Technical - Intel Pushpin Test and Review
What is a backplate worth in temperature reduction?
In 2006, Intel released the Socket 775 (T) platform featuring the Land Grid Array socket. With each retail boxed processor, Intel sold its customers a CPU Cooler in each package that utilized so-called pushpins. Each pushpin consists of two pieces; a clear retaining portion that is "pushed" through the mounting holes of a motherboard, and a "pin" that is plunged into the clear retainer to expand it - forcing it to grab the backside of your motherboard PCB for pressure. The CPU Cooler included for current desktop processors have a copper core with a radial aluminum heatsink, and come standard with a preinstalled amount of thermal interface material that doesn't even cover the contact area completely.
The pushpin installation method on these CPU Coolers is just barely acceptable when users install it properly. For those that choose to use them, these two half-solutions combine into a massive migraine when installed improperly. Take the following complaint threads for example, here, here, here. Also this one, and here, and, well we could probably spend all day reading them. In spite of this, Intel has doubled down on this solution for its Sandy and Ivy Bridge processors as well as the fourth generation, Socket 1150, featuring the Haswell line. For the stock cooler, the Intel pushpin system has now been used for sockets 775, 1150, 1155, 1156, 1366 and perhaps others. Poor contact via half missing thermal interface material and a retention system that provides "just enough" pressure results in our 3770K processor throttling back within minutes on turbo mode even in a cool open-air ambient environment. Still, many installers do use this CPU Cooler, and if you are one of them, we explain how you can drop your temperatures simply by removing Intel's evil pushpins and using a sturdy spring retention system and backplate. For our non scientific test, we chose to use the Shinetsu PCS-TC-11T-13 35x35 millimeter square thermal pad. This allows for a consistent application of TIM to the processor, where thermal grease can be subjectively installed too thick or too thin from one install to the next.
Our purpose of this test is a comparative one. We want to see what, if any, benefit there is to using a backplate instead of Intel's pushpins. The monitoring software, Realtemp, offers a calibration mode to adjust core temperatures at idle based on recommendations above ambient. With our objective in mind, we did not calibrate Realtemp since we are not interested in the actual temperatures themselves, merely the temperature differences between a Pushpin and a Backplate installation.
Motherboard: Asus P8Z77-VLX CPU: Intel I7-3770K Ivy Bridge Processor, SR0PL CPU Cooler: Factory Intel RAM: 32GB GSkill DDR 1866 HD: x1 OCZ Vertex 3 30GB PSU: Silverstone 1000W Bench: High Speed PC Top Desk Station Bios: Default settings (Normal Performance mode) OS: Microsoft Windows 7 SP1 x64 Sensor Software: Realtemp 3.70 (uncalibrated)
After installing the thermal pad and CPU Cooler, the system was set to run Prime 95 for 30 minutes. The maximum values recorded by Realtemp are shown below. Removal of the CPU Cooler produced the shown imprint on the pad. What matters not is the thermal material left on the CPU Cooler, but the imprint of the copper core on the processor. Here we can see that less than half of a circle imprint is evident. The reason for this is likely due to inconsistent and insufficient pressure exerted by our four pushpins.
Test 2: Socket 1150 / 1155 / 1156 Backplate and Spring Screws
After removing the stock cooler and cleaning our processor with Arcticlean, we reinstalled a new Shinetsu pad and removed the pushpins from the cooler. As a precaution, one should always test fit any backplate to make certain no solder joints will contact the plate when installed. On the Asus P8Z77-VLX there are two rows of leads that are relatively close, but not touching. Advanced users can choose to trim the leftover leads that the motherboard manufacturer didn't bother to. Electrical tape around areas close to solder joints is an added insurance.
Our Prime 95 torture test was repeated. Removal of the CPU Cooler shows a healthier, more circular imprint on the IHS from the Intel copper core.
The results show a consistent temperature reduction among all cores. Cores 0 and 1 show 10 degrees less while 2 and 3 show 9 degrees less.
This simple test shows how backplates can drastically improve performance when a thermal pad is used, by up to 10 degrees Celsius. In fact, mounting a cooling device to an Intel processor using any method other than Intel's plastic pushpins should be the goal of any user seeking to ensure his/her processor does not throttle back from extreme temperatures. If using an aftermarket CPU Cooler is not in your budget, the optimal configuration for using the stock CPU Cooler is with a good quality thermal grease and a backplate.