A7reYu Posted January 31, 2022 Share Posted January 31, 2022 I ordered the XH305i because it was on sale on Amazon and the price was actually cheaper than ordering things individually. First off, I have to say that the bending kit is trash. The hacksaw cuts like garbage and the jig scratches the tubes very easily causing me to order new ones from Corsair. No clear, so satin clear. This actually really annoyed me as Corsair uses the worst shipping company in existence, USPS. Anyways Through all of it the XD5 pump seems to have some like adhesive on the bottom and Amazon will do a exchange for the kit but honestly its a lot of work sending it back. Draining the loop, taking rads out ect. I ordered a second rad and paired them with noctua fans. https://imgur.com/a/qAkovdi My question is, the 12900K seems to hit 100c in a few seconds of Cinebench at 5.1GHZ. I don't believe the XC7 is making bad contact, I even did the washer mod prior to the loop. Under normal game loads the CPU stays around 60-70 stable and passes P95. When gaming in GOW at 2880x1620 (the dldsr 4k res) the Vram hits 64c max, core 60 on the GPU, CPU hits 75C and the coolant temp his about 41C. Is this a good temp for the coolant in a loop like this? or is my pump maybe faulty? at 3500 RPM it makes a bad noise but seems fine at every other RPM/higher RPM. Link to comment Share on other sites More sharing options...
LeDoyen Posted January 31, 2022 Share Posted January 31, 2022 you may want to remove that washer mod. reducing mounting pressure can cause the opposite effect. in fact it will do most of the time. The GPU connection looks weird in the photo, it looks like water is coming and going out from the same connection. photo trickery oooor... 41°C coolant looks okay, without knowing what your ambient temp and fan speeds are of course. If it's below 16°C in a chilly room, it's toasty and you should get those fans working. if you're in a cozy 24°C, that's fine with two rads. Link to comment Share on other sites More sharing options...
c-attack Posted January 31, 2022 Share Posted January 31, 2022 You hit 100C instantly in Cinebench but pass Prime95? I would check the Vcore being applied in CB. That seems like extra voltage being applied for those instructions. It’s pretty common for D5 pumps to have a weird resonance in a short RPM section somewhere between 2700-3600 rpm. Most D5s I’ve had over the years do this in one spot. Program your pump curve around it. You don’t need to micro-manage the pump speed anyway. Small adjustments have zero effect on end temps. Set it to something low like 2000 rpm at idle coolant temp and then have it kick up to 4000 or something above the bad noise section at 35C coolant that should only happen when gaming. You could also do the same thing by manually setting it to a fixed rpm before/after gaming. You also will want off the quiet/balanced/extreme presets on the Commander Pro. They don’t work for this or anything in particular. It’s running a water cooling fan curve using cpu temp as the control source. Go the cooling tab and hit + to create a custom curve. Graph appears below. Click on on one of the shape tools in the corner. These are meant for water cooling. Then change the sensor to Commander Pro Temp 1 or whichever port you connected the pump liquid temp sensor. Now apply that curve to all fans. That should knock them down at idle. You also can freely adjust this curve to tweak it to your requirements. Link to comment Share on other sites More sharing options...
A7reYu Posted January 31, 2022 Author Share Posted January 31, 2022 9 hours ago, LeDoyen said: you may want to remove that washer mod. reducing mounting pressure can cause the opposite effect. in fact it will do most of the time. The GPU connection looks weird in the photo, it looks like water is coming and going out from the same connection. photo trickery oooor... 41°C coolant looks okay, without knowing what your ambient temp and fan speeds are of course. If it's below 16°C in a chilly room, it's toasty and you should get those fans working. if you're in a cozy 24°C, that's fine with two rads. The photo just looks misleading. Theres a 90 going down and a 90 going up. A tube from the xd5 to the bottom rad, other port to the back of the gpu to the 90 going up to the cpu. My aio was making poor contact on the bottom of the cpu and thats why I went with the washer mod. I definitely dont think its a mounting pressure issue and majority of people having issues with the xc7/xc7 pro did the mod to fix performance issues. I think the block just isn't the greatest and I will probanly just order an EK block. I dropped my coolant temps nicely just by removing the dust covers in the Lian Li 011. Its like they were just trapping the hot air in, you could feel a huge difference of air flow after removing them. Averaging 35-36c on the coolant now. Link to comment Share on other sites More sharing options...
LeDoyen Posted January 31, 2022 Share Posted January 31, 2022 arg yea, since it's all exhaust you want those filters out ! they make a massive difference as you saw 😛 As for the EK blocks, the cold plate is machined to be slightly convex, and not flat, to conform better to the IHS average shape, so you won't have to use the washers mod to have good contact. you want good mounting pressure with those. And i'm not even sure there's a way to do a washer mod with the new blocks anyway. Link to comment Share on other sites More sharing options...
A7reYu Posted January 31, 2022 Author Share Posted January 31, 2022 All in all its sad to see the xc7/pro have a lot of issues cooling the 12900K. Yeah -5c in the loop with 50 more watts in the GPU 400 oc vs 350 also 1000mhz on the vram. Link to comment Share on other sites More sharing options...
LeDoyen Posted January 31, 2022 Share Posted January 31, 2022 well the XC7 seems to work well with most applications, but when you start to pump 250 - 300W into a CPU, efficiency starts to matter. A block that cools 4 or 5° more can be the difference between throttling or not. It could pay off to check what C-attack said about Vcore. the motherboard may be a bit liberal with it.. some push the power limits way higher than what Intel recommends. Link to comment Share on other sites More sharing options...
A7reYu Posted February 1, 2022 Author Share Posted February 1, 2022 at 1.3V pulling 240W at 5.1/4.0 in R20 the CPU hits 100C in a few seconds. I removed the washer mod with zero difference in performance. Link to comment Share on other sites More sharing options...
c-attack Posted February 1, 2022 Share Posted February 1, 2022 (edited) Not what I was expecting you to say. While a different CPU, my 10900K at 5.2x10 has a Vdroop min of 1.29v when running R20. I am using the new XC7 Pro made for the LGA 1700 socket. The 10900K pulls 288-291W during the short test and I peaked at 75C on the highest core. It's a very short test so coolant temp is irrelevant, other than as a baseline starting point. We know your coolant is not 25C warmer than my 24C. So for me that leaves two possibilities. 1) You are using the old XC7 made for the 115x series and this does have a detrimental effect on the contact. Millimeters do matter. Or 2) I have misjudged the value of Vcore on the new Alder Lake series. Every once in a while they shift the relative value of Vcore and "1.25v" is really like 1.35v on the prior CPU. I have not been following the 12900K overclocking in great detail, so if someone is aware of a value shift on Vcore, please chime in. Edited February 1, 2022 by c-attack Link to comment Share on other sites More sharing options...
A7reYu Posted February 2, 2022 Author Share Posted February 2, 2022 (edited) I believe what the issue is possibly the die of the 12900K is longer top to bottom. The way I have the XC7 positioned is with the cold plate/logo sideways as if it were for older 115x. So basically its possible the outside cores are getting warm because the IHS isn't making 100% contact with the fins on the block. I am contemplating just ordering an EK block that's designed for 12th gen that I know for a fact makes proper contact regardless of the washer mod. Edited February 2, 2022 by A7reYu Link to comment Share on other sites More sharing options...
LeDoyen Posted February 2, 2022 Share Posted February 2, 2022 you can check that easily by looking at the temperature difference between cores when loading the CPU. During a cinebench run what's the hottest/coldest temps you see? the delta shouldn't be too high, typically less than 10°C. if you have cores at 75 and others at 100, you know there's a paste or a waterblock issue Link to comment Share on other sites More sharing options...
A7reYu Posted February 2, 2022 Author Share Posted February 2, 2022 (edited) In time I will know. When I removed the washer mod it seems even at 5.0ghz the temps are now worse. I believe it was actually doing something and the spread of the paste looked good when I removed it. I ordered an EK velocity 2 and I will update this thread with results. Edited February 2, 2022 by A7reYu Link to comment Share on other sites More sharing options...
c-attack Posted February 2, 2022 Share Posted February 2, 2022 I think the problem here is the water cooling kit has the old CPU block. Corsair released a "XC7 Pro" version specifically for the 1700 socket and larger die size on the new CPUs. It has more cooling fins and a new design. You may be loosing out on contact fit and surface area. Link to comment Share on other sites More sharing options...
A7reYu Posted February 3, 2022 Author Share Posted February 3, 2022 1 hour ago, c-attack said: I think the problem here is the water cooling kit has the old CPU block. Corsair released a "XC7 Pro" version specifically for the 1700 socket and larger die size on the new CPUs. It has more cooling fins and a new design. You may be loosing out on contact fit and surface area. They did but if you look on reddit you will see people posting nothing but issues with that block. I don't even see it available anymore meaning maybe they halted production. Link to comment Share on other sites More sharing options...
c-attack Posted February 3, 2022 Share Posted February 3, 2022 I have not heard about any issues with the new Pro version. It is readily available in the US, but I don't know about supply elsewhere. You can use any block you want, but I suspect the root issue here is the older block with the wrong mounting mechanism. Link to comment Share on other sites More sharing options...
A7reYu Posted February 4, 2022 Author Share Posted February 4, 2022 On 2/2/2022 at 5:37 PM, c-attack said: I have not heard about any issues with the new Pro version. It is readily available in the US, but I don't know about supply elsewhere. You can use any block you want, but I suspect the root issue here is the older block with the wrong mounting mechanism. If that were the case my arctic freezer would have performed poorly on my strix d4 board. It has the cut outs for the 1151 socket and seems to work fine with it. The 12900K is an extremely hot chip but honestly the temps are really bad with this block vs my arctic freezer II. Just search xc7 pro reddit and you will see some complaints about that block too. Link to comment Share on other sites More sharing options...
c-attack Posted February 4, 2022 Share Posted February 4, 2022 I did and found two people with questions and no clear cause. I also own this block along with 6-7 others. Any time you are seeing a large difference between cooling methods/blocks in instant load temperature, something is wrong with the mounting. You can’t chalk 25C differences up to physical properties or design. Again you can use whichever brand you like and I’m not trying to sell you a block. I would like to figure out the root issue. There is a whole other thread of people that force mounted the old block on LGA1700, claim it “works just fine”, and Corsair is needlessly trying to sell them more stuff. I tend not to believe it and this is the type of problem I would expect. Link to comment Share on other sites More sharing options...
A7reYu Posted February 21, 2022 Author Share Posted February 21, 2022 So an update. The EK Velocity 2 completely fixed the issues I was having. The cpu at 5.2ghz/4.1 on a 30 minute r23 loop peak at 86c. Before I couldn't even run the benchmark with an xc7 without a BSOD. It 100% was not a mounting pressure issue. I would suspect more of an orientation issue of the cold plate. Its crucial the cold plate aligns with the 12th gen die. If the corsair logo is facing the bottom, you are 100% making contact with 70% of the die on the fins/more surface area of the cold plate. Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now