c-attack Posted January 27, 2017 Share Posted January 27, 2017 Both do not fit underneath unfortunately they push into a motherboard cooler plate. The case is very modular though and I could move the motherboard down 5-10CM, but that would really squish my cable management setup. Thanks, that is what I was wondering about. Similar problem with SB800. While modular, it isn't always practical to shift it downwards. I think you are in a good temperature zone now and don't need to worry. Link to comment Share on other sites More sharing options...
skxc Posted January 27, 2017 Author Share Posted January 27, 2017 I think you are in a good temperature zone now and don't need to worry. Thanks for the confirmation! @Jackietools my highest temp with the XTU benchmark was 57C, but I don't have my 6700 OC'ed, it runs at 4.0Ghz Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Coolant temperature (H100i v2 or H115i Temp) is one of the data points to consider. When you spin the fans up, you can't cool the CPU directly. That is done at the cold plate and the ability to transfer heat from the CPU is down to the conductivity of the materials (CPU/TIM/Cold plate). The coolant then transports the heat to the radiator where it is expelled (hopefully). However, heat travels both ways across the cold plate, so your coolant temperature effectively determines your minimum CPU temperature. If performing action "X" causes your CPU temp to rise +20C, the lower the coolant temp, the lower the final CPU temp (25+20C vs 35+20C). Your coolant temp cannot be below the room temperature and that (although more specifically local case ambient temperature) determines its baseline. You will always be 8C warmer for everything in a 28C room compared to a 20C room. 29C is certainly not a terrible coolant temp, unless it is 18C in your room. However, I am still more interested in how your actual use temps work out. In stock CPU settings and with C-states/EIST enabled, you will only put out a trickle of heat at idle. The temps should be good regardless. Whether they stay way when stressed and if the coolant is able to come down in temp after are more telling questions. In the original screen shot, the coolant temp was 37C and that is much different. SKXC had his radiator above the frame. I have not looked at the case manual nor do I have it for inspection. Is it necessary to put the fans or radiator above the frame? From the pictures, it looks like both would fit underneath, but that is often misleading. That was "skxc" who submitted the photo of coolant temp of 37. Mine is pretty steady at idle around 29-30. Exactly how should I set up the C-staes? Of course EIST is enabled but should I leave C state on Auto? Any help with these configs appreciated. As far as install "SKXC" had his radiator on top of the case. When the plastic top cover is put back on it would not allow much clearance between fans and cover to dissipate the heat. I believe the correct way is, as I have mine and "skxc" has now changed his is with the radiator mounted to underside of case with fans exhausting air out. Another idea is taking the CPU mount and resetting it with Arctic Silver instead of Corsair's generic paste. I don't want to be "flogging a dead horse" but I tend to be a perfectionist almost o a detriment. I really appreciate your help as well as everyones here. What about that 75 degree temp. Isn't that to high?? Link to comment Share on other sites More sharing options...
c-attack Posted January 27, 2017 Share Posted January 27, 2017 Looks like a 75C is kind of high is it not? Don't run the benchmark. It is too peaky for meaningful data and it won't move the coolant temp even 1C. Instantaneous CPU core temps are mostly dependent on the voltage applied in that moment, the type of instruction, and the thermal transition through the CPU and cold plate. Cooler settings won't make a difference and the heat generated must always pass through the CPU before it enters the cooling system to be removed. I don't want to make any generalizations about how Kaby Lake CPUs behave under synthetic stress tests. However, they are likely to be similar to SkyLake and be very rangy under the auto volt stock conditions. Instead of the benchmark, run a short 5 minute test. Before you do, go into the wrench settings and configure the display to show your 4 individual cores in the line graph. This is more useful than "CPU Temp" and can show odd behavior or sometimes mounting issues. The XTU graph will show your end CPU temps, but also note the starting and ending H100i v2 temperature. That will tell you how effective the cooler was at removing heat or whether it is 'stuck in the loop'. EDIT: I know it was skyx's 37C, but that is the problem we are trying to address. Mounting position, fan positions and speeds will affect the coolant temperature. 29C coolant temp in 23C room is not terrible, but I would like to confirm it won't be compromised under load. Tweaking your BIOS settings and overclock is something else. You don't need to change any of those at this point and it would be better if you didn't for comparison purposes. As for AS5 vs the stock paste, I don't think you will see any difference. AS5 is not exactly the top of the pile any more and there are some alternatives, but we are talking about 1-2C at most for a 95 TDW CPU. This can be done as some point down the line if you wish. Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Don't run the benchmark. It is too peaky for meaningful data and it won't move the coolant temp even 1C. Instantaneous CPU core temps are mostly dependent on the voltage applied in that moment, the type of instruction, and the thermal transition through the CPU and cold plate. Cooler settings won't make a difference and the heat generated must always pass through the CPU before it enters the cooling system to be removed. I don't want to make any generalizations about how Kaby Lake CPUs behave under synthetic stress tests. However, they are likely to be similar to SkyLake and be very rangy under the auto volt stock conditions. Instead of the benchmark, run a short 5 minute test. Before you do, go into the wrench settings and configure the display to show your 4 individual cores in the line graph. This is more useful than "CPU Temp" and can show odd behavior or sometimes mounting issues. The XTU graph will show your end CPU temps, but also note the starting and ending H100i v2 temperature. That will tell you how effective the cooler was at removing heat or whether it is 'stuck in the loop'. EDIT: I know it was skyx's 37C, but that is the problem we are trying to address. Mounting position, fan positions and speeds will affect the coolant temperature. 29C coolant temp in 23C room is not terrible, but I would like to confirm it won't be compromised under load. Tweaking your BIOS settings and overclock is something else. You don't need to change any of those at this point and it would be better if you didn't for comparison purposes. As for AS5 vs the stock paste, I don't think you will see any difference. AS5 is not exactly the top of the pile any more and there are some alternatives, but we are talking about 1-2C at most for a 95 TDW CPU. This can be done as some point down the line if you wish. Just did another 5 minute test but set fans to performance in Corsair link and highest temp 72C was only 3 degrees cooler.. Last setting was on quiet, the fans run way too loud on performance. but my coolant temp went down to 29C immediately. Why wont the fans throttle down. May have to configure them manually. I am going to run it again using your advice. Thanks. Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Just did another 5 minute test but set fans to performance in Corsair link and highest temp 72C was only 3 degrees cooler.. Last setting was on quiet, the fans run way too loud on performance. but my coolant temp went down to 29C immediately. Why wont the fans throttle down. May have to configure them manually. I am going to run it again using your advice. Thanks. Coolant temp never went above 30C and all cores showed the average same temp throughout test. Had Link set on default fan setting. What do you think? Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Disregard the last picture Link to comment Share on other sites More sharing options...
c-attack Posted January 27, 2017 Share Posted January 27, 2017 OK, good. That's all you can do. It looks like the fan and radiator placement is not a restrictive issue. You'll have to see if this changes in a negative way when extended GPU loads are present, but there is no reason to expect that it will be a problem. As for actually reducing the voltage induced CPU temp side of the equation, that is something you can address through the BIOS. Even if you are not going to overclock, setting a specific adaptive voltage often keeps everything on a tighter rein than leaving the Vcore on AUTO. If you haven't already, take a look at a solid professional review for the CPU. I found this one interesting and the stock CPU temps were higher than I would have expected. However, the Kaby's seem to clock right up without major penalties and the CPU was clearly designed for the temperatures you are seeing now. Keep in mind the hardware set-up in that review is an open bench, and as such, the coolant temp won't have the usual 4-6C penalty we all do with it in an actual case. Therefore, add +5C or so to their mark to estimate a case environment. You might look for an end user overclocking guide that will give you some idea of what other people see with various cooling solutions and voltages. It looks like the usual thread has already started at Overclock.net. Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Thanks. So the way it looks now my cooler is functioning properly and I am within safe limits? Just wish it ran cooler like my old 17-3770K but guess it's mot meant to be. You have been very helpful and Ill check out those articles you sent. Thanks again. I read the article ( i am also a Tom's Hardware member. Good stuff) and according to the article (please correct me if I am mistaken) the temps seem to be in the limits of their tests correct? I hoe so as I am going crazy thinking what I can do to cool this down. I do have room for a 140 MM intake fan at the bottom if that would help. Thanks Link to comment Share on other sites More sharing options...
c-attack Posted January 27, 2017 Share Posted January 27, 2017 Jackie, what did you decide to do with the radiator fan control point? Did you opt for the pump control or the motherboard? If you went for pump control, you will likely want to make a custom fan curve. The included default curves were designed for a 20-23C room and most us get outside that window at some point during the year. You can simply slide the whole scale up a few degrees or come up with something that suits you. This is the preferred means of control and it offers more consistent fan speeds compared to CPU temp control. Fans should be linked to H100i v2 temp group. If you go from the board, that is fine as well. There is no negative side to the cooling performance, however the inevitable side effect is frequent changes in fan speed as they try to keep up with the very dynamic CPU temp. You don't need the fans to do this and water coolers can absorb a fair amount of heat before CPU temp (and coolant temp) is affected. If it comes down to a choice between MSI software and Link, I understand that and would probably opt for board control for now as well. Hopefully, MSI has some built fan delays to prevent the frequent fluctuations. Unfortunately, while present on my Asus X99, it does not work for the CPU temp based control. Link to comment Share on other sites More sharing options...
c-attack Posted January 27, 2017 Share Posted January 27, 2017 Thanks. So the way it looks now my cooler is functioning properly and I am within safe limits? Just wish it ran cooler like my old 17-3770K but guess it's mot meant to be. You have been very helpful and Ill check out those articles you sent. Thanks again. I read the article ( i am also a Tom's Hardware member. Good stuff) and according to the article (please correct me if I am mistaken) the temps seem to be in the limits of their tests correct? I hoe so as I am going crazy thinking what I can do to cool this down. I do have room for a 140 MM intake fan at the bottom if that would help. Thanks More case fans well help case ambient temperature, but likely nothing for CPU temps, particularly since I think Kaby Lake (and all other recent CPUs) are limited by voltage temperature and not how much heat the cooler can dissipate. I am not sure exactly what you are asking about the temps in the article. However, most of those tests are extreme and do not reflect real world use. The one thing that did catch my eye was higher than expected CPU temps for gaming (Watch Dogs 2). Obviously, gaming temps are quite dynamic but it was higher than I would have guessed. Perhaps it is because Intel has made more of the core power available for those tasks whereas on older CPUs, the core would lie dormant. I do think your temps are similar to everyone else's, but I am hopeful you can knock them down a little by setting a specific Vcore. What that value should be, I have no idea. The overclock arena may be able to give you a target that is successful for them. Link to comment Share on other sites More sharing options...
Jackietools Posted January 27, 2017 Share Posted January 27, 2017 Jackie, what did you decide to do with the radiator fan control point? Did you opt for the pump control or the motherboard? If you went for pump control, you will likely want to make a custom fan curve. The included default curves were designed for a 20-23C room and most us get outside that window at some point during the year. You can simply slide the whole scale up a few degrees or come up with something that suits you. This is the preferred means of control and it offers more consistent fan speeds compared to CPU temp control. Fans should be linked to H100i v2 temp group. If you go from the board, that is fine as well. There is no negative side to the cooling performance, however the inevitable side effect is frequent changes in fan speed as they try to keep up with the very dynamic CPU temp. You don't need the fans to do this and water coolers can absorb a fair amount of heat before CPU temp (and coolant temp) is affected. If it comes down to a choice between MSI software and Link, I understand that and would probably opt for board control for now as well. Hopefully, MSI has some built fan delays to prevent the frequent fluctuations. Unfortunately, while present on my Asus X99, it does not work for the CPU temp based control. Right now I am running the cooler fans thru Corsair Link. Going to leave it for a while this way and see how it doed when I fire up flight simulator. Link to comment Share on other sites More sharing options...
Jackietools Posted January 24, 2019 Share Posted January 24, 2019 It’s been 2 years and I’ve finally had enough. I installed a Noctua DH-15 and CPU is super cool. In the last 2 days I checked and re installed the CPU plate using Arctic thermal paste. I reversed fan direction at top of case. I took out empty hard drive cages all to no avail. Heavy gaming and Intel Tuning Utility topped out temps of at least 100C with thermal throttling. Noctua only enters the 60’s and at idle about 31C. I think the problem is definitely the CPU plate not making full contact with the cooler plate. I tried resetting it 3 times and each time I removed it it looked like paste was not compressed as it should be. I have read about others experiencing this problem. Don’t know why I waited 2 years but glad I put the Noctua in. I’ll try to send pics. Link to comment Share on other sites More sharing options...
Jackietools Posted January 24, 2019 Share Posted January 24, 2019 https://photos.app.goo.gl/nHQDPKadnyP5X2M96 https://photos.app.goo.gl/WKQ9MX5CFxQD6LTS7 https://photos.app.goo.gl/e1Uw2fmpmBtXPVML8 https://photos.app.goo.gl/Q4Pgk88o5mzSXMV97 Link to comment Share on other sites More sharing options...
Jackietools Posted January 24, 2019 Share Posted January 24, 2019 Hopefully these pics uploaded Link to comment Share on other sites More sharing options...
D34DMetalPC Posted January 24, 2019 Share Posted January 24, 2019 Yikes use the rad as intake on the front of the case. Link to comment Share on other sites More sharing options...
Jackietools Posted January 24, 2019 Share Posted January 24, 2019 Yes that was an option I tried. It idled about 5 degrees cooler but under load still topped out at 100C. That is why I suspect a problem with CPU plate contact. Photo shows Noctua DH 15 installed Link to comment Share on other sites More sharing options...
Jackietools Posted January 24, 2019 Share Posted January 24, 2019 Didn’t matter anyway. Coolant temp always stayed around 45-50 degrees under full load so coolant was getting cooled or not enough heat was being transferred from CPU to water jacket Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.