RX 480 is apparently killing pcie slots

razor1 said:
Yeah the guy came on to ask for help on what he should do and he saw the issues with the card, and he got blasted for it, instead of trying to help him he got accused...... its was pathetic.
Click to expand...

Will undervolting the GPU significantly affect PCI-E power draw? Seems that it's overvolted and has caused throttling due to thermals, but if voltage lowered and the card doesn't throttle, it would consume more power anyway, correct? Would it be a wash?
 
yes and it will drop the power usage considerably going from 1.3 to 1.1 volts I would say maybe as much as 30 watts or so?

But the thing with undervolting still gotta be careful because now we don't know if AMD did those volts because the GPU needed those volts for stability reasons, need more tests.
 
Lord Risky said:
Not to beat a dead horse here, but I still don't think that is correct. Anyone else like to give their input?
Click to expand...
HELP! Please... Is a 6+2 pin connector same as 8 pin?
Can I fit a 6+2 pin into an 8 pin connector?
Is 6+2pin the same as 8pin?
(common question at toms)
https://forums.geforce.com/default/topic/472194/pci-e-power-cable-8-pin-vs-6-pin/
6+2 pin pcie cable same as native 8 pin on seasonic m12? | Overclockers Forums

If you didn't want to beat a dead horse, you could have googled it quickly
 
Lord Risky said:
Pretty sure I am right and I think tyns' post proves it. We are all guilty of posting fud's from time to time and should research/verify our understandings before posting.
Click to expand...

As i said, pretty certain NONE of my PSU's actually work this way, and all the tear down's I've ever seen never support this diagram either. So not so sure PSU makers actually follow this or not.
 
razor1 said:
yes and it will drop the power usage considerably going from 1.3 to 1.1 volts I would say maybe as much as 30 watts or so?

But the thing with undervolting still gotta be careful because now we don't know if AMD did those volts because the GPU needed those volts for stability reasons, need more tests.
Click to expand...

Users seem to be getting higher clocks (1300+MHz) but GPU-z is reporting 160W and that's the chip only, correct?
 
Lord Risky said:
Not to beat a dead horse here, but I still don't think that is correct. Anyone else like to give their input?
Click to expand...

If you look at the 8 pin connector, look at how the last two are removable and you will also notice they are extensions of the other neighboring wires.
 
Tyns said:
Will undervolting the GPU significantly affect PCI-E power draw? Seems that it's overvolted and has caused throttling due to thermals, but if voltage lowered and the card doesn't throttle, it would consume more power anyway, correct? Would it be a wash?
Click to expand...
If it genuinely is over volted then simply reducing the voltage will reduce heat, power use and most likely allow higher clocks as well.
I doubt that the GPU is over volted though, unless AMD really have lost a screw.

ps
Heat is directly proportional to power used.
Power is proportional to voltage squared.
A small change in voltage can net a fair change in power and heat.
 
Nenu said:
Lets see if I am full of it as you so eloquently put :)
In a perfectly set up system you can go way beyond power specs for many parts of a circuit. Others wont cope so well.

There has to be leeway in the spec for typical problems such is a bad connection.
A bad connection can be caused by many problems.
A bent pin.
A badly inserted connector.
A corroded connector.
Dirt on a connector.
A frayed wire.
...

There are 2 power supply rails and 2 ground rails on a 6 pin connector.
A problem on one of these can increase the load on its partnering wire, up to 100%.
A bad connection on any of these it will insert a resistance.
As current flows it drops a voltage across any resistance, dropping power over the the resistance.
This causes the bad connection to heat up.
The higher the current flow, the hotter the bad connection will get.

Losing one wires connection completely, one wire will have to take the full load which will cause the remaining wire and connections to heat up.
As long as the current flow is within spec and the remaining wires have good connections, this might not cause a problem.
If power draw is outside spec the extra heat generated could cause a serious problem.

If there is a bad connection with current still flowing through it, this will generate a lot more heat.
Keeping power use within spec this can still burn out. I've had this happen.
With power above spec, even more damage can occur.

Either of the above situations can induce fire in the wrong circumstance.
Click to expand...


coolstorybro, I've no use for strawmans, we're done conversing. A problem with life can cause death, we're not talking about plugging something into faulty hardware, it's faulty, period. And that's the last I bother with you Mr.I'mSoSmartAndAwesome.
 
Tyns said:
Users seem to be getting higher clocks (1300+MHz) but GPU-z is reporting 160W and that's the chip only, correct?
Click to expand...


I'm not sure about GPUz's power reading but if its still getting 160 watts, its probably hitting the same problem with the pci-e bus. But they are overclocking it so.....
 
Vercinaigh said:
coolstorybro, I've no use for strawmans, we're done conversing. A problem with life can cause death, we're not talking about plugging something into faulty hardware, it's faulty, period. And that's the last I bother with you Mr.I'mSoSmartAndAwesome.
Click to expand...
lol.
Wonderful comeback :)

Reading is a useful skill, it will help you create a coherent response.
Perfectly functional hardware can exhibit exactly those problems as well.
And most PSUs fail safes will not protect against these types of problem because they do not overload the PSU.
 
razor1 said:
yes and it will drop the power usage considerably going from 1.3 to 1.1 volts I would say maybe as much as 30 watts or so?

But the thing with undervolting still gotta be careful because now we don't know if AMD did those volts because the GPU needed those volts for stability reasons, need more tests.
Click to expand...


Some german reviewers tested the theory out and undervolted the gpu, power draw dropped and performance actually went up lol. Maybe there is something to the idea that AMD baked in too much voltage scaling in the bios?
 
Nenu said:
Its already exceeding the 6pins specced power.
Its not a matter of pushing more power through that connector, its a matter of fixing the card or withdraw it from sale, then motherboards wont get burnt out.
The reason the card needs fixing is because there is a serious quality issue with it.
Click to expand...
You should really read posts that give specs. Most 6 pins can handle a full 192W, as most have 3 12V wires and these are the lowly 8( cant remember the units, posted it yesterday) whereas the higher ones are up to 11.

So spec be damned and that logic you live by. AMD needs to get that PCI-slot power down and moving to the PCI-e cable is not only smart but nearly risk free.
 
thesmokingman said:
Some german reviewers tested the theory out and undervolted the gpu, power draw dropped and performance actually went up lol. Maybe there is something to the idea that AMD baked in too much voltage scaling in the bios?
Click to expand...


Maybe just don't know :)
 
JustReason said:
You should really read posts that give specs. Most 6 pins can handle a full 192W, as most have 3 12V wires and these are the lowly 8( cant remember the units, posted it yesterday) whereas the higher ones are up to 11.

So spec be damned and that logic you live by. AMD needs to get that PCI-slot power down and moving to the PCI-e cable is not only smart but nearly risk free.
Click to expand...
You need to provide a citation for that.
Perhaps you can show me the specs I should read, that back up what you said.
 
JustReason said:
You should really read posts that give specs. Most 6 pins can handle a full 192W, as most have 3 12V wires and these are the lowly 8( cant remember the units, posted it yesterday) whereas the higher ones are up to 11.

So spec be damned and that logic you live by. AMD needs to get that PCI-slot power down and moving to the PCI-e cable is not only smart but nearly risk free.
Click to expand...
Why is it that a bunch of uninformed Nvidiots reside in AMD subforums .
 
So let me summarize how I see everything based on the information that I know

Video cards historically have loaded and (in articles ppl listing for checking power such as 960GTX as comparisons) overloaded system components outside of their specs, be it power supply or mobo power draw. In the past GPU makers would (over)load up the PSU on the 6-8 pin and just sip power from the PCI-E slot.

AMD with the new card has decided to however split the power draw between the 6 pin and the PCI-E slot evenly, which is causing at/over limit power draw from the PCI-E bus reportedly and its killing mobo's, at that all the mobo's I see reported "Dead" have been bargain basement ASROCK's etc and have ironically shown that a lot of motherboards are crap. This does not bode well for AMD which is marketing the 480 as "Cutting edge bargain bang for your buck" which is resulting in ppl sticking them into shitty computers like DELL's etc and might cause some components to explode other than the PSU in the past.

Now coincidentally by AMD pushing the PCI-E spec to its limits they've opened up a whole can of worms for how shoddy the motherboard market is, so now AMD has a choice to either change the 6 pin to an 8 pin, or for ppl to realize how terrible their bargain basement computer parts are and have to upgrade
 
Nenu said:
You need to provide a citation for that.
Perhaps you can show me the specs I should read, that back up what you said.
Click to expand...
Seriously you just like being difficult.
upload_2016-7-2_0-41-32.png


Read it and weep.
 
 
When people on this forum, or any other forum that attracts experienced PC enthusiasts, start documenting their destroyed mobos, then I'll take notice.

Until then I'm dismissing the claim that RX 480 melts PCIe slots as the conclusions of inexperienced system builders and teenage gamers - similar to the "Nvidia driver bricked my GPU", "Crimson driver bricked my GPU" and "Denuvo killed my SSD" nonsense we've seen in the past.

We need to ask ourselves, "Why does every one except us experience problems with drivers bricking GPUs, GPUs burning mobos and other catastrophic failures"? Why are these serious problems so widespread on Reddit and on the AMD forums and the Nvidia forums but nonexistent on Hardforum? Considering we run our machines ragged, shouldn't Hardforum see more complaints of failed hardware then the average forum?

I'm aware of the photos from the AMD forums. But literally the first thing that caught my eye with those pics was that the case looked like it had been dropped out of plane. It also looks as though it has no exhaust fans and it also seems the PC was never dusted. The handwriting on that piece of paper he attached looked childlike. Right now the only facts we can ascertain from those photos is the machine was owned by a young person, was poorly maintained and ventilated and may have been damaged when it was assembled.

So, I'm taking these claims with a grain of salt for now. When [H] starts documenting fried mobos, then I'll lend it more credence.
 
  • Like
Reactions: N4CR
like this
Nenu said:
Difficult?
I asked you to prove what you said, perfectly reasonable to anyone else.

Where did you get that from?
Context may matter a lot.
Click to expand...
I went looking for specs to better understand what the PCI-SIG spec sheets showed. Found a guy on OCN that found that spec years ago and held onto them for reference.
 
trandoanhung1991 said:
Lol, people still refuse to believe that the 480 causes issues even though there is actual video proof for it.



At 2:00 mark he starts to explain the issue.

Quick summary: He's building a cheap $350 AMD system with an AM2 motherboard. With the 480 in it, it would shut down every time (without damage it seems, lucky guy). Here's the kicker: he puts in a 980Ti and no problems.

And the 480 works in his Asus Z170-A motherboard, obviously, because those use overkill components for high overclocking headroom. It's a $165 motherboard. So obviously it can at least sustain that kind of load, but for how long?

RX 480 is such a hipster, because it went mainstream before it was cool :ROFLMAO:
Click to expand...

He's using an old AM2 motherboard, which I assume is used. He blames the power, but not really testing anything. He's assuming. I would assume that some component on the motherboard was near the end of its life or really worn out. I doubt he spent $165 for that AM2. It might have been at some point worth that much. If he did, someone should have told him a 860K + motherboard is far cheaper and better setup. He's full of it.
 
crusty_juggler said:
When people on this forum, or any other forum that attracts experienced PC enthusiasts, start documenting their destroyed mobos, then I'll take notice.

Until then I'm dismissing the claim that RX 480 melts PCIe slots as the conclusions of inexperienced system builders and teenage gamers - similar to the "Nvidia driver bricked my GPU", "Crimson driver bricked my GPU" and "Denuvo killed my SSD" nonsense we've seen in the past.

We need to ask ourselves, "Why does every one except us experience problems with drivers bricking GPUs, GPUs burning mobos and other catastrophic failures"? Why are these serious problems so widespread on Reddit and on the AMD forums and the Nvidia forums but nonexistent on Hardforum? Considering we run our machines ragged, shouldn't Hardforum see more complaints of failed hardware then the average forum?

I'm aware of the photos from the AMD forums. But literally the first thing that caught my eye with those pics was that the case looked like it had been dropped out of plane. It also looks as though it has no exhaust fans and it also seems the PC was never dusted. The handwriting on that piece of paper he attached looked childlike. Right now the only facts we can ascertain from those photos is the machine was owned by a young person, was poorly maintained and ventilated and may have been damaged when it was assembled.

So, I'm taking these claims with a grain of salt for now. When [H] starts documenting fried mobos, then I'll lend it more credence.
Click to expand...
I agree that more scientific approaches are best. But my biggest worry is some will attempt to make negative issues more prevalent and murky the issue beyond reality.
 
DukenukemX said:
He's using an old AM2 motherboard, which I assume is used. He blames the power, but not really testing anything. He's assuming. I would assume that some component on the motherboard was near the end of its life or really worn out. I doubt he spent $165 for that AM2. It might have been at some point worth that much. If he did, someone should have told him a 860K + motherboard is far cheaper and better setup. He's full of it.
Click to expand...
I don't feel he is trying to be unfair at all. I like his frankness and attempts at honesty. At least he doesn't seem to have an agenda.
 
DukenukemX said:
He's using an old AM2 motherboard, which I assume is used. He blames the power, but not really testing anything. He's assuming. I would assume that some component on the motherboard was near the end of its life or really worn out. I doubt he spent $165 for that AM2. It might have been at some point worth that much. If he did, someone should have told him a 860K + motherboard is far cheaper and better setup. He's full of it.
Click to expand...

He falls into the same bucket as the Crusty_Juggler described. Young little knowledge etc..., but the entertaining part of it all is the experts at hand...:) Beating a dead horse over and over just because some clowns have access to the interwebs.
All of this must be true. How could AMD release a card that actually works...lol
 
DukenukemX said:
He's using an old AM2 motherboard, which I assume is used. He blames the power, but not really testing anything. He's assuming. I would assume that some component on the motherboard was near the end of its life or really worn out. I doubt he spent $165 for that AM2. It might have been at some point worth that much. If he did, someone should have told him a 860K + motherboard is far cheaper and better setup. He's full of it.
Click to expand...

Considering the whole system works with a 980ti, it is the video card preventing it. Apart from the system crashing when the benchmark becomes demanding, what else would you assume it to be?
 
JustReason said:
I went looking for specs to better understand what the PCI-SIG spec sheets showed. Found a guy on OCN that found that spec years ago and held onto them for reference.
Click to expand...
The inbuilt headroom is to reduce the chance of catastrophic failure.

Most PSUs cannot detect a bad connection or a single missing connection so its critical to protect from them generating a lot of heat.
ie the PSU will not shut down when encountering a bad connection or missing connection on a 6 or 8 pin power plug.
By having higher capacity on the wires, when one is in trouble, the system may still operate safely.

I gave my take on this earlier.
RX 480 is apparently killing pcie slots
 
Trimlock said:
Considering the whole system works with a 980ti, it is the video card preventing it. Apart from the system crashing when the benchmark becomes demanding, what else would you assume it to be?
Click to expand...

There is no prove it is happening he does not have it recorded. He just mentions it. This is enough to discredit his credibility
 
Trimlock said:
Considering the whole system works with a 980ti, it is the video card preventing it. Apart from the system crashing when the benchmark becomes demanding, what else would you assume it to be?
Click to expand...
definitely the card but wasn't it a Foxconn MoBo? I am not seeing this as great mobo purchase, I usually look for ASUS MSI or Gigabyte when I look for a new mobo, but they do exist so the issue needs to be fixed.
 
Nenu said:
The inbuilt headroom is to reduce the chance of catastrophic failure.

Most PSUs cannot detect a bad connection or a single missing connection so its critical to protect from them generating a lot of heat.
ie the PSU will not shut down when encountering a bad connection or missing connection on a 6 or 8 pin power plug.
By having higher capacity on the wires, when one is in trouble, the system may still operate safely.

I gave my take on this earlier.
RX 480 is apparently killing pcie slots
Click to expand...
None of that changes the fact that AMD needs to alleviate the power on the PCI slot and seeing that running even 100W on the PCI-e cable will cause no real harm we have a very real solution. You seem to be making a ton of excuses for no fix at all but continue to complain about the issue.
 
JustReason said:
None of that changes the fact that AMD needs to alleviate the power on the PCI slot and seeing that running even 100W on the PCI-e cable will cause no real harm we have a very real solution. You seem to be making a ton of excuses for no fix at all but continue to complain about the issue.
Click to expand...

You have misunderstood the need for the specs headroom.
Its to protect against heat damage, and fire.
Its not sensible to deliberately flout the specs unless you know exactly what you are doing.

You cannot assume you have the ability to exceed the spec.
No manufacturer can otherwise they open themselves up to liability claims when things go wrong.
 
JustReason said:
definitely the card but wasn't it a Foxconn MoBo? I am not seeing this as great mobo purchase, I usually look for ASUS MSI or Gigabyte when I look for a new mobo, but they do exist so the issue needs to be fixed.
Click to expand...

hmm... not sure about the MB, you could have a point there. I don't really care to re-watch the video but the presenter, at least to me, did a very good job.
 
JustReason said:
definitely the card but wasn't it a Foxconn MoBo? I am not seeing this as great mobo purchase, I usually look for ASUS MSI or Gigabyte when I look for a new mobo, but they do exist so the issue needs to be fixed.
Click to expand...

You can't really get any better since the budget in that attempt was $350. Even if he had another $50 I doubt he'd spend more on the MB since the 480 already took $240 off the budget. Only about $100 for CPU+MB+RAM+PSU. Not much leeway there.
 
trandoanhung1991 said:
You can't really get any better since the budget in that attempt was $350. Even if he had another $50 I doubt he'd spend more on the MB since the 480 already took $240 off the budget. Only about $100 for CPU+MB+RAM+PSU. Not much leeway there.
Click to expand...
I think it was a re-use, in either way this is a real legitimate situation to purchase the card for (breathing new life into an ancient/cheap system) and proposes a limitation. That is, if it isn't a fluke.
 
You must log in or register to reply here.
Back
Top