To the unlucky burnt Athlon CPUs

We invite you to join our investigation of contemporary high-performance CPU cooling systems in order to figure out, which cooler is the best. Now and then new faster Athlon and Pentium processors crop up, each of their extra hundred megahertz contributing to the global climate warm-up and sparking our interest to CPU cooling matters. In spite of the cold weather outside, burning processor hearts of the latest technologies offsprings heat up our souls, and not just figuratively but in real degrees. So, this is not a purely overclockers' headache to take care of PC stability providing an adequate cooling system. This is a universal PC users' duty.

Of course, we cannot supply an overall material covering every cooler model you may get hold of nowadays. We doubt that there are that many enthusiasts buying every novelty that appears in stores. This roundup was written for those who decided to buy a new PC, to upgrade or overclock the CPU, who are not satisfied with their old cooler or who would simply like to keep up with the latest cooler news.

Testbed and Methods

In this roundup we stuck to the same methods as in the previous part called Roundup: Choosing the Best Socket A Cooler, so that we could easily compare the results. For this test session we took ViewStation 701 PC case with a 250W power supply unit from Shark Technology, ABIT KT7E mainboard based on VIA Apollo KT133E chipset (VT836E/VT82C686B) and used no accessory coolers. In order not to damage the CPU, we took a Thermaltake protection cover which didn't produce any notable influence on the temperatures.

We ran all the tests with AMD Athlon 1400MHz CPU with an unlocked multiplier. It worked in turns at the nominal frequency and lowered (1000MHz) clock frequency with 266MHz FSB and constant Vcore. After the temperature leveled out as the CPU worked in idle mode, we started BurnP6 and left the PC working for another half an hour. As the practice showed, it was quite enough to define the top temperature of the CPU. To get the charts we used the data supplied by MBProbe hardware monitoring program version 1.30. As a thermal interface for all the coolers we took KPT-8 thermal paste. The indoor temperature rested within 19-21 degrees Centigrade, the atmosphere pressure and humidity remained unchanged.

Testing Participants and Results

Well, let us have a look at the testing participants one by one. Their technical parameters and dimensions are available in the summary chart. The first group of coolers we tested is famous for its tender names: HTPB, HTEB and HTEBS. All of them are made by the same manufacturer, Apotheosis Honor, and have a highly inconvenient fastening mechanism: the lever is too small and there is no hole for a screwdriver. By the way, the first edge damages on our processor (and the last ones) appeared after installing and removing these particular coolers. HTEB is an aluminum case with a copper cylinder (33.5mm in diameter and 8mm high) embedded into it for the sake of better heat dissipation. The heatsink edges are rather thick. Below you can see what it looks like and what we obtained during the tests:

    

In the very beginning of the tests, shortly after Windows booting is complete, the CPU is quite warm (it's warmer than 40 degrees Centigrade). As the PC keeps on working in idle mode, the temperature sets at about 30^oC. After that we launch BurnP6, which brings about a sharp temperature increase freezing at its peak. After that the return process begins.

If you combine HTEB heatsink with a 1cm higher fan, you'll get HTEBS cooler. There are no other differences between them, except the irritating noise, HTEBS produces.

    

A cylinder Orb-like case and the hugest copper cylinder at the foot (the same 33.5mm in diameter and good 23.5mm high) - here are the distinguishing features of HTPB.

    

The next group of testing participants includes two coolers from Titan Computer Co., Ltd. They are intended for Intel Pentium III Coppermine, Tualatin, Celeron, CeleronII, AMD Duron / Athlon Thunderbird CPUs with clock frequency up to 1400MHz and Athlon XP CPUs (up to 1800+).

TTC-D8T with its standard heatsink and common fan doesn't represent anything extraordinary. The only thing to mention here is an awfully badly polished heatsink footing, which had a negative effect on the test results, of course:

    

The foot of the other cooler from Titan Computer Co., TTC-D7T, is radiantly shining. Its huge aluminum heatsink gets wider at the top, which contributes to better heat dissipation and makes the device easier to install. Nevertheless, be careful: on the upper photo you can see that the heatsink foot is not square but rectangular that can prevent you from installing the cooler on a board with closely located high capacitors:

  

  

Another racers are children of a well-known manufacturer, CoolerMaster. It seems to have been inspired by "The one as known as Prince": one of the coolers has an inscription on the box saying "AAC-001 as known as DP5 6131C-A1". We hope, there will follow no rows or lawsuits if we call it just AAC-001.

The retail version of this cooler is quite easy to recognize: it has a non-elaborate but wholesome aluminum rectangular case and the box saying "Recommended by AMD". It is intended for Pentium III Coppermine CPUs from 1.13GHz and higher, Tualatin up to 2.0GHz, Celeron from 850MHz and higher, Athlon from 1.4GHz, Athlon XP up to 1.7GHz, Duron from 900MHz and higher. This cooler owes its high results to the manufacturer's particular attention paid to quality matters. Though we were a little bit disappointed with the produced noise, then we assumed that it could have happened just by mischance, since the noise level of the elder brother of AAC-001, CoolerMaster HCC-001, was pretty low.

    

Let us pass over to the other child of CoolerMaster, HCC-001. It is intended for Pentium III Coppermine CPUs from 1.13GHz and higher, Tualatin up to 2.2GHz, Celeron from 850MHz and higher, Athlon up to 1.5GHz, Athlon XP up to 2.0GHz, Duron from 900MHz and higher. Unlike the above discussed regular-looking coolers, this product is more impressive. It has a monolith cooper foot with a thin aluminum ribbing. For better heat dissipation and rigidity it has accessory holes and lugs. All this stuff is hidden into a thick housing with a fan fastened to it. This construction proved just excellent, but we would like to stress another thing here. Both coolers from CoolerMaster have won our hearts not only with their high-quality design, but also for easy installation. Have a look at the cooler photos: the long lever arm and convenient rest for your finger would allow you to install it without any screwdriver, gently and softly, with no risk to damage the CPU. CoolerMaster engineers and developers haven't invented anything new, but have merely created a wise solution, which is sure to ease the life of every user. Well done, guys!

    

The last word will be devoted to the coolest coolers from Zalman, a new but already widely known brand. To be more exact, Zalman is known for its efficient and low-noise coolers. They don't differ much in construction, but their heatsinks are made of different materials. We tested CNPS3000-Plus, made of aluminum with a copper heart; CNPS3100-Plus, a sole copper device; and CNPS3100-GP, a copper golden (!!!) plated cooler. Just look what simple and smart solution the company offers (by the way, this solution already has a number of patents). Thin metal petals the heatsink is made of are tightly pulled together at the foot and blossom out freely at the top. The flower relation has not come to us just by chance: the manufacturer itself calls this product FHS - "Flower HeatSink". Its foot is perfectly polished. FHS is attached to the socket with a tight clip. The latter has special holes for a screwdriver and a smart "Installation Tool" (a wire mechanism with a handle at the one side and a rest at the other one. It serves to lock and ease the clip. A 92mm fan is fastened not directly to the heatsink but to a special bracket holder, and the user is free to choose any of the numerous installation opportunities. Alongside with the Installation Tool, there is a syringe with thermal paste, an illustrated installation manual, and a set of screws (with one-inch and one-millimeter threading step) to have the holder fastened in any PC case.

    

  

Of course, a gigantic fan with 2800rpm rotation speed is almost noiseless and provides substantial airflow enough to cool down the CPU together with the mainboard. The nicest thing is different however: the cooler is shipped with an adapter for the cooler power supply cable equipped with a built-in resistor, which allows to decrease the rotation speed down to 1600rpm! Frankly speaking, we could hardly hear CNPS3100-GP working even in the "loud" mode (the HDD and PSU fan appeared much noisier). In the quiet mode we didn't manage to hear anything at all. However, every extra decibel is crucial for a home PC, so, we tested the coolers in both modes. Here we would like to draw your attention to something written in the manual: before installing the cooler, you are recommended to disable cooler rotation speed monitoring in the BIOS Setup and other monitoring utilities. Otherwise you'll be annoyed by hysterical beeps from the PC speaker and PC standstill, or it can simply fail to boot, because the fan is hardly working: 1600 or even 2800 rotations per minute seem just a trifle for today's CPUs.

Now we would like to provide some explanations to the latest graph, that is, to the fifth "1000MHz (no fan)" curve. The matter is that Zalman coolers have a really large dissipation area (2000sq.cm), so we made up our mind to try a fan-free variant. To be more precise, we didn't use the fan above the cooler. We fastened a regular fan (80mm in diameter) blowing the air out at the rear side of the PC case and removed the nominal cooler fan. Well, the experiment turned out a complete failure: when the CPU got over 85^oC warm, we stopped BurnP6. To be fair we should note that no one promised the cooler would stand such extreme conditions, but we were too eager to try…

Cooler Specifications Summary Chart*

* - All the characteristics are given according to the manufacturers.
** - Standard mode / quiet mode

Conclusion

We won't confine ourselves to the traditional "it's up to you to decide" commentary, though it's a good thing to do. From our point of view, the following issues are worth taking into consideration:

• As you remember, two similar fans from CoolerMaster may have different noise level. In other words, it implies that if you buy a praised "noiseless and powerful" you may actually get a "powerful but very noisy" one - it all depends on the very item you buy. Likewise, if you get an indisputable outsider of this roundup, TTC-D8T, but it happens to be an item with a well-polished foot (or you will do it yourself), you'll see that it is not so bad indeed.
• We were deeply impressed by both the exorbitant price and stylish outlook of Zalman coolers. Now we can recommend them to those who have:
  • Too much money to spend;
  • A transparent PC case;
  • Their PC located too close to a child's bed, or to a sleeping mother-in-law, or wife, or cat, or else, so the computer is not allowed to produce any noise at all.
  • Please, don't treat the numeric results we obtained as a universal truth. We surely couldn't avoid some measuring errors, which are inevitable in winter with all those drafts from wind-non-proof windows, which make it impossible to maintain a constant temperature throughout the test session. Besides, it is of great importance what CPU you have, what thermal paste you use, how the cables are stretched inside the PC case, what case model it is and even where it is located.
• All in all, all the coolers we tested this time have coped with their task successfully: they cooled down the processor. On the one hand, if they all do well, why would you pay more? On the other hand, there are too many other factors to consider. This way, our last word still will be "it's up to you to decide".