XFX Black Edition's and temps

Chrisroman

Gawd
Joined
May 29, 2011
Messages
692
Sharing this info that I stumbled on, for those that may not know.

I ordered a XFX Black edition last week and thought I was getting the slick looking one with dual fans, but it turns out there are different Black Editions. Mine is this model, with the single fan, external exhaust design

FX-797A-TNBC

ThumbS3.ashx


The other one, that was always talked about, and reviewed on various sites is this one.

FX-797A-TDBC

ThumbS3.ashx


Both are Black Editions, just different cooling. The dual fan design (aka DD, Double Disipation design) exhausts air INSIDE the case. This is a more efficient design, assuming your case has good exhaust airflow. But there seems to be an issue with the DD design causing the VRAM to get very hot, possibly causing stability issues.

http://forums.guru3d.com/showpost.php?p=4215399&postcount=141

Not sure if it's a design issue of the DD or the fact that the case needs proper internal airflow for this type of card cooling.

Just fyi...
 
yeah this is a good thing to bring to people's attention. see my post here in another thread where i discussed the cooling of my BEDD. i guess it's common practice for non-reference coolers to avoid direct cooling of the VRM's (case in point). the BEDD cards actually do have direct VRM cooling and people discuss this in the anandtech link given in my original post. as long as the VRM temps are within spec i'm not concerned with it.
 
All of the reviews state that these cards are a huge disappointment (the black editions I mean) and that the cooling in most cases is actually WORSE than stock. So glad I didn't those and went with full reference coolers instead.
 
All of the reviews state that these cards are a huge disappointment (the black editions I mean) and that the cooling in most cases is actually WORSE than stock. So glad I didn't those and went with full reference coolers instead.

it's a mixed bag. some reviews are getting better results and giving the card a 'gold award' etc., some are finding it cools about the same or worse. i'm curious to see what [H] thinks of it.
 
yeah this is a good thing to bring to people's attention. see my post here in another thread where i discussed the cooling of my BEDD. i guess it's common practice for non-reference coolers to avoid direct cooling of the VRM's (case in point). the BEDD cards actually do have direct VRM cooling and people discuss this in the anandtech link given in my original post. as long as the VRM temps are within spec i'm not concerned with it.

What do your VRM temps look like under load in the DD unit?


All of the reviews state that these cards are a huge disappointment (the black editions I mean) and that the cooling in most cases is actually WORSE than stock. So glad I didn't those and went with full reference coolers instead.

Hyperbolate much?

I did a google search and had a hard time finding anything negative about the Black Edition. Besides, the point of this thread was to point out that there are TWO different cooling designs for the Black Edition. And the case your using is a factor. One type of cooling design may give you better results based on your case/setup.


it's a mixed bag. some reviews are getting better results and giving the card a 'gold award' etc., some are finding it cools about the same or worse. i'm curious to see what [H] thinks of it.

Thing is, these reviews never mention the case. Not mentioning the case, fan, and airflow configuration is like comparing water cooling and not comparing the radiators.
 
Last edited:
Thing is, these reviews never mention the case. Not mentioning the case, fan, and airflow configuration is like comparing water cooling and not comparing the radiators.

Mot sites probably test them open on a bench, instead of inside a case.
 
With the dual-fan you're getting lower noise levels with slightly higher temps.

It's the same scenario with my XFX HD 6950 XXX (Dual-Fan). I get around 75C on Crysis and the fan hardly sounds different compared to idle. A stock 6950 is slightly cooler, but screams like a banshee...

Anyways I've just ordered this 7970 BE DD, hopefully it's quiet as well.
 
I've got a coolermaster HAF case, fairly decent airflow.

I'm going to go for this Black edition DD card. To use those fans for more cooling I might just turn up their speed when overclocking it. Noise should be ok since I use a headset.
 
Installing mine today in an HAF932. Great airflow in that case so I'm not worried. From what I've heard/read on other threads though, having 2 of of the BEDD cards can cause some serious heating issues with the top card, from the bottom blowing hot air into it. I'd actually like to see some temps from guys running 2 with a large, max airflow case like the HAF932. If they run just as hot as I've seen on the other threads I'll actually be a little upset that I didn't just spend a few less dollars getting the reference model.
 
Hot VRMs, not VRAM is an issue with most axial fan design coolers. These coolers are actually cheaper to make than reference designs and often lack the large heat plate that covers all the VRMs and VRAM that dissipate the heat into the body of the shroud and the heatsink to be completely exhausted out of the system by the high static pressure of the blower fan.

VRM temps may be in spec but it's actually VRM temps that limit your overclock these days moreso than anything regarding the actual core voltage.
 
What do your VRM temps look like under load in the DD unit?

i don't have any way of measuring the VRM temps. there's no software that i'm aware of that allows you to monitor VRM temps and i don't have an IR gun.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
All of the reviews state that these cards are a huge disappointment (the black editions I mean) and that the cooling in most cases is actually WORSE than stock. So glad I didn't those and went with full reference coolers instead.

LOL...yeah OK!
 
i don't have any way of measuring the VRM temps. there's no software that i'm aware of that allows you to monitor VRM temps and i don't have an IR gun.

just grabbed an IR gun from work. i'll test tonight and see what kind of temps i'm getting on the back of my card at load.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Just got mine, it's sitting here next to me at work.....so what if it runs hot? It's a hot card. In the sense that it's sexy, way better looking than all the reference model cards. The tradeoff is totally worth it..........................................................
 
Friend of mine got his BEDD and it is a beautiful card, much better looking than my black edition with the reference external exhaust. But After playing with both I have to admit the external exhaust is a better design, especially when go with 2x or 3x cards. But then agaIn, anyone to 2x or 3x configurations is probably going to water cool anyway.

That said, a either design in a single card configuration, in a properly ventilated case will perform just fine. In my HafX case I have

2x 120mm intake in rear
1x 200mm intake side
1x 230mm intake front
2x 200mm exhaust top

With this case configuration both cards cool within a few degrees of each other, but the DD version was quieter. When I disabled the 200mm side fan, temps went up on both cards, but the DD version suffered the most. And when I disabled the 2x 200mm exhaust fans on top along with the 200mm side fan the DD edition became unstable in furmark after about 10 minutes.

Case air flow need to be thought out for the BEDD,
 
Wow that sucks man, didn't they have a pciture of it when ordered? Or they must gave you that one instead? Anyway thanks for telling us, I planned on buying a new card.
sig.jpg
 
With 5870s, all 5 or 6 VRM temps could be seen in MSI Afterburner. I don't know about the 7970s.

Mine were constantly going over 125° C which is why I ditched the Arctic Cooling Accellero. Cooled the core wonderfully but had dinky sinks for the VRMs that failed miserably and the card was unstable as hell.
 
Relevant thread in the AMD subforum:

Thermal pictures on page 2:
http://hardforum.com/showthread.php?t=1666122&page=2

I'll bet you $10 that this is the reason XFX released it at launch instead of of waiting. It was the only non-reference card on the market at release so people didn't know better. None of the launch reviews of this card mentioned anything about the card temperatures so people just bought them blindly and are now going to have to deal with having a card that is not only more expensive, but runs way hotter than even a reference card.
 
i grabbed an IR gun from work and measured the VRM temps on my XFX 7970 BEDD. i did this at the card's stock settings with overdrive disabled. my case has a side fan and i pulled the side panel off and set it about 12" from the card, still pushing air in that general direction.

case with side panel removed:
IMG_0380.jpg


IR gun:
IMG_0379.jpg


Idle at Desktop
Core temp: 54 C (reported by CCC)
VRM temp: 48 C (measured)

STALKER CoP benchmark, 4th run
Max Core temp: 79 C (reported by CCC)
Max VRM temp: 78 C (measured)

Metro 2033 benchmark, 6th run
Max Core temp: 77 C (reported by CCC)
Max VRM temp: 76 C (measured)

3DMark11, standard settings free edition
Max Core temp: 73 C (reported by CCC)
Max VRM temp: 67 C (measured)

Furmark, 10 minutes
Max Core temp: 86 C (reported by CCC)
Max VRM temp: 89 C (measured)

there's too many uncontrolled variables to determine why my numbers are different than hardware.fr. all i know is that i feel better about my VRM temps at the moment. i'm going to test them after a couple hours of BF3, and also with the card overclocked which should change things quite a bit.
 
Last edited:
i grabbed an IR gun from work and measured the VRM temps on my XFX 7970 BEDD. i did this at the card's stock settings with overdrive disabled. my case has a side fan and i pulled the side panel off and set it about 12" from the card, still pushing air in that general direction.

case with side panel removed:
IMG_0380.jpg


IR gun:
IMG_0379.jpg


Idle at Desktop
Core temp: 54 C (reported by CCC)
VRM temp: 48 C (measured)

STALKER CoP benchmark, 4th run
Max Core temp: 79 C (reported by CCC)
Max VRM temp: 78 C (measured)

Metro 2033 benchmark, 6th run
Max Core temp: 77 C (reported by CCC)
Max VRM temp: 76 C (measured)

3DMark11, standard settings free edition
Max Core temp: 73 C (reported by CCC)
Max VRM temp: 67 C (measured)

Furmark, 10 minutes
Max Core temp: 86 C (reported by CCC)
Max VRM temp: 89 C (measured)

there's too many uncontrolled variables to determine why my numbers are different than hardware.fr. all i know is that i feel better about my VRM temps at the moment. i'm going to test them after a couple hours of BF3, and also with the card overclocked which should change things quite a bit.

You are aiming at the back of the PCB. The actual VRMs are on the other side. You are just measuring the heat going through the back. The thermal imaging captures much more of the heat through the material to a depth your IR gun is not penetrating.
 
Last edited:
You are aiming at the back of the PCB. The actual VRMs are on the other side. You are just measuring the heat going through the back.

just like hardware.fr did, as shown in the original post.

The thermal imaging captures much more of the heat through the material to a depth your IR gun is not penetrating.

is that so? didn't know thermal imaging did anything more than surface temps, good to know. regardless this is my only means of testing since there are no sensors for AIDA64 etc. to report VRM temps. have a suggestion for a better way?
 
Your idle temp is even higher than the 7970 that I RMA-ed. My reference 7970, at 1200/1575 w 1.22v (it easily passed furmark at 1200 at 1.17v but would sometimes spazz out when doing nothing but displaying the desktop), idles at 35-38C as reported by both CCC and Trixx (was having issues with Afterburner). On the flip side, even loading a youtube video seems to fully engage full speed and cores so that temps spike into the low-mid 40s.

I should know on Wednesday evening how my incoming BEDD does.
 
I actually think the BEDD has worse temps than the core unit it seems.
Mine is in the process of RMA because the fans rattle, but I'm actually thinking of returning it for a refund and going with something else, such as the Gigabyte OC Windforce edition.
 
Your idle temp is even higher than the 7970 that I RMA-ed. My reference 7970, at 1200/1575 w 1.22v (it easily passed furmark at 1200 at 1.17v but would sometimes spazz out when doing nothing but displaying the desktop), idles at 35-38C as reported by both CCC and Trixx (was having issues with Afterburner). On the flip side, even loading a youtube video seems to fully engage full speed and cores so that temps spike into the low-mid 40s.

I should know on Wednesday evening how my incoming BEDD does.

do you have multiple monitors plugged into a single card? i do and that raises idle temps. here's my post showing a comparison.
 
One 2560 only, but if that memory reading is accurate, then it is not powering down properly at "idle." I have had similar issues.
 
One 2560 only, but if that memory reading is accurate, then it is not powering down properly at "idle." I have had similar issues.

yep, with multiple monitors the idle core clock raises from 300 to 400 MHz, and the memory always stays at 3D clocks. it's a well known thing, i was just surprised at how much it affects idle temps.
 
With 5870s, all 5 or 6 VRM temps could be seen in MSI Afterburner. I don't know about the 7970s.

Mine were constantly going over 125° C which is why I ditched the Arctic Cooling Accellero. Cooled the core wonderfully but had dinky sinks for the VRMs that failed miserably and the card was unstable as hell.

... wouldnt the ideal solution be to use the stock base-plate for RAM / VRM cooling and your accelero for the GPU? Not sure if this is an option, or was, but i'm pretty sure it'd give you what you want / wanted.
 
Last edited:
Just thinking outloud... but if you are using bedd, wouldn't you want your side panel fans set for exhaust?
 
Just thinking outloud... but if you are using bedd, wouldn't you want your side panel fans set for exhaust?

well it really depends on the case setup and the inlet/outlet fan configuration...but in general no you want the side fan blowing air directly at the card.
 
well it really depends on the case setup and the inlet/outlet fan configuration...but in general no you want the side fan blowing air directly at the card.
Doesn't the card exhaust towards the side panel?

You'd think at least one site would've done analysis of this type of coolers by now.
 
Doesn't the card exhaust towards the side panel?

the card exhausts in essentially all directions. i think of my case as a pressure boundary considering all of the in/out fans as a functioning system.
 
Doesn't the card exhaust towards the side panel?

You'd think at least one site would've done analysis of this type of coolers by now.

I will be installing 4x 120mm fans on the side of my HAF 932. I plan to have them as intakes due to dust concerns, but I'll try with them as exhaust in a few days just to see what happens.
 
I thought I ordered the BEDD because the site I got it from didn't have pictures, but I got the non DD Black Edition. Actually kind of happy I did after seeing the VRAM temps on the BEDD. Pushed it to max on CCC with +20 and no artifacts or anything crazy thus far.
 
have a suggestion for a better way?

just to clarify Spare-Flair, this was a sincere question. if you know of a better way i'd like to try it. it sounds like i'm limited by the IR gun but if there's a better approach i'm all ears.
 
the card exhausts in essentially all directions. i think of my case as a pressure boundary considering all of the in/out fans as a functioning system.
So yeah, I would design my new setup with positive air pressure, taking care that the intake cfm > exhaust, but it makes sense that the majority of the air coming off those dual fans be evacuated by the most direct route, rather than travelling through the rest of the case. Would mean beefing up the intake fans I think, hmmm... Do they still sell canned smoke? We used to use it for air flow testing back in the day...

Ghoststalker said:
I will be installing 4x 120mm fans on the side of my HAF 932. I plan to have them as intakes due to dust concerns, but I'll try with them as exhaust in a few days just to see what happens.

Definately keep us posted on that...
 
Well I just installed the 4x120mm as an intake. I have a custom fan profile, temps maxed on the core at 80c. The fan on the BEDD was at about 82% at this speed, so that was definitely audible. On a positive note, I didn't crash the 15 minute test. Previously the only way I could pass was running the test with the fan at 100% and temps hit 84c. So I lost about 17% fan speed and 4-5c (temps just barely reached 80). This was also pulling air through a magnetic filter I have to reduce dust, which reduces CFM a little bit.

Worth noting that when I ran with my side panel completely off and the fan on the BEDD at 100% temps were at about 82c.

I will try again later with it setup as an exhaust from the side, though I don't know how that will work in my case. I currently have a 230mm front intake, 140mm exhaust, and 4x 120mm fans exhausting out the top through the H100 radiator. This will basically mean my only intake is the front, though I could reverse the 140mm in the back as well.
 
Back
Top