Arctic Cooling Accelero Xtreme 5970 now available!

Hope that AC is reading this and seeing their lost sales. I'd also like them to reply to me one day this millennia...

I was about to buy one aswell, but then I realised the problem they had with the older ati's HD4870x2 ... And they never did anything to fix the broken ones.

Then I investigated a bit and got to this forum, thank you everyone for posting here, I was about to order one 2 days ago until I saw this thread.

Thanks!
 
I was about to buy one aswell, but then I realised the problem they had with the older ati's HD4870x2 ... And they never did anything to fix the broken ones.

Then I investigated a bit and got to this forum, thank you everyone for posting here, I was about to order one 2 days ago until I saw this thread.

Thanks!

Any news on Artic Cooling emails or something?

I really want some new cooler for my HD5970 to take off some noise, but I can't afford a liquid cooling RL...
 
Any news on Artic Cooling emails or something?

I really want some new cooler for my HD5970 to take off some noise, but I can't afford a liquid cooling RL...


You have about a 90% chance of at least killing one DVI port if not bricking the whole card with this cooler. I would not even be considering it. I have not received any reply at all from the two emails I sent them.
 
Any news on Artic Cooling emails or something?

I really want some new cooler for my HD5970 to take off some noise, but I can't afford a liquid cooling RL...

They have one more day to reply to me before a lodge a complaint with the BBB and begin a chargeback through Discover Card. If they ignore Discover like they're ignoring me, then Discover will flat out refund my money. I have no tolerance for this nonsense. At least send me a reply saying "Sorry, we fucked up, we know about it, here's a refund" - I still won't be happy, but I won't be looking to spread the word about how shitty they are.

Hello once again,

This is my fourth attempt to get a reply from you. If I go one more day without any reply, I am going to initiate a credit card chargeback through Discover Card who will then deal with you directly, and file a complaint with the Better Business Bureau in the United States as well. I am willing and able to understand a manufacturing defect; I cannot understand and will not tolerate any lack of support.

Multiple people on the Hardforums are stating you are not replying to them, and given you are charging a fair amount for this cooler, I would expect some form of reply.

Thank you once again, and I look forward to your reply,
FOOBAR
 
I put the cooler on my Gigabyte 5970 and yep one of the DVI ports died.

But the displayport and the other DVI port works fine, a bit pissed off but not too much as I dont really use eyefinity.
 
I put the cooler on my Gigabyte 5970 and yep one of the DVI ports died.

But the displayport and the other DVI port works fine, a bit pissed off but not too much as I dont really use eyefinity.

Welcome to the club. That's six that we know of now.
 
Why hasn't the title of this thread been changed as a warning to people yet? It should be a news item.
 
hmm, AC has ignored all of my emails, but just sent me an email survey with questions about their webshop and product that i purchased.
 
It's hard for me to be patient because I've got a small LAN with some friends coming up in just a couple weeks...but lets give Arctic Cooling some time to sort this out. I know from professional experience that responding to emails can take a lot of time away from actually solving a problem. It does seem like they could be a little more responsive to some people, but how they resolve the issue in the end is what really counts.

I received an email from Arctic Cooling early this morning, and they asked for high-resolution photos of my card with the cooler installed. I just got that put together and sent off to them. Here's the link to my pics if anyone is interested: http://bit.ly/9hAcCR

Edit: I happened to notice just now that the 5970 cooler is no longer listed on their international shop, it was still there a day or two ago. Maybe it's just sold out, but hopefully they've suspended sales while they figure out what's going on.
 
Last edited:
Dear Sir,

I'm sorry to hear, that you've made a bad experience with one of our products.

To help us finding a solution for this problem it would be helpful, if you send us some detailed pictures of your VGA board with our cooler installed and without. Please provide high resolution macro pictures of your card. Of special interest would be the backside area of both GPUs, sideview pictures and if you uninstalled the heatsink already some detail shots from both GPUs.

We now have taken the product from our webshop in order to prevent further damage. We are already reviewing the situation with our Engineers and also with Sapphire who are using this cooler as stock cooling solution on their card without any problems.

Also the circumstance that you've reported, that the DVI Port is not working, but is working when using the adapter is so strange.

Other users report, that everything went fine with their boards and that they didn't have any problems.

We are currently collecting all available information to find the reason why this is happening to some customers. This is why we need your help here.

They replied finally. I sadly can provide them with no photos, but hey, I'll take a reply.
 
i finally got a reply. good to see them acknowledging the issue and that a problem exists ...

Dear Sir,

thank you for your quick response. We are currently reviewing the situation and
will inform you once we have found a solution for the problem that is existing
at some customers boards. As Sapphire is using the cooler in large scale without
any trouble, we are very worried about what is going wrong with some endusers'
board.

Also some customers report, that their installation went fine without any problems.

We will keep you updated once we found the reason for the trouble and have a
proper solution for it.

Best regards / Mit freundlichen Grüßen
 
It's not on either the International or the US webshop page so it does look like they pulled it.

Regards
Fred
 
I was contacted by Artic Cooling with the following reply:

"Dear Sir,

We are currently reviewing the situation and will inform you once we have found a solution for the problem that is existing at some customers boards.

As Sapphire is using the cooler in large scale without any trouble, we are very worried about what is going wrong with some endusers' board.

Also some customers report, that their installation went fine without any problems.

We will keep you updated once we found the reason for the trouble and have a proper solution for it.

Could you try to use the defective dvi board with a dvi to dv sub 15 connector and see if this is working for you?"


I would like to try this but I don't have the converter that is needed, so if anyone can test and report back it would be appreciated.
 
Last edited:
Thanks for sharing your photos, I've being reading this topic with interest.
In this photo:
http://www.icedesigns.com/srv2/ac5970/full/2.jpg
It looks like the leftmost pin (at left of the left gpu) is shorter than the others and it appears to bend the board.
Could you take a look and see if this pin is shorter or is this just from the picture?
It looks shorter. thanks.

It also looks like the vrm heatsink is pushed up tight to the main heatsink on the left of the board which is bending it.

Fred
 
It also looks like the vrm heatsink is pushed up tight to the main heatsink on the left of the board which is bending it.
Fred

Yes, the vrm heatsink is also bending the board.
Maybe if the pin was taller (like the others), the board would be 1 or 2 mm far away and that would not happen...
 
Last edited:
I'll know more tonight when I remove the cooler to take more pictures, but Mr.Wolf and Fred, I believe you are both correct. The cooler has 4 mounting points that go through the board around each GPU, where tension is set by the spring brace on the back of the board. In addition it has 3 standoffs (like those for attaching a motherboard to the case) which I would expect should be touching the board when tightened. Mr.Wolf commented about one of these standoffs in pic #2, the two others are on the opposite side of the board and are not easily viewed after installation (even in person) due to the lip on that side of the backplate (you cannot see these two at all in the photos).

You can't really tell from the picture, but it turns out that the screw that goes into the standoff visible in pic #2, which seems to be bending both the cooler and the board, is not actually even tightened enough to bring the standoff into contact with the board -- there is a tiny gap. The standoffs on the opposite side, which cannot easily be seen, are even further away from the board after being tightened to what I felt was secure -- where I thought the two pieces being joined had made contact and I tightened it just a bit more. I think I can take pictures from a different angle later to show this.

If you look very closely at pic #7, you can see that the VRM heatsink only contacts the main heatsink on the side that would be installed closest to the motherboard (the side that has the standoff visible in pic #2). So it may be that if the standoffs were the correct height, the VRM heatsink might not be a problem.

Thanks for the comments to help analyze this issue. I'll look into things further tonight and add some new photos.
 
So is XTREME 5970 killing 5970s just like the 4870x2 XTREME killed a lot of 4870x2s?

Is the problem because the stock mounting screws are too short or is it something else?
 
So is XTREME 5970 killing 5970s just like the 4870x2 XTREME killed a lot of 4870x2s?

Yeah, it seems so... I was about to buy one of these coolers, but then I got with this forum...

Is the problem because the stock mounting screws are too short or is it something else?

There's no answer on that yet, so we don't know why they're killing the HD5970's.
 
More pictures have been added here...same page, new pics start at #11. The first few are close-ups of the three standoffs that have been discussed in recent posts. When I was installing the cooler, I was tightening by feel as one would tighten a normal screw where one expects the surfaces being joined to make contact. So those closeups represent what happened to feel "normal" to me, at the time I didn't know that none of the standoffs were actually touching the board. I see very similar bending to what I experienced on the second photo on this page posted by geekusa earlier in this thread. I searched for similar angles on Sapphire's Toxic 5970, and the only decent angle I could find was in the OCC review (direct image link)...there I see the same bending, looks like the standoff may be touching completely, you can clearly see the thermal pad going downhill. I have no way of knowing if that is how they received it or if that is how they reinstalled it after removing the cooler for pictures.

Even with the screw mentioned above loosened, the VRM heatsink still contacts the main heatsink. Earlier I thought that might have been due to the screw into the standoff on the main heatsink because there appeared to be a bit of space between the two heatsinks on the opposite side, but now I believe that tiny gap was just due to very slight variations in the height of the fins on the VRM heatsink. So it appears that contact between the two heatsinks contributes to slight bending of the card.

I took pics of the underside of the factory cooler and the Accelero for comparison. I really don't see any good opportunities for a short with the Accelero, but maybe someone will catch something. The biggest difference in fit I see is the standoffs. I have no measurement equipment to give any numbers, but the depth between the standoffs and what contacts the GPUs is greater on the stock cooler than the Accelero, and furthermore most thermal pads on the Accelero are at least five times thicker, making most of them flush with the standoffs. That means there's no way the standoffs on the Accelero will touch the board without the heatsink exerting much greater force (and causing something to warp). Right now I can't see a reason for this...it seems either the standoffs should be several mm longer, some Loctite or something should have been provided to keep the screws in place with minimal tension, or Arctic Cooling should have provided some kind of spring-loaded screw system or something where the screw would bottom out and the spring would exert the proper force. There is no mention of any specific instructions regarding the standoffs in the installation instructions.

Lastly, Arctic Cooling asked me to take detailed pictures of the card itself and the GPUs to look for any damage. Nothing looks abnormal to me. One final note, they also asked me to try using the VGA adapter on the failing DVI port, and that does work on my damaged card as it has for a few others with this problem.
 
Wait wait...

So you're telling me if you use a DVI-to-VGA converter it works (On your killed dvi port) ??? o_O

Ok now I am shocked....
 
One thing I noticed from the photos on this review:
Sapphire Radeon HD 5970 4GB Toxic Review

The Sapphire cooler looks like the accelero cooler:
Accelero cooler
Saphire cooler

But in the sapphire cooler, we can see a small yellow tape near one of the standoffs.
It looks like a insulator, maybe you can take a look, to see if the heatsink hits any component in that part.
I can´t see it clear in your photos, because the backplate covers part of the side view.

Also there are a lot of differences from the sapphire 4GB board and the ati 5970 board, but the cooler looks the same...it's worth investigating if the cooler touches anything.
The 2 big rubber supports at the end of the board are a nice touch.
ATI 5970 2GB board
Sapphire 5970 toxic 4GB board

This can be nothing, but it was the only thing I could add.
 
Comparing the stock ati cooler with the accelero cooler:
http://www.icedesigns.com/srv2/ac5970/target14.html

We can see that one of the standoffs base in the accelero cooler looks huge in size compared to stock cooler.
This is the standoff that Sapphire insulated with the yellow tape in the 4GB model.

edit1:
It also looks like they cuted a small part of the standoff in your accelero cooler (comparing your cooler photo to accelero site photo and to sapphire photo).
I can't see this clear in the photos.

edit2:
Another difference is the area around the GPUs contact, it's lower in the stock cooler and it's also insulated with a transparent plastic sheet.
You could also try to see if there are any tall components in that area...
 
Last edited:
Also there are a lot of differences from the sapphire 4GB board and the ati 5970 board, but the cooler looks the same...it's worth investigating if the cooler touches anything.
The 2 big rubber supports at the end of the board are a nice touch.
ATI 5970 2GB board
Sapphire 5970 toxic 4GB board

This can be nothing, but it was the only thing I could add.

Actually the two boards seem to be identical if you look at component placement, its only the extra memory which causes it to look massively different.

Good spot on the yellow tape.

Fred
 
Wait wait...

So you're telling me if you use a DVI-to-VGA converter it works (On your killed dvi port) ??? o_O

Ok now I am shocked....

Yep, that's right...It is surprising but doesn't seem completely impossible...I'm sure the VGA adapter included in the card doesn't actually convert a digital DVI signal to analog VGA, it just passes analog output coming from a different pathway in the card, and connects to the DVI port to save space. rampantandroid reported the same experience here, I can't recall whether someone else may have mentioned it too.


Probably board bending due to short screws?

If bending was causing the problem, in this case it looks like it would be due to contact between the main heatsink and the VRM heatsink, or due to the short standoffs on the main heatsink...the standoffs are open (so the screw will not bottom out), and the stock screws have no trouble reaching them.


edit2:
Another difference is the area around the GPUs contact, it's lower in the stock cooler and it's also insulated with a transparent plastic sheet.
You could also try to see if there are any tall components in that area...

I'm not sure about this one...the only transparent plastic sheet I see is on the stock backplate (pic #17), which is reused. The copper part that contacts the GPUs themselves on the Accelero appears to me to be more raised than on the stock cooler, though this may not be clear in the photos.


We can see that one of the standoffs base in the accelero cooler looks huge in size compared to stock cooler.
This is the standoff that Sapphire insulated with the yellow tape in the 4GB model.

edit1:
It also looks like they cuted a small part of the standoff in your accelero cooler (comparing your cooler photo to accelero site photo and to sapphire photo).
I can't see this clear in the photos.

Good eye...I added a couple new pictures to show this area. (Note that the baseplate is not present and the card is just laying on the cooler -- no screws.) In this case, it looks like my Accelero cooler is a revised and much safer design: the cutout gives much better clearance. As you noted, the base of the Accelero standoff is also much "wider" (extends further toward the center of the card)...it does not look like this would cause any problem. There is a component or two on the surface of the board under it, but they are extremely short...even if the cooler had been tightened completely, the height of the round extension of the standoff should be enough to provide clearance. BTW, notice the thermal paste in the photo you linked to on the Arctic Cooling site? Made me chuckle... ;) (for those unfamiliar, look at the orientation vs. the GPUs on a 5970)

After doing further comparison to the Toxic cooler, I noticed something else interesting...it appears the VRM heatsink is a completely different piece between the two. In the TweakTown review (direct image link) or Bit-Tech review (direct image link), you'll see the mounting points do not extend to the edge of the cards like mine do. More importantly, if you look at the HEXUS.net review (direct image link) it appears the fins on that VRM heatsink are shorter and do not contact the main heatsink.

This led me to further examine the fit of the VRM heatsink on my card, and I found one very tight clearance issue and one almost certain undesirable contact. There is very little clearance between the fins of the VRM heatsink and the chip on the board just to the left of the power connector, but in close examination of my initial images, I am confident they were not touching. Check out that capacitor though in the last two images I added. If you offset the heatsink as much as possible in the screw holes away from the capacitor, you might be able to create enough space to slide a business card through, but by the time you put the baseplate on and tighten everything, and the main heatsink pushes down on the VRM heatsink, it seems there's an overwhelming chance it's going to be touching. What's the chance that could be the culprit?
 
stingray, wolf - thanks for all of the effort you guys have put into this so far. :)
 
Thanks for the detailed photos, they are very good.
I've measured the metal housing of some solid state capacitors and they are non conductive (from a different video card).
There is some kind of non conductive varnish painting the metal housing of the capacitor.
 
Guys I've decided to sell my 5970 and get one of the 4Gb Toxic versions. My supplier tells me it will be available in 5-10 days.

If this is unresolved still I'll take any pictures you want of it.

Regards
Fred
 
It blows my mind that forum members care more than the company. Thanks to everyone that is helping.
 
Back
Top