Sanity Check - RX5700XT Junction Temp

Samuel T. Cogley

Hi guys - I know the Junction temp is high by design but I have a Powercolor Red Devil and it's hitting the high 80s / low 90s. Not the best intake at the moment in my P400A Mesh as I'm waiting for fans. In fairness it's only really concerning me in FurMark and the Campaign Map of TotalWar Warhammer II, which also hammers GPUs for reasons best known to Creative Assembly.

Everything is stock even the power limit. The card doesn't seem to like even a mild undervolt sadly, but I'll have a bit of a play with that.

K.O.Kiki

That does sound a bit high - how is the temp with side panel off? Also remove the PCI covers underneath the GPU.

Also you could try limiting frames to monitor refresh rate.

Samuel T. Cogley

Thanks K.O. Kiki, pretty much the same with the side panel off, maybe a degree or so lower. There's noting under the card except a wifi card but that's in the very bottom slot. Only thing I can think of is the side closest to the Motherboard there is an NVME SSD and a chip set (Asus Prime X570-P) fan maybe stopping air coming out that side, otherwise it's got fairly unobstructed.

Samuel T. Cogley

Actually I tell a lie about everything stock, I have a 60% fan limit otherwise it sounds like a jet engine.

the_syco

Samuel T. Cogley wrote: »

Actually I tell a lie about everything stock, I have a 60% fan limit otherwise it sounds like a jet engine.

Personally, I have a pair of Audio-Technica ATH-M20x, so I hear what I need to hear, and not my fan noise. My exhaust fans max RPM is 3000, but only hit 1700 when gaming in PUBG (temps would be 45 degrees).

Look at changing the max to 70%, and see what difference it makes.

Samuel T. Cogley

the_syco wrote: »

Personally, I have a pair of Audio-Technica ATH-M20x, so I hear what I need to hear, and not my fan noise. My exhaust fans max RPM is 3000, but only hit 1700 when gaming in PUBG (temps would be 45 degrees).

Look at changing the max to 70%, and see what difference it makes.

Thanks! Both the junction temp and the GPU temp are at 45C?

VeVeX

Undervolting those cards helps with temperatures and fan curves without impeding performance to any noticeable degree.

Homelander

Those temps aren't dangerous or anything but that is high. The cards fans just aren't running enough obviously since you've limited them. If you look at reviews it seems to be a good 5700XT in general.

But, high 80's or 90c under 100% load with the cooler limited at 60% sounds about right. Fine, even.

Samuel T. Cogley

Right I've been doing more testing tonight.

Temps are slightly better with the side panel on now that I've good airflow. My card is definitely not behaving like the one in the GamersNexus review. They had fans normalised at 40dBA and were getting 78C IIRC. I'm more and more convinced I've a bad cooler mount. The cooler isn't getting that warm and it's not spitting out a lot of hot air into the case compared to my old 1080ti.

So far my little 'all red team' project isn't going so well.

I've decided to run all stock and ramp up the fans a bit. At around 1875Mhz and 2500RPM fan speed I'm getting mid 80s*. Guess that'll have to do until I can summon up the courage to disassemble the card. At least these Red Devil's are only 6 screws.

*With the GPU edge temp barely hitting 60C which only reinforces it for me that it's a hot-spot due to a bad mount.

BloodBath

I wouldn't worry about junction temp too much. Nvidia don't even have a sensor to report this area for this very reason. It looks bad but it's not a problem.

A 5700xt uses a fair bit less power than a 1080ti which might explain why it's not spitting out as much hot air.

No harm in a rebuild though. You might improve temps a bit with better mounting and paste.

Samuel T. Cogley

Thanks for all the help so far guys.

I know silicon lottery and all that, but is it usual for these cards to have very little undervolting 'head' room? Mine is set to 2033Mhz boost @ 1200mV and I can't even drop it 50mV.

BloodBath

Yep. Very little wiggle room with mine as well. The days of getting big boosts are over. The manufacturers push them pretty hard out of the box.

Samuel T. Cogley

This thing is wrecking my head. I'm getting random crashes in GTA 5. From what I can tell it's not temp related. Could it be my PSU - The only one I could get at the time as an Aerocool 600W unit.

Samuel T. Cogley

Doing some reading and it's not a great PSU from what I can gather. I suppose not for the price! Can anyone suggest a good PSU for a R5 3600 and a 5700XT? Ideally that's in stock with Amazon. I'm thinking 600W is probably enough? Just 1 NVMe SSD so no mad number of HDDs or anything else I can think of that would take a lot of power.

Samuel T. Cogley

Went for a Gigabyte P650B Kolink Enclave 700W. I'm getting random lock-ups on the desktop so thinking more that it's this ****ty PSU as it's pretty random rather than under load.

simonp1

Samuel T. Cogley wrote: »

Went for a Gigabyte P650B Kolink Enclave 700W. I'm getting random lock-ups on the desktop so thinking more that it's this ****ty PSU as it's pretty random rather than under load.

Hey Samuel

Ive the same gpu and I'm getting similar crashes in Battlefield v ex. on loading this morning and one while in a game, then could play for another hour without issue. COD warfare never has an issue for some reason but reckon its driver related.

My junction temps are pushing 90+ under load too

Simon

Samuel T. Cogley

simonp1 wrote: »

Hey Samuel

Ive the same gpu and I'm getting similar crashes in Battlefield v ex. on loading this morning and one while in a game, then could play for another hour without issue. COD warfare never has an issue for some reason but reckon its driver related.

My junction temps are pushing 90+ under load too

Simon

Thanks Simon. I'm getting random crashes all over the place now. I'm really hoping it's the PSU and not the card as I bought it second hand.

daheadonu

Have had very similar issues as the OP rma'd the card and have been told nothing bis wrong with it,bought a 850w seasonic bit Overkill but stock seems to be low atm.

Samuel T. Cogley

daheadonu wrote: »

Have had very similar issues as the OP rma'd the card and have been told nothing bis wrong with it,bought a 850w seasonic bit Overkill but stock seems to be low atm.

Stock is dire for PSUs at the moment. That's why I ended up with that aerocool unit! Knew I should have looked harder. Thanks for the feedback! Hope it sorts it for you.

Samuel T. Cogley

Just looking at the 12V monitor on the MB it's switching from 11.98V to 12.076V. Can some one check my math here of 96mV ripple? That's high isn't it?

Thanks all I know this might be getting a bit boring :pac:

Edit: HW Monitor has it pegged at 12V but if I run Furmark it drops to 11.712V.

Samuel T. Cogley

New PSU (KOLink Enclave 700W) wish me luck!

daheadonu

Where did you pick it up?my order got changed to unknown when in stock...

Samuel T. Cogley

daheadonu wrote: »

Where did you pick it up?my order got changed to unknown when in stock...

Amazon, I back ordered it. There was only one left when mine got allocated.

Samuel T. Cogley

So kids - without wanting to jinx myself - don't spend £40 on your PSU.

Samuel T. Cogley

The saga continues with this bloody thing. About the only way I can get it stable is keep the temps under 80C with 100% fan under load. I've taken the cooler off an reapplied the thermal paste - No joy. I bought the card second hand - do you think it's been heat damaged?

BloodBath

That's supposed to be 1 of the best coolers for that card. No idea why yours is performing so badly.

What case are you using?

Samuel T. Cogley

BloodBath wrote: »

That's supposed to be 1 of the best coolers for that card. No idea why yours is performing so badly.

What case are you using?

Phanteks P400A, 3 intake 1 exhaust. I've even got the side panel off at the moment. It keeps the card nice and cool at 100% fan speed, it just sounds like a jet engine. I just can't work out why it crashes once it gets warm - I thought these AMD chips appeared to run that bit warmer because the hot-spot temp is exposed. Sorry to be clear the 80+C is the Junction temp still. I think Gamers Nexus had theirs running at 78C. GPU temp is in the 60s even with sensible fan settings.

Thanks for the reply BloodBath!

BloodBath

Hey no problem, bit of a head scratcher alright.

Are you sure it's not a driver issue? What driver are you using?

Have you tried lowering the clock speed and voltage a bit? Not the ideal solution but if it stops crashes then great.

Samuel T. Cogley

BloodBath wrote: »

Hey no problem, bit of a head scratcher alright.

Are you sure it's not a driver issue? What driver are you using?

Have you tried lowering the clock speed and voltage a bit? Not the ideal solution but if it stops crashes then great.

Aye, tried several drivers - DDUing each time. Tried down clocking, reduced power limit - which did help somewhat but only in so far as it kept the temps down for a bit I think.

I've even tried just installing the drivers without all the bloat and running MSI afterburner instead. No joy.

The kind of hard, complete system, lock ups it's generating leads me to believe it's hardware. I've another one on the way from Amazon so I'll try that. Just kicking myself I'm paying 2080 Super money for a RX 5700XT, well a 5700XT and a door stop.

Thanks again for your input, desperately hoping there's something I've overlooked.

K.O.Kiki

It's starting to sound like you got a bad card

Can you contact the seller to get the original receipt?

Samuel T. Cogley

K.O.Kiki wrote: »

It's starting to sound like you got a bad card

Can you contact the seller to get the original receipt?

I broke the warranty seal

Azza

The cashes your getting, is a crash to desktop or is your system locking up or power off?

K.O.Kiki

Samuel T. Cogley wrote: »

I broke the warranty seal

https://europa.eu/youreurope/citizens/consumers/shopping/guarantees-returns/faq/index_en.htm

EU law should have you covered (but I am not a lawyer).

Samuel T. Cogley

Azza wrote: »

The cashes your getting, is a crash to desktop or is your system locking up or power off?

Complete lock up and restart - I get the very occasional crash to desktop or dreaded AMD black screen crash.

Samuel T. Cogley

K.O.Kiki wrote: »

https://europa.eu/youreurope/citizens/consumers/shopping/guarantees-returns/faq/index_en.htm

EU law should have you covered (but I am not a lawyer).

Thanks K.O. Kiki! I'll have the argument with them at least. Hopefully the guy who sold me the card will be of some help, but then the bar-steward sold me a duff card so I might be a bit stuck there. In fairness I suppose there is a slight chance it was the crappy PSU I was using.

Azza

I have a 5700XT a few months now and in the last month or so my system has just turned itself off while gaming. No warning just power down and restarts albeit slowly where it seems to load the by the BIOS screen 2-3 times before getting back to windows. Also the poweroff disabled XMP for my ram and I have to re-enable it.

Bit puzzled as to whats going on.
I don't believe my PC is over heating and at 700w my PSU should be adequate though its abit of a no name brand.. When playing any game the issue would happen within 5 to 15 minutes.

I rolled back several AMD gpu drivers until about the March driver and that seemed to have fixed it. A week or so with no issue but its started again today.

Wondering if it could be updating to Windows 10 2004 or if my PSU isn't up to the job.

Samuel T. Cogley

Azza wrote: »

I have a 5700XT a few months now and in the last month or so my system has just turned itself off while gaming. No warning just power down and restarts albeit slowly where it seems to load the by the BIOS screen 2-3 times before getting back to windows. Also the poweroff disabled XMP for my ram and I have to re-enable it.

Bit puzzled as to whats going on.
I don't believe my PC is over heating and at 700w my PSU should be adequate though its abit of a no name brand.. When playing any game the issue would happen within 5 to 15 minutes.

I rolled back several AMD gpu drivers until about the March driver and that seemed to have fixed it. A week or so with no issue but its started again today.

Wondering if it could be updating to Windows 10 2004 or if my PSU isn't up to the job.

Thanks for sharing you're experience of it. I might try an even older driver than the one I'm using. I cant imagine a 700w unit isn't enough though. Running a 700w unit here too. I'm still on 1909.

Homelander

It would depend very much on the unit. 700W isn't always 700W, with the much cheaper brands instability can happen at half that wattage.

Cuddlesworth

Its not windows or the drivers. They can't make your system power off, Blue screen sure, but not actually power off or interact with the bios like that. That's a electrical fault, linked to heat.

As others have said, cheap power supplys are never worth it. Would be worth checking out if there are any pro reviews on that model.

Samuel T. Cogley

The one I have now is a decent unit it's a KOlink Enclave 700w - it's fairly high up on the PSU Hierarchy. I was wondering if the old unit might have damaged the card in some way rather than having been duped into buying a dud - but I think it's just the latter.

Cuddlesworth

I was referring to the other poster whose system is straight up restarting, not bluescreening. For random BSOD's like you have, the best option is to swap out parts till you find the culprit using some deductive reasoning.

Eg, swapping the 5700 for a 750ti might tell you the graphics card was causing it, but might not be pointing out it was the 250 watt draw from the graphics card causing the psu to drop voltage.

Azza

Yeah, I'm leaning towards PSU issue. I don't think its overheating though. Temps never seemed that bad. I also opened the case but it made no difference.

Interestingly the only thing I've done differently is connected an Xbox One controller via USB. Perhaps my power requirements are so borderline at the moment the Xbox controller tipped it over the edge. I'll try without connecting the controller.

Am leaning towards a replacement PSU at this stage. No freezing, no bsod, no warning, just poof power off.

Perhaps as a temp measure I should try undervolting my GPU?

EoinHef

Azza wrote: »

Yeah, I'm leaning towards PSU issue. I don't think its overheating though. Temps never seemed that bad. I also opened the case but it made no difference.

Interestingly the only thing I've done differently is connected an Xbox One controller via USB. Perhaps my power requirements are so borderline at the moment the Xbox controller tipped it over the edge. I'll try without connecting the controller.

Am leaning towards a replacement PSU at this stage. No freezing, no bsod, no warning, just poof power off.

Perhaps as a temp measure I should try undervolting my GPU?

Do you get the crashes with XMP left turned off?

Reminds me a bit of my old 3570K with 1666Mhz ram. XMP profile caused the PC to crash. Setting the ram to its rated speed and timings manually solved the problem. Never did figure out what was wrong with the profile.

Azza

I think so, not 100% sure but I'm pretty certain I wasn't aware that XMP was turned off after the power off issue first started happening and was using it without it turned on, but was still getting the restart.
No issue outside of gaming with the restart/power off.

Samuel T. Cogley

Cuddlesworth wrote: »

I was referring to the other poster whose system is straight up restarting, not bluescreening. For random BSOD's like you have, the best option is to swap out parts till you find the culprit using some deductive reasoning.

Eg, swapping the 5700 for a 750ti might tell you the graphics card was causing it, but might not be pointing out it was the 250 watt draw from the graphics card causing the psu to drop voltage.

Im getting complete lock ups and restarts also. Think it's just a dud card.

Azza

So far so good with the undervolt. No crashes yet but I did have one case of the dread black screen that's been plaguing 5700XT cards. In that regards I think I've been lucky.

Have noticed the recommended PSU for my GPU the MSI RX 5700XT Gaming X is 750w . So yeah I suspect the power downs are just the PSU been tapped out by the card. The undervolt seems to be reducing the power draw by around 20w. Hoping that's enough for now.

Samuel T. Cogley

Right so maybe it will crash but I've the card in the second slot (16X) and it seems more stable and seems to be drawing less power. I've checked to make sure I didn't accidentally switch BIOS, and I'm still on the OC BIOS. What occurred to me is the Asus Z570-P I'm using has the chipset fan directly in behind the GPU. The fan is completely covered by it. Could it be the chipset overheating or perhaps a dodgey 16x slot?

Samuel T. Cogley

Yeah ignore my last, Furmark crashed it although it was a soft crash.

Samuel T. Cogley

Gotta love AMD cards, I genuinely think it's a lot more stable now but I'm just getting the odd AMD Gremlin with the odd soft lock and some odd behavior quitting out of games. Anyone think the chipset could be the issue? Might order a new motherboard as the design of this one is a bit silly.

BloodBath

I've convinced there was a faulty batch of these cards. No point wasting more money on other parts. It's most likely the card itself.

Samuel T. Cogley

BloodBath wrote: »

I've convinced there was a faulty batch of these cards. No point wasting more money on other parts. It's most likely the card itself.

Thanks Bloodbath. I think I'll try the new GPU. If it does the same thing at least I can swap the motherboard over FOC.