Reviews
PC Peripherals and Accessories
Spire Diamond Cool LGA775 CPU Cooler
| Spire Diamond Cool LGA775 CPU Cooler |
|
|
|
| Written by Gray Rentz | |
| Saturday, 21 January 2006 | |
|
Spire Diamond Cool LGA775 CPU Cooler Recently we upgraded our test rig to use a LGA775 processor (instead of the Socket 478 we were using before), and thus we contacted Spire Coolers and received one of their LGA775 Coolers for review. The model we received is the Diamond Cool SP507B7-U. This model is very interesting looking with its copper heat sink, heat pipes coming out of the one side, and a large fan with a UV Reactive casing. Right out of the box the Diamond Cool feels very well made and is quite hefty. Being all copper, the heat sink is very good looking, plus should yield very good performance. Little touches finish off the good looks as well with things like the UV reactive fan case, and the power cord which is wrapped. In the package you will find the Diamond Cool itself, the plastic plate that goes on the underside of the main board, a tube of thermal paste, and a small installation guide. To start out installing the Diamond Cool, you will want to remove your current fan/heat sink setup, and then remove the main board screws. This is the main thing that I tended to not like very well about the Diamond Cool, you must remove your main board or get it out slightly so that you can slip in a X shaped back plate that the Diamond Cool screws into. While this is common among coolers, it is still an annoyance that is not on the stock Intel cooler, which uses a twist lock mechanism that does not require you to remove or get under the main board. Backplate: Once the Diamond Cool back plate is behind the main board, you will want to apply the thermal paste. Next, you line up the four corners and screw the screws into the back plate. The screws work as thumb screws, but also have a phillips top so you can use a screw driver to finish putting them in. They also have springs so that they keep an even pressure on the processor. A few things to note about the Diamond Cool is that it is a bit larger than the stock Intel cooler. It is a bit higher and sticks out from the edges a bit more. This is something to consider if the fit in your case is already very close with the stock cooler. Also, the Diamond Cool fan is not controlled by the LGA control circuit that the stock fan uses (this is the 4th pin on the power connector), this circuit is used to control the fan speed based on temperature, but as you will see later, it is not a downfall that the Diamond Cool does not have this. Diamond Cool and Stock Intel Cooler: Diamond Cool sticks out over the edge of our test rig main board: To test the Diamond Cool vs. the stock cooler we used a Intel 3.4Ghz w/HT Processor on a PCChips M957G main board along with 1gb of DDR333 memory, and nothing else installed on the main board. We had this setup on our bench out of a case to take the whole case out of the equation. The power supply and hard drive were moved away from the main board and processor to prevent those from interfering as far as temperature goes. Picture of the test rig with stock Intel mounted: Then, using Speedfan, we took the temperature readings. To provide a even load and but the CPU under maximum strain we used prime 95's torture test. Since this is a HT machine, we started off by running one instance of prime 95 (which provides 50% load) and then stepped it up to two instances of prime 95 (which provides 100% load). This is how we came to the results/temps/graphs you will be seeing below. First up was the stock Intel cooler. After idling for about 10 minutes after startup, the temperature stabilized at 120F. One instance of Prime 95 was then started and temperatures were taken after 5 minutes (160F) and 10 minutes (162F). Next a second instance of Prime 95 was started to supply full load to the processor, and again temperatures were taken after 5 minutes (163F) and 10 minutes (163F). Finally after a two minute cool down (both Prime instances stopped) 133F. This temperature graph shows the machine idling and then going up to
the load of one instance and then two instances of Prime 95: and in Celsius: Here we see the graph of the cool down period. Notice how from when the
instances are stopped, it makes a nice sweeping cool down: and in Celsius: Next up is the Diamond Cool. After idling for about 10 minutes after startup, the temperature stabilized at 116F (4 degrees less than the stock Intel cooler). Test rig with Diamond Cool mounted: One instance of Prime 95 was then started and temperatures were taken after 5 minutes (152F) and 10 minutes (153F). Next a second instance of Prime 95 was started to supply full load to the processor, and again temperatures were taken after 5 minutes (153F) and 10 minutes (154F). Finally after a two minute cool down (both Prime instances stopped) 117F. This temperature graph shows the machine idling and then going up to
the load of one instance and then two instances of Prime 95: and in Celsius: Here is the graph of the cool down period: and in Celsius: From looking at the graphs and readings the Diamond Cool seems to stay at a more constant temperature and have a more consistent heat up and cool down. It also seems to top out a full 10F less than the stock Intel cooler. Not bad considering to my ear it also seems to be much quieter. The consistent heat up and cool down also would most likely put less stress on your processor as it is not changing temperatures so often. To get an even better idea of the comparison, here is a rough overlay of the graphs (in Celsius). (NOTE, the Diamond Cool (blue line) has been pushed a bit ahead in time of the stock Intel (green line) on the graph so that you can see and compare them. This makes it look as though the Diamond Cool takes longer to heat up, but that is incorrect, it is just because it is pushed ahead for easier comparison)
Diamond Cool: Blue Line The graphs confirm the Diamond Cool topping out lower, and also cooling down quicker than the stock Intel cooler. You also can see how the Intel cooler jumps around a bit more in temperature at full load where the Diamond Cool stays steady. Also, the Diamond Cool returns to a lower idle temp quicker than the Intel cooler. So from here I would say that the Diamond Cool is a decent replacement for the stock Intel cooler and performs quite well, plus is very quiet (quieter than the stock Intel). It also features some great looks that if you have a case window, would make quite a neat statement. Now there is one small thing that I would like to mention, on LGA style boards, the capacitors and resistors are placed rather close around the processor. This is so that when using the stock Intel cooler (which is round in design), air is pushed over the capacitors and resistors to help keep them cool (as they tend to put out a lot of heat on their own). With the Diamond Cool however, the air is pushed only out of the two sides and thus does not blow over some of the capacitors and resistors, plus it is heated more as all the air is being pushed out of the two sides. This put a bit more heat on the capacitors and resistors around the processor as can be seen in this pictures of a infrared thermometer pointed at the capacitors and resistors around the edge of the processor:
The same heat is put on the memory as well: Now, this may seem very hot, but it is only about 5F-10F hotter than with the stock Intel cooler, but it is something to consider. It also depends on your board for where the capacitors and resistors are placed and if the hot air discharge blows over them. Overall I feel that the Diamond Cool is a good upgrade from the stock Intel cooler. It offers a good performance increase, looks great, and is quiet! There really were no downfalls to the Diamond Cool besides the bit of extra heat that the surrounding capacitors and resistors received, but that is common for many cooling systems like water cooling and should not have any serious effects (for the few degrees hotter that the capacitors and resistors run). Thanks to Spire Coolers for the Review Sample!
|
|
| Last Updated ( Saturday, 21 January 2006 ) |
| < Prev | Next > |
|---|