Search the Community

This is the 2nd; vastly improved methodology testing round of using household things as thermal interface material. Variables have been removed and some major changes have been made, as detailed below. Voltage; In this round of testing, voltage was locked close to the stock voltage of this CPU [I5 3570K] at 1120 millivolts. At times this voltage rose to 1128 millivolts, as measured by a multimeter. Cooler; In this round, the stock Skylake cooler has been replaced by a Corsair H55, While the TIM is what we're testing, the water heating until equilibrium should be negligible, as well these tests took place back to back, so the water never cooled beyond the initial 20 heat test conducted with commercial TIM. IHS TIM; This CPU was delidded and had it's stock TIM replaced, for better or for worse, it now has the same TIM as was tested, Jetart Nano Diamond. This was just a paste I bought at a local store and is currently what is in use, it is competent. Test; This test was conducted using all 4 threads of Prime 95 small FFT, for 10 minutes, with a few minutes of before and after. This is very close to before. Results; Graphs have been provided of arbitrary CPU temps as read out by MSI Afterburner, as well as the Min/Max of the coolest and hottest core, on this CPU it was always the 3rd core, for both. Dead Components; This test did claim a victim, my [already faulty] Maximus V Formula has now been rendered very unstable but this, and will be definitely be recycled, If you want it for your wall or what ever, message me and I'll ship you the board [you pay p&p], but remember it's not really working and not repairable, I even washed it to finish the tests. Body Fluids Tests; NOPE, never again, I stopped after the 3rd one, not viable, I won't ever be doing them again. Methodology; Methodology is simple, MSI Afterburner will be used to take a screenshot of arbitrary CPU temps, Realtemp will be used to show max core temps, and throttling. Prime95 small FFT on all 4 cores will be used to generate heat. A Corsair H55 will be used as the cooler, with Push/Pull 120mm corsair fans, the motherboard is an ROG Maximus V Formula. The CPU is an i5 3570K at stock clocks, voltage locked at 1.12v (it's stock varies between 1.096 and 1.112). The RAM, Video Card, HDD and PSU are somewhat irreverent, so I won't list them. Test will run for 10 minutes OR until 2 cores thermal throttle together, 3 cores at any one point throttle, or the CPU permanently throttles to under 3GHz (stock turbo boost is 3.6GHz) The test method has been massively updated and the unrealistic 1.4volt on a skylake stock cooler are gone, in favour of this set up, with temp graphs for the 10 minutes. Actual Results CONTROL, Jetart Nano Diamond, Tests 1 and 2 Test one Pass, H=74°C, L=27°C Normal thermal compound passed fine, this CPU was delitted and uses this compound under the IHS. This is the control, repeated twice Test two Pass, H=74°C, L=28°C second test was consistent with first test. Nothing (no TIM) Pass, H=103°C, L=33°C Not as bad as I thought it would be, I unpowered the pump for a moment to see if it was working, the temp rose on the hottest core from 98 to 103, it still got to 101-103 at highes. Petroleum Jelly Pass, H=81°C, L=27°C Test was only 9 minutes, calculation error. This works well but not quite as good, but grease is sometimes part of TIM, so maybe petroleum grease is close. Bleach Pass, H=77°C, L=27°C Chlorine might be a good conductor of heat in this form, but the corrosion was visible even from this just 12 minutes, do NOT use. 24 Karat gold leaf, Tests 1 and 2 Test Tne Pass, H=77°C, L=32°C *I think I selected the blend of tests for this one* This test, while not reflected in the min/max, was better than the actual TIM in test 1, this is probably because of Gold's thermal conductivity and maluability. Average score was really good Test Two Pass, H=91°C, L=29°C I suspect I selected the blend first time by mistake. Gold work hardens, I used the same sheet, so it's possible that was also a contributing factor in the poor performance. Fairy Washing liquid Pass, H=78°C, L=31°C performed alright, not much to say. Lithium Axle Grease Pass, H=92°C, L=32°C slow temp rise was of concern, might not be the greatest. PVA Glue Pass, H=78°C, L=30°C I bet this will be fun to clean off the CPU [it was indeed a pain]...... For this reason I wouldn't advise this. AquaFresh Toothpaste Pass, H=79°C, L=33°C Classic one that get's quoted a lot, not actually that great it seems Sensodyne Strontium Toothpaste Pass, H=79°C, L=28°C* Room temperature fell substantially for a moment (I had to open a window due to glue being spill), results not affected as I waited for the room to heat back up. Visible in graph Baking Soda and Both Toothpasts Fail, H=105°C, L=27°C DNF. Room temperature is still a little lower. After almost 10 minutes, clocks fell to 2990MHz, after throttling to 3000-3200GHz, and did not rise after 2 updates. absolute fail. WD40 Pass, H=103°C, L=31°C One single core, for less than a second, hit 103, but this is an isolated incident, the coolest core idle is also the hottest under load. Thermal Pad Fail, H=105°C, L=43°C DNF. Test was attempted 3 times, each time throttling caused under 3GHz, and was instantaneous. These pads are very silicon like, more like a dry thermal pad, I'm not 100% if they are, I'll look into it. Shaving Foam Pass, H=80°C, L=29°C what can I say, these all seem to work well... Bodily Fluids..... Ew Out of three tests, two caused a CPU socket short and controlled shutdown. This result in the permanent instability at booting of this board. I actually washed it with WD40 followed by soapy water to try and remove urine and semen traces from the CPU pins, it did end up posting after fiddling, but each boot requires find the DIMM slot that actually works this time. Saliva This is the only 100% and verified test. Pass, H=81°C, L=28°C Nope, these are kinda gross, but they're only going get worse. Urine...... DNF, Lose of data urine seeped into socket pins, causing controlled shutdown, will not be reattempting Semen DNF, Lose of data ew, no. Fun to get, but caused a short in the socket after becoming less viscous... Required a complete socket clean and probably bricked this board, what was left of it, and again, ew I have been up now for close to 14 hours with this, I'm going to bed, please don't hate this too much, A hell of a lot of work, and the third test will introduce a temperature stable environment and a new CPU/Board combo, as well as different thermal interface materials to test, mostly chemicals, metals, and other stuff that I have to sign to get, maybe, it might be hard to actually test it.