EK-FCX1900 mini review - is DangerDen in ... danger ?!

I was considering the option to upgrade from my trusty 7800GTX to X1900XTX. At the end, especially as we have now sweet price for those X1900 cards, I decided to give X1900XTX a try, despite the Nvidia loyalty fact. I managed to hunt down very nice Edvard König (EK-FCX1900 ) X1900 full cover waterblock (CPU+MEM) for some nice wc overclocks and to complement the card. :wink: 7950GX2 is still the king of the hill, but sadly it was outside of my price range (not to mention super expensive waterblocks for this card). I felt that package from the above is really good value for money. It should give nice performance jump for the eye candy 1680x1050 WS resolution.

I would warmly suggest Edvard König (EK) GPU water blocks. Just installed one and build quality, look and surface finish are pretty much impressive. Proper DD alternative for half the price. DD is primarily expensive (here in EU) because of the direct import from US. EK is imported from and manufactured in Slovenia (EU). My X1900 block was £49 (including high quality barbs), compared to DD £90 price (and without barbs). All in all, HIGHLY recommended.

Here are the pics:

http://62.204.32.7:10973/X1900/dev_x19_01.jpg

EK-FCX1900 (top/bottom) + naked X1900XTX + (of course) 20WGX2 - finest LCD monitor built to date. Even now, after almost half year of usage, I'm still loving it to pieces. :D Also, I have my desk back now - even for the experiments like this :) X1900XTX core was cleaned properly - look at the NEC logo reflection! Also, EK-FCX1900 does have aluminium Vreg cooler attached. This is the reason why I didn't touch the Vreg thermal strip on the card itself (white line on the pic). Memory was cleaned and then I rub the Artic Silver Ceramique - that's why it looks dirty.



http://62.204.32.7:10973/X1900/dev_x19_02.jpg

EK-FCX1900 - top side & naked X1900XTX. Really, really surprisingly good build quality of the waterblock.



http://62.204.32.7:10973/X1900/dev_x19_03.jpg

First pic is showing the chambers in the waterblock, just beneath the GPU core. It should trigger some nice water turbulence. It's always a good thing. Also, EK-FCX1900 is definitely not restrictive and it should give really low (overall) pressure drop. Other two pics are - BLING! Very nice mirror-like surface finish of the core/memory pads. Shiny & reflective. No complains here. Some minor milling marks are present, but nothing to worry about as finish is already above my expectations. Thumbs up for EK!


X1900XTX preparation and EK-FCX1900 installation:

http://62.204.32.7:10973/X1900/dev_x19_04.jpg


As you may notice from the first pic, I always rub the heatsinks (core/mem pads) with the thermal paste (AS Ceramique in this case). It should fill nicely those microscopic holes and imperfections and additionally improve thermal transfer. Such fine layer is the reason why waterblock pads are now appearing discoloured. On the card itself, proper (not thick) layer of the AS Ceramique paste was applied to GPU core and memory.

Next two pictures are showing the final result - X1900XTX with attached EK-FCX1900 :) Perfect! and nice looking too. My usual practice is that I'm mounting the block two times. First time is the test to verify if waterblock (core/mem pads) are making the proper thermal contact with the core/mem on the card itself. It's simulation, as I don't like surprises than card overheat because of this. In this case, I had perfect marks on the waterblock from the core and 8 memory blocks (even text from the core and memory was slightly printed on the waterblock!). It's again showing us that waterblock build is uniform and precise. As simulation was OK, I reapplied the thermal paste and prepared the waterblock again (as in the first pic) and final attachment was next.

Few things to mention. My block version is prepared for crossfire. That means that fittings (barbs) can be used on both side of the block (up and down)! Apart from the crossfire compatibility, it does offer significant advantage, help and flexibility for the proper routing of your tubes. Block is heavy, that's around 580g of pure cooper baby!, but it doesn't bend the card ... when fitted properly and without the too much pressure. I was scared at first that this would happen, but no .. it's just fine. Just remember: don't use excessive force for those screws (like with any waterblock). You will know when it's time to stop.

I already mentioned that waterblock surface finish is very good, so you will not have much work apart from the cleaning and thermal paste application. My usual practice for the final preparation is: polish->artic silver tm remover->artic silver tm purifier->rub the thermal paste (as ceramique or pure as).

Here is my "cleaning kit":

http://62.204.32.7:10973/X1900/dev_x19_clean.jpg

It never let me down :) Even for demanding tasks, involving hugely stained, corroded, tarnished and oxidated cooper blocks.


For some live action of the waterblock, check out this thread:

http://www.xtremesystems.org/forums/...d.php?t=114624

Well Done! I also did the switch from a trusty 7800GTX from Asus to an Saphire X1900XT :) and i must say that i was firstly disapointed about the drivers but i got used to it... and i was also amayzed by the perfromance gain from that 7800GTX to that X1900XT which is all i was asking for. But i only run my games at 1280 x 1024 for now.... and use 6x AA and 8XF depending of the Game ..of course... but most of the time it's thoses settings. I bet that if i had a 1600 x 1200 (or higher res) monitor i could use 4x AA insetad of 6x AA or maybe even less, right ?

Heres what i did whit my X1900XT ;) :

http://images6.theimagehosting.com/IMG_2146.23a.th.jpg

Very well done!
Thumbs up! ;)

Quote:

---

Originally Posted by Eddy_EK

Very well done!
Thumbs up! ;)

---

Eddy, would it be possible to see the Delrin version of this block - I'm considering very much buying one, but just wanted to see one in the flesh...

creidiki got a block in delrin...

The Delrin top is extra http://ekwaterblocks.com/shop/index.php?cPath=21_47

Those EK Full cover blocks are pure watercooling :banana::banana::banana::banana: :p:
Great price, awesome looks, excellent quality and finish

where did you buy it from may i ask... as i currently have a mcw55, but dont want the need for a flan blow on the ram and voltage regulator...

nice little review btw...

You can buy directly from Eddy at www.ekwaterblocks.com or many other resellers of the blocks like www.over-clock.com. Depends on where you are in the world! :p:

Allways an eye pleasure watching EK's full cover block... :clap:

I don't think dangerden is in danger since EK is in europe and DD is in the USA. although too bad EK isn't in the USA.

He has low shipping costs however... ;)

i love my EK block...

cost £35 less than a DD Tyee when it came out, and it looks nice IMHO...

I can't install the block yet... (FC-79) but that's due to several reasons.... one of them is time ;)

I just cleaned it with ketchup LOL... shiny like... ehrm, well, something that's very shiny :p:

And for the price... you'd definitely NOT be buying DD blocks...

Quote:

---

Originally Posted by Garrett

I can't install the block yet... (FC-79) but that's due to several reasons.... one of them is time ;)

---

Yeah, tell me about the time ;) My EK waterblock was on my desk for two freaking weeks, just sitting there and starring at me. It was unbearable :eek:

I have installed the block now... very nice indeed... I didn't bother checking contact by mounting the block twice, I've installed tons of waterblocks over the years now and so far all contact was good.

I screwed all the screws in about equally and I could see small amounts of ceramique coming 'out' of the corners of the bga mem chips and gpu, this must mean that the force of the block pressing on causes the small ceramique residue to 'exit' via the sides of the components cooled by the block :)

I no longer have artifacts while running the memory, can even run at 800 now, before at 750 I'd get artifacts, must purely have been a heat issue...

You made me reopen my case to have a look! Very nice photos.

Quote:

---

Originally Posted by Garrett

I have installed the block now... very nice indeed... I didn't bother checking contact by mounting the block twice, I've installed tons of waterblocks over the years now and so far all contact was good.

---

I'm procedural and detailed (in life) and I guess that because of such karma I need double the time to install anything. ;)

Quote:

---

Originally Posted by Garrett

I screwed all the screws in about equally and I could see small amounts of ceramique coming 'out' of the corners of the bga mem chips and gpu, this must mean that the force of the block pressing on causes the small ceramique residue to 'exit' via the sides of the components cooled by the block :)

---

Usually, good indication that pressure is good and equal ;)

Quote:

---

Originally Posted by Garrett

I no longer have artifacts while running the memory, can even run at 800 now, before at 750 I'd get artifacts, must purely have been a heat issue...

---

That's the reason why I like full cover blocks, as memory overclocking is usually much more exciting :D

I used to think that DDR3 didn't get so hot as DDR1 mem on videocards... omg I almost burned my finger when touching those mems while gaming... I quickly removed my finger again, I guess it was 65c or hotter :eek:

With this gorgeous EK block I know 100% sure that THOSE temps won't occur anymore ;)

Hi there :)

What's your average vreg temps during the full system load (3dmark06 or something else which is really utilising all those pretty pixel shaders) :) Here, it's in the range 63*C - 66*C (as reported by AtiTool) and my case is not very well ventilated (ambient temps are bit high). Hopefully, it's OK, Also, I guess that EK-FCX1900 is not designed to cool those 7 black voltage regulators, with the "Pulse" logo on top - just slightly left from the small ones? Reason why I'm asking this is because there is very little clearance between the EK alu vreg cooler and them, so I guess that block is not design to cool them and and the same time I hope that there is enough clearance so that they are at least ventilated slightly (by ambient air). With the stock ATI cooling, they are not passively cooled at all.

Quote:

---

Originally Posted by DVL73

Hi there :)

What's your average vreg temps during the full system load (3dmark06 or something else which is really utilising all those pretty pixel shaders) :) Here, it's in the range 63*C - 66*C (as reported by AtiTool) and my case is not very well ventilated (ambient temps are bit high). Hopefully, it's OK, Also, I guess that EK-FCX1900 is not designed to cool those 7 black voltage regulators, with the "Pulse" logo on top - just slightly left from the small ones? Reason why I'm asking this is because there is very little clearance between the EK alu vreg cooler and them, so I guess that block is not design to cool them and and the same time I hope that there is enough clearance so that they are at least ventilated slightly (by ambient air). With the stock ATI cooling, they are not passively cooled at all.

---

Hi!
It would help if you apply some thermal paste where alu and copper part are together.

Let's see which retailer or manufacturer is smart enough to license or carry these blocks in the US.

Any way to make these blocks more Future Proof as in not having to buy a whole new block for every new gen of vidcard that comes out?

Quote:

---

Originally Posted by Eddy_EK

Hi!
It would help if you apply some thermal paste where alu and copper part are together.

---

Hi Eddy. You see ... I thought about that when I was testing (and mounting) the block and I was convinced that thermal paste was already applied on that joint between the alu vreg cooler and cooper part. Actually, alu vreg cooler was already attached (stock set-up) when I received the block and this is the another reason why I didn't' verify this :(

Anyhow, as now I have to dismantle everything in order to fix this, I think that I will not bother as vreg temps are not that high. However, other users should acknowledge this important hint, as it makes perfect sense. Heat from the vreg chips will be properly transferred via alu vreg cooler to the main cooper part if you use the proper layer of thermal paste in between ;)

Quote:

---

Originally Posted by DVL73

I thought about that when I was testing (and mounting) the block and I was convinced that thermal paste was already applied on that joint between the alu vreg cooler and cooper part.

---

You got the point!

Just to clarify this for other users. If you observe the first picture under the "X1900XTX preparation and EK-FCX1900 installation" section (above) or second picture from the top, you may notice the two screws where the alu vreg cooler is attached to the cooper part. This is the joint. Just unscrew the alu vreg cooler, apply the thermal paste properly on that part and screw back again. Easy. :)

couldnt resist so i just bought this from oc-uk, but was just wondering if you noticed that when you put the standard barbs in the thread protrudes into the water route/cavity by 1 or 2mm, and was wondering if this would cause a restriction in flow???

Im new to WC so not too sure myself... i got round it by putting a some DD high-flow barbs o their which have a bigger rubber washer... its just something i noticed though...

i did a little paint work to show you what i mean (hopefully):-

Quote:

---

Originally Posted by houlty

couldnt resist so i just bought this from oc-uk, but was just wondering if you noticed that when you put the standard barbs in the thread protrudes into the water route/cavity by 1 or 2mm, and was wondering if this would cause a restriction in flow???

---

The difference would be unmeasureable... I am sure...

Quote:

---

Originally Posted by safan80

I don't think dangerden is in danger since EK is in europe and DD is in the USA. although too bad EK isn't in the USA.

---

Don't worry.... We'll soon fix that... :D

Quote:

---

Originally Posted by Eddy_EK

The difference would be unmeasureable... I am sure...

---

i thought it wouldnt be a problem, but thought i'd let an expert assure me...

Quote:

---

Originally Posted by Petra

Don't worry.... We'll soon fix that... :D

---

Are you going to stock EK's new 7950GX2 blocks?

Quote:

---

Originally Posted by houlty

couldnt resist so i just bought this from oc-uk, but was just wondering if you noticed that when you put the standard barbs in the thread protrudes into the water route/cavity by 1 or 2mm, and was wondering if this would cause a restriction in flow???

i did a little paint work to show you what i mean (hopefully):-

---

Hi,

I know what you mean, exactly ;) That waterblock picture was not from the final installation. It was just there, well, for the show off :) Anyhow, yes, barbs are protruding slightly (on the cooper side) and they are not affecting the flow BUT more importantly it's fixable if you are pedantic (like me). Take a look at this picture:

http://62.204.32.7:10973/X1900/x1900_bits.jpg

You will get this with the EK-FCX1900 and you may ignore my DD G1/4 PerfectSeal barb on the left. Notice the two metal "washers", on the bottom/right. I used them in order to move slightly back the thread part from EK barbs (for the perfect fit), by placing this between the barb and the water inlet (cooper side). Also, DON'T remove the washer from the EK barb when you are doing this, as otherwise connection will leak. Here is the final installation:

http://62.204.32.7:10973/X1900/x1900_barbs.jpg

You see what I mean :cool: Plexi side is not a problem as it's deep enough. Also, as shown in the picture, this is my preferred set-up for single card version (if you have crossfire prepared block). You may notice the white stuff inside of the block, where the threads are. That's PTFE tape on the barbs/caps threads. I like to be on the safe side, apart from the fact that caps/barbs do have the o-ring seal.

Quote:

---

Im new to WC so not too sure myself... i got round it by putting a some DD high-flow barbs o their which have a bigger rubber washer... its just something i noticed though...

---

Which DD high flow-barbs did you used ? Please note that threads are G1/4 BSPP straight and you have to use the appropriate barbs. Block is coming with nice EK barbs and DD PerfectSeal G1/4 barb (as shown in the picture above) is also nice option. Threads on DD are slightly shorter, which is good.

Quote:

---

Originally Posted by Petra

Don't worry.... We'll soon fix that... :D

---

Looks like Petras-Tech Shop will be the smart vendor to carry these.

Quote:

---

Originally Posted by Petra

Don't worry.... We'll soon fix that... :D

---

Then you'll have a new customer. Just make sure you get the G80 or r600 blocks in quick lol

Quote:

---

Originally Posted by DVL73

Hi,

I know what you mean, exactly ;) That waterblock picture was not from the final installation. It was just there, well, for the show off :) Anyhow, yes, barbs are protruding slightly (on the cooper side) and they are not affecting the flow BUT more importantly it's fixable if you are pedantic (like me). Take a look at this picture:

http://62.204.32.7:10973/X1900/x1900_bits.jpg

You will get this with the EK-FCX1900 and you may ignore my DD G1/4 PerfectSeal barb on the left. Notice the two metal "washers", on the bottom/right. I used them in order to move slightly back the thread part from EK barbs (for the perfect fit), by placing this between the barb and the water inlet (cooper side). Also, DON'T remove the washer from the EK barb when you are doing this, as otherwise connection will leak. Here is the final installation:

http://62.204.32.7:10973/X1900/x1900_barbs.jpg

You see what I mean :cool: Plexi side is not a problem as it's deep enough. Also, as shown in the picture, this is my preferred set-up for single card version (if you have crossfire prepared block). You may notice the white stuff inside of the block, where the threads are. That's PTFE tape on the barbs/caps threads. I like to be on the safe side, apart from the fact that caps/barbs do have the o-ring seal.

Which DD high flow-barbs did you used ? Please note that threads are G1/4 BSPP straight and you have to use the appropriate barbs. Block is coming with nice EK barbs and DD PerfectSeal G1/4 barb (as shown in the picture above) is also nice option. Threads on DD are slightly shorter, which is good.

---

sorry, regarding the barbs i actually meant the DD perfect seal as shown in your top picture.

Also, i may go for the PTFE tape as you did yourself... my dads a pipe fitter so has miles of the stuff...

ive got my ek - x1900 running on 10mm tubing.

once id assembled the whole thing i noticed that the thread was a little long. just trimmed them with the dremel so they fit perfectly. sealed it all up with plumbing tape.

maybe it should be pointed out in the instructions that the strip of thermal gunk over the vrm's should be left there when installing the ek block? i suspect that if it were not there may be problems?

mine idles at 35 degrees in windows. i think ive seen it make it to 42 while gaming. Very impressive.

still running stock clocks cause i cant figure out how ati_tool adjusts the 3d speed. :rolleyes:

Quote:

---

Originally Posted by pawl

once id assembled the whole thing i noticed that the thread was a little long. just trimmed them with the dremel so they fit perfectly. sealed it all up with plumbing tape.

---

Are you referring to the barb thread part ? Yes, they are a bit long for the cooper side (plexi is fine). I used the metal washers, as explained above. Anyhow, Dremel is a good option too (and wonderful little device) :cool: On one of my EK barbs, there was slight misalignment (error) in one thread circle. I used my Dremel to fix and polish the problematic thread part - it was perfect afterwards.

Quote:

---

Originally Posted by pawl

maybe it should be pointed out in the instructions that the strip of thermal gunk over the vrm's should be left there when installing the ek block? i suspect that if it were not there may be problems?

---

I'm quote positive that you can use the thermal paste too, for better heat transfer. Thermal strip is fine if you don't want to bother. In my case, I decided to leave the thermal strip for the purpose of alu vreg cooler preassure "sponge". Not sure if those vreg chips are prone to cracking?! as they do look gentle. Also, applying the thermal paste to cooper/alu joint is good candidate for the instruction manual.

By the way, how are your vreg temps (as reported from AtiTool) ?

Quote:

---

Originally Posted by pawl

mine idles at 35 degrees in windows. i think ive seen it make it to 42 while gaming. Very impressive.

---

You have very nice temps there. Is that stock X1900XTX (rest of the WC loop may be useful too) ? Good option for the proper stress test of the card is that you install the ATI tray tools, configure the background monitoring graphs and fire up ,for example, 3DMARK06. When it's finished, you may review the graphs and that will show you the idle and peak temperatures.

Quote:

---

Originally Posted by pawl

still running stock clocks cause i cant figure out how ati_tool adjusts the 3d speed. :rolleyes:

---

Yeah, tell me about it. As I migrated from Nvidia, ATI overclocking and 3D speed connection was pretty much confusing to me. Anyhow, few hints from me:

- Install the latest catalyst drivers (with or without CCC)
- Disable the following ATI services: ATI smart & ATI HotKey Poller (it's responsible for switching the clocks in 3D mode).
- Install the AtiTools ver 0.25 beta 14
- Create the new OC profile and set the OC voltages. When saving the OC voltages, you may select the OC profile which will use them.
- Open the AtiTool settings and go to the 3D-Detection
- Configure which OC profile to use when 3D application is detected and set the default profile when 3D application is closed.

Quote:

---

Originally Posted by DVL73

Are you referring to the barb thread part ? Yes, they are a bit long for the cooper side (plexi is fine). I used the metal washers, as explained above. Anyhow, Dremel is a good option too (and wonderful little device) :cool: On one of my EK barbs, there was slight misalignment (error) in one thread circle. I used my Dremel to fix and polish the problematic thread part - it was perfect afterwards.

I'm quote positive that you can use the thermal paste too, for better heat transfer. Thermal strip is fine if you don't want to bother. In my case, I decided to leave the thermal strip for the purpose of alu vreg cooler preassure "sponge". Not sure if those vreg chips are prone to cracking?! as they do look gentle. Also, applying the thermal paste to cooper/alu joint is good candidate for the instruction manual.

By the way, how are your vreg temps (as reported from AtiTool) ?

You have very nice temps there. Is that stock X1900XTX (rest of the WC loop may be useful too) ? Good option for the proper stress test of the card is that you install the ATI tray tools, configure the background monitoring graphs and fire up ,for example, 3DMARK06. When it's finished, you may review the graphs and that will show you the idle and peak temperatures.

Yeah, tell me about it. As I migrated from Nvidia, ATI overclocking and 3D speed connection was pretty much confusing to me. Anyhow, few hints from me:

- Install the latest catalyst drivers (with or without CCC)
- Disable the following ATI services: ATI smart & ATI HotKey Poller (it's responsible for switching the clocks in 3D mode).
- Install the AtiTools ver 0.25 beta 14
- Create the new OC profile and set the OC voltages. When saving the OC voltages, you may select the OC profile which will use them.
- Open the AtiTool settings and go to the 3D-Detection
- Configure which OC profile to use when 3D application is detected and set the default profile when 3D application is closed.

---

looking at the tolerance to the teeny surface mount resistors next the the vreg's id say that the ati strip is also insulating them.

i alt_tabbed out of a 4 hour WOW session and the vreg's were at 61 degrees. the GPU temps were 39 and the cpu was 36. i guess the ambient temps were nice and cool ( i was still in shorts and tshirt ) - but those are still sweet temps.

the rest of the loop is basicly an eheim alphacool pump - radiical cwx block - ek 1900 - dual 120mm radiator - xspc 250mm reservoir

mine is only a vanilla x1900xt - and my powersupply is crapping out. ive got a busy wokring week next week - so il mailorder the new powersupply - and then im gonna get medevil on its asss.

it seemed that this was what i would need to do to get the 3D clocks up. but yoikes its all a bit complicated.

ran a two hour session of wow - with everything turned to the max for the video.

ran the card at 650 / 750 and got the vrms to 61 / core to 41

brilliant stuff!! can it get better than this?

Quote:

---

Originally Posted by pawl

ran a two hour session of wow - with everything turned to the max for the video.

ran the card at 650 / 750 and got the vrms to 61 / core to 41

brilliant stuff!! can it get better than this?

---

Very good. I see that you have the dual rad, so combine that with the good ambient temps and that would explain the nice temps you have. My so-so ambient temps, "restrictive" coolant, single PA160, 3/8" tubing, volt controlled DDC and plexires (as desribed in another thread) are coping just fine ... althought it could be 4*C - 6*C better if I make an effort, but not really bothered honestly.

http://62.204.32.7:10973/x1900/x1900_temp.jpg

This graph represent the 3DMARK06 benchmark session, probably most stressful session the card may have. That pause (lower temps) you see in the middle of the graph is during the 3DMARK CPU stress tests (GPU is not used).

My OC camp, for now:

http://62.204.32.7:10973/x1900/x1900_3dm06.jpg

Not too shabby :) It's going very well. Games are very smooth at 1680x1050, compared to my previous card and with the eye candy ON - especially Oblivion & Prey. Next few days I will push the card further and try to find where the limits are. 7000 is not that far also :D - although in some practical sense there is really no need for more throttle. Card is now nicely overclocked and I can safely say that now I have the performance of the X1950XT on the Athlon 64 FX-62 or Core2Duo E6600 @ 3.33GHz. Currently, I'm keeping the clocks at 725/850 for a day or two so that card can adapt to this speed. It's feeling a bit like climbing to Mount Everest, with a pause at each camp, so that card can acclimatise to the new heights :)

Only worrying thing is that 63*C Vreg temp and I'm slightly hesitant to continue until I see if I can do something about it - as it will probably affect the extreme OC. During the heavy card stress max Vreg temp is in the range of 66*C - 69*C. I was thinking to redo the EK wateblock and apply thermal paste to the cooper/alu joint and see how it goes, but not sure if I should bother if vreg is already designed to hit 100*C ... naturally. On the other hand, it may affect the serious OC :confused:

Quote:

---

Originally Posted by Eddy_EK

Hi!
It would help if you apply some thermal paste where alu and copper part are together.

---

Be careful because AS5 will act like a threadlock, dont let it get in the screw threads, I cant remove my vreg cooler from the main block anymore because the AS5 locked the screws shut - i'll probably have to send it eddy's way if I ever want it removed, me + tools often = broken hardware :lol2:

Quote:

---

Originally Posted by creidiki

Be careful because AS5 will act like a threadlock, dont let it get in the screw threads, I cant remove my vreg cooler from the main block anymore because the AS5 locked the screws shut

---

Hmmm ... interesting. I did have couple of cooper block "joint" AS5 applications many moons ago (with air GPU coolers) and don't remember the situations like that. Anyhow, I'm using the AS Ceramique which I guess is less sensitive to the differences in surface tension - resulting in such strong bound.

How are your Vreg temps creidiki ? as your joint is thermal paste - enabled :)

guys. i bought my 2x fc79 coolers from ek. yaaaay. im gonna join the EK full cover team!
/end OT

I have tried clocking My X1900Xt card, with the non CF compatible EK block, but with voltage regulator "block" mounted, and I hit an alarming 93*C om the regulators. This was with AS5 all around, så perhaps I should try putting some thermal compound between alu and cupper, to get a better cooling of the woltage regulators.

DVL73 how many turns did you have to turn each screw when mounting the block? I fasten each screw (the 4) around the GPU approx. 720° (two turns), and the 3 at the RAM approx. 900° (two and a half turn), where as the voltage regulators are turned almost 1080° (3 turns).
Whitch other heat sources do you have in the GFX loop, resulting in those temps?

It is not the numner of turns.
It is the feeling, how hard you screw the screws.
When you have the feeling the screw has contacted the pcb try to be sensitive ;)

Eddy_EK I Know, but I have that problem, that I'm afraid that I havn't tightened the screws enough, so that they will fall out, 'cause of vibrations.
More to get a hint about, how much he had to turn the screws, to se if I'm tightening them way too much.

Quote:

---

Originally Posted by Steensen

Eddy_EK I Know, but I have that problem, that I'm afraid that I havn't tightened the screws enough, so that they will fall out, 'cause of vibrations.
More to get a hint about, how much he had to turn the screws, to se if I'm tightening them way too much.

---

I see your point (and it's perfectly right)... OK, I don't remember the exact number of turns, however use the smaller screwdriver and turn the screws until you feel that you can't go further without slightly, slightly-moderate force. After that, half turn or so is usually just enough. Obviously, it's hard to describe this as it's just the "feeling" about the things. Anyhow, your card will band just slightly and this is safe & normal (especially at the GPU core part and where the memory is on the top, opposite the PCI-E connector) but don't go into the extremes. If some part of the card is over tight you will se the band and you have to relax the pressure on that part.

Now, if you are still unsure what you are doing, my advice is that you detach the waterblock from the card and verify the pressure. If the thermal paste trace from each of the memory chip and GPU core itself is properly "stamped" on the waterblock (and in the best case scenario you may see the printed text from the chip too!) ... you are fine. If not, you have to repeat the mount and check again until you find the perfect mount and pressure.

Hope this helps.

Quote:

---

Originally Posted by Steensen

I have tried clocking My X1900Xt card, with the non CF compatible EK block, but with voltage regulator "block" mounted, and I hit an alarming 93*C om the regulators. This was with AS5 all around, så perhaps I should try putting some thermal compound between alu and cupper, to get a better cooling of the woltage regulators.

---

That Vreg temp is very high. If you removed the stock thermal strip from the Vreg chips and applied the AS5 instead, it may happen that there is no proper contact between the alu-block and them, as thermal strip is much thicker. You may inspect this visually too. This is the reason why I think that it's not crucial that people remove the stock thermal strip and probably application of the thermal paste on the cooper/alu joint is more beneficial. I'm also interested to see what are the temp differences with and without thermal paste at that joint, so that I possibly redo my installation. My Vreg temps are shown above (63*C - 66*C during the heavy load) and that's without the thermal paste at the joint and with ambient temps on the high side.

Quote:

---

Originally Posted by Steensen

Whitch other heat sources do you have in the GFX loop, resulting in those temps?

---

I have X2 3800 @ 2.6 @ 1.55V and that's pretty much all. My room ambient and case ambient temps are on the high side (and that's hurting the radiator performance), radiator is also surrounded by nearby walls and furniture (they are obstructive and affecting the radiator air flow) and recent changes in my loop (new coolant and res placement) are slightly affecting the temps also, In best case scenario, if I make an effort (I'm utterly lazy nowadays and spare time is just not there), I guess that expectable temps improvement is 5*C - 7*C (if not slightly more).

DVL73 Thx for the Info.
Perhaps I'll try using the original thermal pad on the Vregs, and before mounting the block again, then I'll put some paste at the cupper/alu joint.
I'm Lazy too, so I'm (sadly for you) not going to mount the block twice, with and without paste in between the alu and cupper. Sorry.

I've used simmilar way to mount the block, so the answer must be in the thermal pad, and grease between alu and cupper.
But first, I'll need to get my GFX back from RMA.

I have finally gotten a GFX back from RMA, and I can only say, that what a difference it does, to have put thermal paste between the alu and copper blocks, and using the thermal pad, in stead of thermal grease.
Currently I have clocked my X1950XTX to 800/1143 F@H stable. :D

Quote:

---

Originally Posted by Steensen

I have finally gotten a GFX back from RMA, and I can only say, that what a difference it does, to have put thermal paste between the alu and copper blocks, and using the thermal pad, in stead of thermal grease.
Currently I have clocked my X1950XTX to 800/1143 F@H stable. :D

---

Hey Steensen, that's very nice OC. How are your VREG temps after you applied the thermal paste between the alu and cooper part ? I'm thinking to do the same, but just want to be sure if there are any major benefits - as I have to dismantle the wc loop and card in order to do this :( I have a hard time to cross the 730CPU/860MEM OC boundary for my X1900XTX ... as ATI tray tools artifacts test is complaining. It's either the card itself ... or Vreg cooling is really important factor.

When you say thermal pad, I guess that you are referring to stock thermal pad on those small Vreg chips, cooled by alu part of the block ?