RX 5700 Fan Control Issue...

[PowerPie5000]

My new PowerColor RX 5700 Red Dragon arrived yesterday and I got round to installing it today. Installation was easy enough as usual (completely removed the previous drivers for my RX 580). This is where the simplicity ends.

I got about 5 minutes in to playing The Outer Worlds and my GPU temp hit 95C, the memory temp hit 90C and the GPU Hot Spot reached 110C... Not Good! Turns out the fans weren't spinning up at all. The fan sensor then disappeared completely from HWinfo and the fan reading in Wattman went up to 65535rpm (when they weren't even moving).

I've tried auto and manual fan speeds, different driver versions, double checked all connections and tried both the OC and silent BIOS via the switch on the card. It seems the fans pretty much come and go when they feel like it. Sometimes they work, sometimes they don't (sensor reading ends up greyed or completely missing in HWinfo when they stop responding).

I'm guessing it's RMA time for this card already? I'm pretty sure the fan sensor/control is faulty.... I was hoping it was a driver issue, but I don't think it is.

System specs:

AMD Ryzen 5 3600 (running stock)
Asrock X470 Taichi with the latest BIOS
16GB HyperX Predator DDR4 3333 cl16
512GB Sabrent Rocket M.2 NVME boot drive
2TB Hitachi/Toshiba HDD
750W Riotoro Enigma G2 PSU (basically a Seasonic Focus Gold unit)
Windows 10 and all drivers are up to date.


What do you guys think? I even used separate PCIE power cables from the PSU to the GPU instead of using a single cable with both 8-pin and 6-pin connectors. I know some people say using a single cable with both connectors is ok.

 

[Bounce]

My new PowerColor RX 5700 Red Dragon arrived yesterday and I got round to installing it today. Installation was easy enough as usual (completely removed the previous drivers for my RX 580). This is where the simplicity ends.

I got about 5 minutes in to playing The Outer Worlds and my GPU temp hit 95C, the memory temp hit 90C and the GPU Hot Spot reached 110C... Not Good! Turns out the fans weren't spinning up at all. The fan sensor then disappeared completely from HWinfo and the fan reading in Wattman went up to 65535rpm (when they weren't even moving).

I've tried auto and manual fan speeds, different driver versions, double checked all connections and tried both the OC and silent BIOS via the switch on the card. It seems the fans pretty much come and go when they feel like it. Sometimes they work, sometimes they don't (sensor reading ends up greyed or completely missing in HWinfo when they stop responding).

I'm guessing it's RMA time for this card already? I'm pretty sure the fan sensor/control is faulty.... I was hoping it was a driver issue, but I don't think it is.

System specs:

AMD Ryzen 5 3600 (running stock)
Asrock X470 Taichi with the latest BIOS
16GB HyperX Predator DDR4 3333 cl16
512GB Sabrent Rocket M.2 NVME boot drive
2TB Hitachi/Toshiba HDD
750W Riotoro Enigma G2 PSU (basically a Seasonic Focus Gold unit)
Windows 10 and all drivers are up to date.


What do you guys think? I even used separate PCIE power cables from the PSU to the GPU instead of using a single cable with both 8-pin and 6-pin connectors. I know some people say using a single cable with both connectors is ok.

Yes I'd rma it.

 

[PowerPie5000]

Yes I'd rma it.

Yeah I was thinking it might come to that. Part of me was hoping someone would say it's a driver bug or some other setting. I'll get it sent back... Not sure if I should go with the same model though?

 

[MyBrainz]

Yeah I was thinking it might come to that. Part of me was hoping someone would say it's a driver bug or some other setting. I'll get it sent back... Not sure if I should go with the same model though?

Before you RMA, try the AMDCleanupUtility.exe in the Bin64 folder that you can find in the driver folder.

This will clean install in Safe Mode.. then reinstall the 'latest' driver and see if anything changes.

 

[PowerPie5000]

Before you RMA, try the AMDCleanupUtility.exe in the Bin64 folder that you can find in the driver folder.

This will clean install in Safe Mode.. then reinstall the 'latest' driver and see if anything changes.

Just read this a bit too late as I've already started the RMA process. I doubt it would have made any difference to be honest as I completely uninstalled/wiped the drivers several times. I also tried 3 different driver versions with the last being the current beta. I used the AMD utility and DDU to uninstall everything and then start from scratch.

I think the card I received must have a faulty fan controller. I've opted for a replacement of the same model to be sent and i'm hoping it's not a bad batch.

 

[Freddie1980]

Just a thought could it be something simple like the fan lead cable isn't properly plugged into the card/loose?

 

[PowerPie5000]

Just a thought could it be something simple like the fan lead cable isn't properly plugged into the card/loose?

I checked the two fan connections that I could see (one on the PCB and another along the edge of the heatsink). They seemed to be firmly in place. I'll see how the replacement goes. If it still has issues then I'll get the Sapphire Pulse version, or worst case spend a bit more and get an RTX 2060 Super... I'd prefer not to spend so much on a mid range GPU though.

 

[MyBrainz]

Just read this a bit too late as I've already started the RMA process. I doubt it would have made any difference to be honest as I completely uninstalled/wiped the drivers several times. I also tried 3 different driver versions with the last being the current beta. I used the AMD utility and DDU to uninstall everything and then start from scratch.

I think the card I received must have a faulty fan controller. I've opted for a replacement of the same model to be sent and i'm hoping it's not a bad batch.

Ah ok, was a thought if you hadnt done this already. Lets hope your replacement works as intended when you receive it.

Let us know how it goes with the replacement

 

[Bounce]

To be honest I wouldn't recommend people start messing about with a PCB, when under warranty.

If your not careful you could risk invalidating warranty , it's just not worth the risk.

 

[PowerPie5000]

To be honest I wouldn't recommend people start messing about with a PCB, when under warranty.

If your not careful you could risk invalidating warranty , it's just not worth the risk.

I didn't remove any parts or do anything with the PCB (it has warranty stickers too). The only thing I did do is check to see if the fans were actually connected properly. The fan connections are exposed and easy to access without removing anything.

Im really hoping it was just a one off fault and that the replacement will be fine.

 

[PowerPie5000]

Ok I got the replacement.... And I still have the same issue! Surely I can't have 2 faulty cards in a row unless it's a bad batch? Or dodgy drivers? Sometimes the fans work intermittently and sometimes they don't spin up at all!

I've completely removed every trace of the drivers using the AMD clean utility and DDU in safe mode and powered down. All the connection from the PSU are perfectly fine (separate 8-pin and 6-pin cables). The GPU is firmly inserted in to the slot, all the drivers are up to date including the motherboard BIOS and chipset drivers. The PSU and all my other components are fine and I can't see anything in the BIOS that would affect the GPU fans. I've heard that Wattman is junk with these cards so I'm not using it. It doesn't report the fan speed correctly anyway (again showing 65535rpm when the fans are at 0%).

Now do I abandon AMD and return the card and then grab an RTX 2060 Super instead? It'll end up costing me more though. What else could be causing the fans not to spin up when they should?

 

[Bounce]

Ok I got the replacement.... And I still have the same issue! Surely I can't have 2 faulty cards in a row unless it's a bad batch? Or dodgy drivers? Sometimes the fans work intermittently and sometimes they don't spin up at all!

I've completely removed every trace of the drivers using the AMD clean utility and DDU in safe mode and powered down. All the connection from the PSU are perfectly fine (separate 8-pin and 6-pin cables). The GPU is firmly inserted in to the slot, all the drivers are up to date including the motherboard BIOS and chipset drivers. The PSU and all my other components are fine and I can't see anything in the BIOS that would affect the GPU fans. I've heard that Wattman is junk with these cards so I'm not using it. It doesn't report the fan speed correctly anyway (again showing 65535rpm when the fans are at 0%).

Now do I abandon AMD and return the card and then grab an RTX 2060 Super instead? It'll end up costing me more though. What else could be causing the fans not to spin up when they should?

Faulty motherboard?

 

[drakulton]

Perhaps it's an intermittent signal connection between the card and the mobo. Are all the pci-e contacts clean, is there any carbon on them?

Is there another GPU slot you can try?

 

[PowerPie5000]

Perhaps it's an intermittent signal connection between the card and the mobo. Are all the pci-e contacts clean, is there any carbon on them?

Is there another GPU slot you can try?

I think I've found the problem... It seems the cards BIOS is at fault. I don't usually bother with GPU BIOS updates as I've never found the need, but I thought I'd just give it a go anyway. I found a newer BIOS version for my card dated September 2019 (old one was August 2019) and so far the fans seem to be working perfectly fine and the temps also seem pretty good across the whole card (GPU, VRM and memory etc). It also stays pretty quiet under load just like my previous MSI RX 580 Gaming X .

I'll keep my eye on the fan readings, but so far it's been behaving. I blasted it for a while with Metro Exodus and The Outer Worlds just to make sure

 

[Bounce]

I think I've found the problem... It seems the cards BIOS is at fault. I don't usually bother with GPU BIOS updates as I've never found the need, but I thought I'd just give it a go anyway. I found a newer BIOS version for my card dated September 2019 (old one was August 2019) and so far the fans seem to be working perfectly fine and the temps also seem pretty good across the whole card (GPU, VRM and memory etc). It also stays pretty quiet under load just like my previous MSI RX 580 Gaming X .

I'll keep my eye on the fan readings, but so far it's been behaving. I blasted it for a while with Metro Exodus and The Outer Worlds just to make sure

Where do you find these bios updates?

 

[drakulton]

I think I've found the problem... It seems the cards BIOS is at fault. I don't usually bother with GPU BIOS updates as I've never found the need, but I thought I'd just give it a go anyway. I found a newer BIOS version for my card dated September 2019 (old one was August 2019) and so far the fans seem to be working perfectly fine and the temps also seem pretty good across the whole card (GPU, VRM and memory etc). It also stays pretty quiet under load just like my previous MSI RX 580 Gaming X .

I'll keep my eye on the fan readings, but so far it's been behaving. I blasted it for a while with Metro Exodus and The Outer Worlds just to make sure

Nice one! You could've spent a lot of time and money swapping other components out. I'm glad you got the issue sorted.

 

[PowerPie5000]

Where do you find these bios updates?

Techpowerup VGA BIOS Collection... It's one PowerColor submitted themselves. It's working perfectly and I don't even need to add custom fan curves or use afterburner as the default fan profile is fine. The Red Dragon is actually pretty decent and it's one of the cheaper cards with custom cooling and a factory overclock.

 

[PowerPie5000]

Nice one! You could've spent a lot of time and money swapping other components out. I'm glad you got the issue sorted.

I was ready to give up on AMD and spend more on an RTX 2060 super with similar performance. Glad I decided to test a different BIOS and it's been perfect so far with no crashes or glitches or anything weird (I've heard so many horror stories about AMD drivers with Navi cards).

Im sure PowerColor must be aware of the issue. I had two cards purchased from the same place within 2 days with the same issue, so I'm guessing it's a bad batch with a dodgy BIOS.

 

[drakulton]

I was ready to give up on AMD and spend more on an RTX 2060 super with similar performance. Glad I decided to test a different BIOS and it's been perfect so far with no crashes or glitches or anything weird (I've heard so many horror stories about AMD drivers with Navi cards).

Im sure PowerColor must be aware of the issue. I had two cards purchased from the same place within 2 days with the same issue, so I'm guessing it's a bad batch with a dodgy BIOS.

What version of GPU bios were you having issues with, and what mobo/mobo bios were you using when you had the issue? Just might be a bit of info that folks with a similar setup can relate to if they have a similar issue.

 

[PowerPie5000]

What version of GPU bios were you having issues with, and what mobo/mobo bios were you using when you had the issue? Just might be a bit of info that folks with a similar setup can relate to if they have a similar issue.

My mobo BIOS is the latest available from Asrock for the x470 Taichi (version 3.60). The GPU BIOS with the fan issue was version 017.001.000.049.000000 and is dated and timed '2019-08-13 02:33:00'... The BIOS I'm using now has the same version number, but is dated and timed '2019-09-15 23:31:00'