Printable View
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):-
The difference would be unmeasureable... I am sure...Quote:
Originally Posted by houlty
Don't worry.... We'll soon fix that... :DQuote:
Originally Posted by safan80
i thought it wouldnt be a problem, but thought i'd let an expert assure me...Quote:
Originally Posted by Eddy_EK
Are you going to stock EK's new 7950GX2 blocks?Quote:
Originally Posted by Petra
Hi,Quote:
Originally Posted by houlty
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.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...
Looks like Petras-Tech Shop will be the smart vendor to carry these.Quote:
Originally Posted by Petra
Quote:
Originally Posted by Petra
Then you'll have a new customer. Just make sure you get the G80 or r600 blocks in quick lol
sorry, regarding the barbs i actually meant the DD perfect seal as shown in your top picture.Quote:
Originally Posted by DVL73
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:
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
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.Quote:
Originally Posted by pawl
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.Quote:
Originally Posted by pawl
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:Quote:
Originally Posted by pawl
- 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
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?
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.Quote:
Originally Posted by pawl
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:
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 Eddy_EK
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.Quote:
Originally Posted by creidiki
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.
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.Quote:
Originally Posted by Steensen
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.
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
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).Quote:
Originally Posted by Steensen
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
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.Quote:
Originally Posted by Steensen
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 ?