, and I felt it was a slightly ignorant* comment to have made, given it's very capable of killing a motherboard (and anything attached to it, worse case).
[* in the dictionary definition of the word, not implied as an insult]

EDIT: Forgot to mention that I though this PCPer article on the matter was pretty good:
Power Consumption Concerns on the Radeon RX 480 | PC Perspective

 

[Araxie]

Impressive set up!, were you setting up a testing system before you put things into racks?

each rack tested individually.. is not mine personally I don't like mining, but already have helped to build some large mining farms for friend and clients..

extra pic from below..

Spoiler

EDIT: attaching files is working kinda slow..

 

[Araxie]

Sir.
I don't know if is regarding my video, that you are talking about.
But if your reply is yes, we can tell you you can check with oscilloscopes.
Then we compared, and noticed the difference from highest accurate system and simply, will exist.
If you pay attention, you will see that we have explained clear about the precision, .. about the lose a bit the signal, and even about the reason we made the tool, and let people understand about "how 3,3 and 12v works" "what mean wattage, voltage and current, and so on.
I can assure... and then we can show, the difference exist, for sure...
Depends the tools or the system, we can talk about... 2%....3%....,5,%...let's force.... 8%....10% @@...
We cannot justify more than 65% above, and it's clear for a lot of websites.
Maybe you don't see the reason of the video, and not see we consider a very good competitive card.... and we told the providence shoud be taken fast in order to prevent damage... mainly to simply boards and crossfire.
If is not about my video, sorry.

Best Regards

No sir, im sorry but no, it wasn't meant for you, IDK ever if your response it's to me. hahaha, I was responding to another guy, see the quote..

 

[Trimlock]

[Context: leldra was replying to this comment: "(not to mention that this was done on CX500, oh boy)." from lolfail9001 in post #17]
I've been reading a fair bit of stuff you've been posting, and in general you seem to have a good handle on computing.... However, this comment makes me question that. :\ I don't quite understand how you don't think that a PSU can't fry a motherboard. We have no clue as to how old the PSU is, what it's service life has consisted of, the amount of usage it's had; these all matter.
Situation 1: Once you start driving something out of spec, for instance the motherboard's PCIe power rails, which those components being ran outside of their intended (designed) capacities I can only assume are going to start causing voltage and/or current ripple. Now, imagine what's going to happen if a PSU is old or seen a lot of service... It will eventually drift out of spec and its own ripple will only be amplified in this case.
Situation 2: The PSU's age, capacitors start to dry up, ripple increases, fries the motherboard. Don't think that's possible? I hope you do, since that's exactly what happened to me 7 years ago. I had an Antec 430W (or something like that) which I had bought back in the SMP Socket A days for my Dual Athlon MP system. I continued to run it with my S939 Athlon 64 X2 rig for a few more years. We moved, then I unpacked all my stuff, system ran for awhile, but ultimately that board died. I figured just age, so I plugged my stuff into another S939 I had... that board is dead too... Not only that, so is my X2! *sniff*

I think Dan or someone would be able to better answer this question. I'm pretty sure that if the main 12volt rail were to be blown there are fault tolerances to help prevent damage to other components not to mention the additional 6pin or 8pin that most motherboards use for CPU power.

 

[Ieldra]

running Furmark on full screen completely maxxed out @ 1440p, both cards are on stock fan profile with a slight OC @100% usage as we speak. I also threw like 4 youtube play lists on my second monitor. Now we wait.

the furry-butthole of heat will be run till 7:30 PM est until I get home.

Please post an ominous video later today, showing your pc running for 45 minutes straight, then suddenly fade to black with the crying baby sound from the Brazilian power measurement video. This could make history, think of the memes.

AMD Overdrive warning


You can conclude AMD allows you freedom to blow up your own system if you want - no holding hands in other words.

Unlimited powaaaaaah

 

[Araxie]

I have no business defending anything here. Its all out there just grasp it son.

Oh! clearly your post history prove you are wrong.

 

[Gideon]

 

[Imhotep]

Has it occurred to you that every card has its own power circuit and the reading is a sum of all three. That is why a Rx 480 can draw more than what the specs call for. Just as any other card can and more often than not does.
Have a great weekend Gents

 

[xorbe]

Somehow we've jumped from a 150w video card, to overclocked mining rigs running five cards 24/7 full tilt

 

[Trimlock]

Somehow we've jumped from a 150w video card, to overclocked mining rigs running five cards 24/7 full tilt

Its quite an impressive display of hardware I must say. Too bad its only used for mining.

 

[Imhotep]

Oh! clearly your post history prove you are wrong.

Im Roy Taylor can you keep a secret ...

 

[rbuass]

No sir, im sorry but no, it wasn't meant for you, IDK ever if your response it's to me. hahaha, I was responding to another guy, see the quote..

Ashamed here... hahaha
Post deleted.

 

[Araxie]

Somehow we've jumped from a 150w video card, to overclocked mining rigs running five cards 24/7 full tilt

that's why this card it's the wet dream of any miner.. low wattage at big hash/rates..

 

[Ieldra]

I think Dan or someone would be able to better answer this question. I'm pretty sure that if the main 12volt rail were to be blown there are fault tolerances to help prevent damage to other components not to mention the additional 6pin or 8pin that most motherboards use for CPU power.

On the psu?

[Context: leldra was replying to this comment: "(not to mention that this was done on CX500, oh boy)." from lolfail9001 in post #17]
I've been reading a fair bit of stuff you've been posting, and in general you seem to have a good handle on computing.... However, this comment makes me question that. :\ I don't quite understand how you don't think that a PSU can't fry a motherboard. We have no clue as to how old the PSU is, what it's service life has consisted of, the amount of usage it's had; these all matter.
Situation 1: Once you start driving something out of spec, for instance the motherboard's PCIe power rails, which those components being ran outside of their intended (designed) capacities I can only assume are going to start causing voltage and/or current ripple. Now, imagine what's going to happen if a PSU is old or seen a lot of service... It will eventually drift out of spec and its own ripple will only be amplified in this case.
Situation 2: The PSU's age, capacitors start to dry up, ripple increases, fries the motherboard. Don't think that's possible? I hope you do, since that's exactly what happened to me 7 years ago. I had an Antec 430W (or something like that) which I had bought back in the SMP Socket A days for my Dual Athlon MP system. I continued to run it with my S939 Athlon 64 X2 rig for a few more years. We moved, then I unpacked all my stuff, system ran for awhile, but ultimately that board died. I figured just age, so I plugged my stuff into another S939 I had... that board is dead too... Not only that, so is my X2! *sniff*

I dunno, I'm no electrical engineer, so maybe I'm waaayyy off base here. To me, it seems like a very real possibility that this all could possibly start a chain reaction resulting in board death, that a newer or better (quality) PSU may have been able to prevent. This is all just my take on things, and I felt it was a slightly ignorant* comment to have made, given it's very capable of killing a motherboard (and anything attached to it, worse case).
[* in the dictionary definition of the word, not implied as an insult]

EDIT: Forgot to mention that I though this PCPer article on the matter was pretty good:
Power Consumption Concerns on the Radeon RX 480 | PC Perspective

I think Dan or someone would be able to better answer this question. I'm pretty sure that if the main 12volt rail were to be blown there are fault tolerances to help prevent damage to other components not to mention the additional 6pin or 8pin that most motherboards use for CPU power.

The 12v rail will have overcurrent protection of course, but the limit at which it is set must also allow for fluctuation within spec. Moreover, the damage we are talking about, that could potentially happen, is due to localized power draw; as pcper mentioned it is about power dissipation in pins and connectors.

The 12v rail on the mobo does more than just power the gtx 480