kongfra

dlbsyst
I have always bought EVGA cards because I thought they were the best quality. After this I no longer feel this way. Most likely will be getting Asus top card the Strix. I might pay a bit more but the quality is there.

 
ASUS makes good hardware overall, but good luck ever dealing with an RMA if you ever need it.  Most times ASUS will decline your RMA , or take weeks to get it back if you are lucky, happened to me with their mobo and will never buy ASUS mobos or their hardware ever again. I also had a problem with an ASUS monitor and still in warranty and they declined RMA, so yeah I am a bit sour. The main reason I go with EVGA is peace of mind for their customer service, I only had to use their RMA once for my EVGA 680TI but it was the easiest RMA I ever had to do with any hardware vendor, which is why I stick with EVGA

My friend recently did RMA for his 3080 strix and the process was pretty seam less for him. Not saying the stories are false but it doesn’t always happen. I had a faulty 1440p monitor and ASUS replaced it too.

It also helps their warranty depo is located an hour from us so it’s better to drop off product there than pay $200 CAD to ship product to evga every time the card dies. Once is ok but if this card dies more than once I would take my chances with other oems.

umarehsan

kongfra

dlbsyst
I have always bought EVGA cards because I thought they were the best quality. After this I no longer feel this way. Most likely will be getting Asus top card the Strix. I might pay a bit more but the quality is there.

 
ASUS makes good hardware overall, but good luck ever dealing with an RMA if you ever need it.  Most times ASUS will decline your RMA , or take weeks to get it back if you are lucky, happened to me with their mobo and will never buy ASUS mobos or their hardware ever again. I also had a problem with an ASUS monitor and still in warranty and they declined RMA, so yeah I am a bit sour. The main reason I go with EVGA is peace of mind for their customer service, I only had to use their RMA once for my EVGA 680TI but it was the easiest RMA I ever had to do with any hardware vendor, which is why I stick with EVGA

My friend recently did RMA for his 3080 strix and the process was pretty seam less for him. Not saying the stories are false but it doesn’t always happen. I had a faulty 1440p monitor and ASUS replaced it too.

It also helps their warranty depo is located an hour from us so it’s better to drop off product there than pay $200 CAD to ship product to evga every time the card dies. Once is ok but if this card dies more than once I would take my chances with other oems.

 
When I sent in my mobo it took close to 6 weeks to get it back, I ended up just buying a new mobo so I can get my machine up and running, I guess everyone has different experiences I just know I won't touch them ever again LOL

Anyone else still had OCP tripped after limiting framerate and downclocking/locking the voltage of their cards?
 
I have a FTW3 and ran though all the steps I have seen being done online (Limiting frame rate, underclocking, locking in voltage, setting power limits to lower etc). Still yesterday after about an hour of playtime, the entire screen went black and I had to reset. I assume this was OCP kicking in. I promptly uninstalled the game and installed in on my laptop (3060 card) and its been running just fine.
 
I've already had to RMA the FTW3 once due to the gorilla glue thermal compound ruining a card when installing the waterblock. I don't want to have to RMA it a second time. I just don't see this getting fixed without a driver/vBIOS update to fix the power issue.

I am also using my main rig for WFA purposes which is why I didn't even think about chancing it when OCP kicked in. It happened once, and that was enough for me to uninstall. I knew that whatever their "fix" was, was clearly not an actual fix.
 
Not worth the risk to my daily 8-5, not over a $40 pre-order beta game. Not over ANY game tbh.  Thankfully I just recently bough this asus gaming laptop with good specs (Ryzen + 3060) so it can game reasonably well. Just not ideal to be gaming on a laptop.
 
 

AJKagan
I am also using my main rig for WFA purposes which is why I didn't even think about chancing it when OCP kicked in. It happened once, and that was enough for me to uninstall. I knew that whatever their "fix" was, was clearly not an actual fix.
 
Not worth the risk to my daily 8-5, not over a $40 pre-order beta game. Not over ANY game tbh.  Thankfully I just recently bough this asus gaming laptop with good specs (Ryzen + 3060) so it can game reasonably well. Just not ideal to be gaming on a laptop.
 
 

But its not the game. In other words yes you can avoid New World and your card will continue to work. But its a ticking time bomb. Whose to say the next game released wont also have the same issue and kill your card? Do you just wait on every game before you play to see if it kills these cards and then if it does you never play it? Thats not a functional product. Check out my link in the post below yours

AJKagan
Anyone else still had OCP tripped after limiting framerate and downclocking/locking the voltage of their cards?
 
I have a FTW3 and ran though all the steps I have seen being done online (Limiting frame rate, underclocking, locking in voltage, setting power limits to lower etc). Still yesterday after about an hour of playtime, the entire screen went black and I had to reset. I assume this was OCP kicking in. I promptly uninstalled the game and installed in on my laptop (3060 card) and its been running just fine.
 
I've already had to RMA the FTW3 once due to the gorilla glue thermal compound ruining a card when installing the waterblock. I don't want to have to RMA it a second time. I just don't see this getting fixed without a driver/vBIOS update to fix the power issue.

There have been several posts on the Nvidia reddit thread and also other posts in this thread here where people have had the OCP trip in *AFTER* the patch and AFTER limiting the FPS to their monitor refresh rate manually.  New World's patch "reduces" the likelihood of the problem happening but does NOT remove it.  You can't stop a hardware problem with software without gimping the card, as mentioned before.  Do you remember when Furmark was frying VRM's?  What did Nvidia and AMD do to this day?  They throttled the card by app detection in the drivers.  People found out by renaming furmark as Quake3.exe and seeing much higher power draw after.  
 
I have no idea if that still works to this day.
 
It seems like the only way to stop the card from actually dying is to lower the power limit.

ALL Cards with iCX 

yangning8109
I RMA'd my RTX3090 ftw3 ultra 3 times, so I got 4 different cards, 3 rev 0.1 ones and 1 rev 1.0 (last good one). First one (red lip rev 0.1) dead when I played GTA V at 4K (like those dead in New World); Second one (red lip rev 0.1) has fan rpm, PCIe power balance issue, and random black screen in game; Third one (black lip rev 0.1) still has PCIe power balance issue; Fourth one is black lip with rev 1.0 board, no problem so far. Thus, various problems of FTW3 cards currently reported by media are all exactly true according to my experience.

Are you checking the revision of your card by the PCIe power draw? Or is there a better way? Also what is black lip and red lip and what is their significance?

KeithW19
Are you checking the revision of your card by the PCIe power draw? Or is there a better way? Also what is black lip and red lip and what is their significance?

Look at the right of the PCIe insert on your card by an ear and the revision number is there. 

Is it worth trying to send a 3090FTW3 in for the 1.0 revision or that just not possible?

KeithW19
EVGA's GPU fans are trying to hit 200,000 RPM in New World, that might explain a fried chip or two | PC Gamer
 
That article may be relevant. I have yet to see someone with a Rev 1.0 card to say they had the failure. So I suspect that when the fan tries to spin at 200,000 rpm the card draws too much voltage from the PCIe slot and the card fries. Rev 1.0 cards apparently fixed the PCIe overdraw issue

I am skeptical.  It is a PWM fan.  It can only operate up to 100%.  Telling it to spin at 5,000 RPM is the exact same amount of load as telling to to spin (supposedly) at 200,000 RPM.  Both will simply be 100%.  It can't draw extra voltage.  It can't draw extra current.  That isn't how electricity works.  100% is as much as it can be told to consume.  There is no such thing as 4000% duty cycle.
 
The RPM reporting bug has existed for a long time.  Instead of focusing on a software bug and a theory that doesn't make sense, why don't they instead focus their editorial talent on the ongoing power balancing issues and the associated power spikes and throttling which is occurring as a result?

Hello, I received a beta key for New World Beta today and I own a RTX 3090 FTW3 ULTRA.

I'm curious to see what happens when I play it.

ty_ger07

KeithW19
EVGA's GPU fans are trying to hit 200,000 RPM in New World, that might explain a fried chip or two | PC Gamer
 
That article may be relevant. I have yet to see someone with a Rev 1.0 card to say they had the failure. So I suspect that when the fan tries to spin at 200,000 rpm the card draws too much voltage from the PCIe slot and the card fries. Rev 1.0 cards apparently fixed the PCIe overdraw issue

I am skeptical.  It is a PWM fan.  It can only operate up to 100%.  Telling it to spin at 5,000 RPM is the exact same amount of load as telling to to spin (supposedly) at 200,000 RPM.  Both will simply be 100%.  It can't draw extra voltage.  It can't draw extra current.  That isn't how electricity works.  100% is as much as it can be told to consume.  There is no such thing as 4000% duty cycle.
 
The RPM reporting bug has existed for a long time.  Instead of focusing on a software bug and a theory that doesn't make sense, why don't they instead focus their editorial talent on the ongoing power balancing issues and the associated power spikes and throttling which is occurring as a result?

What he's trying to say is that the same bug that is causing the fan to spin to max RPM (multiple people have already said that even when they use FIXED RPM, when the fan does the 200,000 RPM thing, it's actually spinning at MAX speed (or rather, beyond max speed, similar to the speed when the OCP trips), is the exact SAME bug that causes the OCP to happen.  Read the previous posts.  Several users here have already said that they tested that right before they black screen, the fan spins up to maximum RPM (like in the Anno 1800 game in the building tooltip).
 
It's NOT a 'reporting' bug.  The fans ARE actually revving up, but what's reported is incorrect.  So whatever is possibly CAUSING this to happen could also be causing fuses to blow or VRM's to get damaged.  Again-the ONLY people who know EXACTLY what is going on is eVGA--and they are very obviously not going to make a press release explaining the exact cause of this (since they would be forced to do a recall/RMA of all cards if they did that, like the Intel P67 recall for RMA for "B3" stepping PCH).
 
There are PLENTY of Ultra rich people on this forum (They won't admit it since well...ego).  Yet not a single person has offered to donate a revision 0.1 3090 FTW3 to buildzoid or Elmor--two of the most qualified people on the planet for troubleshooting the cause.  It's sort of disgusting, actually.  You guys want to know the precise cause of the problem yet no one is willing to take one for the team and send a card (not 1.0 version) to Elmor or buildzoid.  Yet people want to whine and complain.
 
Months ago I asked someone to donate a card to buildzoid.  Yet STILL no one has.  And there are PLENTY of people who can get by without a card.
 
Disgusting.  If you don't trust buildzoid, send it to Elmor.

It seems like the entire 3090 line was basically a major screwup.  That model was being pushed to the very limits of what it was capable of doing right out of the box already, which meant that a much larger fraction of the ones that made it into the wild ended up failing - there was a much smaller margin for error than there was with the 3080.  Obviously they didn’t all fail; plenty of people have 3090’s that work just fine, but a higher percentage of them ended up going bad than with the other 3k cards.  

kevinc313

ty_ger07

Kylearan
or rather, beyond max speed

Not possible.  Read PWM specifications.  There is no such thing as greater than 100% duty cycle.  Do they tap into free energy to break the 100% barrier?
 
 
 
OCP will cause the fan to go to 100%.  How can you determine which came first?  The chicken or the egg?  Considering that we know beyond a shadow of a doubt that a big percentage of these cards have a power balance issue and are known to power cap when they shouldn't (or OCP when they shouldn't), doesn't it make more sense that the fan reaching 100% is a symptom of OCP triggering and is unrelated to the fan sometimes displaying erroneous values?
 
I agree that the fan bug is annoying and should be fixed, but I don't believe that the fan bug is killing cards.  The fan simply does not consume much power.  You could jump them from 0% to 100% instantaneously and they won't consume more than a couple dozen watts.  A couple dozen watts is nothing in the grand scheme of things.

It should be pointed out that the fans are typically rated for 3500 RPM and .55 AMP, so they are capped at under 100% for 3000 RPM, but I highly doubt even running at full current at a few extra fractions of an amp would be enough to blow out the individual header each one is on. 
 
The 'Gamers' article misreports Igor's theory, which is already dubious at best, and shows they are technically incompetent not understanding how a PWM fan works.  IMHO the "insane fan speed" is part of the protect mode where it intentionally 'fails' to a solid 5v or nothing on the PWM line, giving full fan speed.

My understanding from Igor's is not that fans are consuming too much power and that's the reason for fuses exploding yet the ICX3 voltage controller is failing maybe to the point of producing a short circuit, and there you get OCP triggered and exploding fuses. But to me that's way too much speculation. I'm closer to the theory of PCIe port consuming too much power due to the balancing issue, and that makes the 10A fuse of the PCIe port power path to explode. That'd explain why only the 0.1 revision is affected.