SocketA Coolers Roundup. Q1`2004

Introduction

Recently, we tested all the most popular coolers for SocketA processors:

P4 Coolers: Q4`2003
Coolers for AMD: Q4`2003
Coolers for P4: Q3`2003
Coolers for AMD: Q3`2003

This platform is already moving towards its decline which will probably start no earlier than the year 2005. It means that for at least one year ahead AMD will be supplying processors for the socket, and other manufacturers will be shipping motherboards and coolers.

As regards the coolers, with all the variety of models the user has almost no choice. Developers so far have failed to create a cooler of balanced performance (efficiency, small mass, low noise level). The main criticism is just about the noise level which is still at an unacceptably high mark for most coolers. That is why we'll be treating this parameter as equal as the cooling efficiency. We'll be bringing in all the other parameters (mass, sizes, fastening scheme, processing quality etc.) only for estimation.

Today, we are reviewing three coolers made by Spire - a new line of Igloo 2510 coolers and a brand-new Titan CU9TB/SC cooler.

Titan CU9TB/SC

Titan is among the largest manufacturers of cooling devices. Its main activity is about producing cheap coolers. But in the sector of high-efficiency coolers, the company has nothing to offer. The only product which could have pretensions for the crown was CU5TB cooler.

However, despite its low price (under $13) and the full copper radiator with a smart aluminum fan, the cooler hasn't gained popularity among the users. Its main issue was the noise level of the fan. Therefore, you could only use such a cooler with a speed regulator only, which had to be purchased separately.

Quite recently, Titan has presented a new model for the public judgment: Titan CU9TB/SC in which, hopefully, all the predecessor's flaws will be fixed.

The cooler is packaged in a big box, and thoroughly wrapped in foam-rubber. Besides, in the box there is a tub with the famous "silvery paste" (highly non-recommended for use) and a speed regulator (highly recommended). The presence of a speed regulator is a big step forward as compared to CU5TB.

On the face of it, the cooler is very nice: the copper radiator is inside a gleaming metal housing, on top of which there is a fan made of transparent plastic. The fan's fins are colored blue, and on the fan itself there is a plastic protective grid.

Note that the fastening clip has a hook on all the six teeth of the processor socket. The cramp is resilient enough, so there aren't any issues during installation. But in so doing you have to use improvised means, a screwdriver.

When it comes to the processing of the base, Titan coolers have always been among the leaders at that criterion. So CU9TB proved no exception: the base is polished to the mirror gleam (10/10 mark).

To protect it against scratches and corrosion, there is a plastic film glued to the base.

Now let's take a closer look at the fan. Its maximum rotational speed is =4200 RPM (with the minimum 1800 RPM). The air flow is 42.08 CFM(18.03CFM), ad the noise level does not exceed 36dBa (~20dBa).

In practice, the fan demonstrated excellent acoustic performance: in the normal mode, the noise was quite acceptable. As the rotational speed dropped to the minimum, the noise from the fan disappeared completely!

And the favorable impression from the cooler disappeared completely right after the very first tests. The cooler showed very poor efficiency: by 6 C worse than its predecessor, CU5TB! After the thorough search I found out the cause: it turned out that engineers at Titan followed a vicious circle towards cost reductions. In particular, the radiator fins are not welded to the base like it is in CU5TB, but glued with some sort of thermal interface (which proved to be of rather poor quality, as the results showed).

This fact made the advantages of the cooler null and void. As a result, the cooler might be of interest to only the users of weak processors who are after noiseless operation of the CPU unit.

Glacialtech Igloo 2510 (Pro)

Glacialtech presented a somehow renewed line of aluminum-copper Igloo 2510 coolers to the public judgment. Along with the generic model, the company offers a more powerful modification - Igloo 2510 Pro.

All the elements of the cooler have undergone changes: the radiator performance, patented technology for fastening the radiator, the fan performance.

The cooler radiator is made of aluminum, with a copper plate attached to the base. The processing quality is mirror-like (9.5/10)! On the base, there is a pre-applied layer of thermal interface which is not inferior to KPT-8 at quality.

The feature of the cooler is a very tight fastening clip. To put on the cooler, you've got to press upon the stop (good it's there) as hard as you can. On the one hand, it is a definite shortcoming: with careless installation there is a probability of damaging the processor core. On the other hand, it is a plus, since due to the strong pressing may seriously increase the cooling efficiency. If we consider the high processing quality of the base, then it's no wonder when upon removing the cooler I found no traces of thermal paste - it squeezed out completely (which is good).

As regards the fans, on the Igloo 2510 model there is a 70 mm fan (height = 25 mm) with the rotational speed 2800 rpm and 26 dBa noise level. The airflow = 24.3 CFM. On the Igloo 2510 Pro, there is also a 70 mm fan of 4800 RPM rotational speed. The noise level level is 35.0 dBA, and the air flow is 40.1CFM. In real operation, the noise from Igloo 2510 cooler is not an issue, which can't be said about the noise from Igloo 2510 Pro. The latter can't be recommended to the amateurs of quiet operation.

However, if to use the regulators of rotational speed, then you can get a truly quiet operation from both coolers (as low as that of Igloo Silent Breeze II). After minimizing the speed, only the most sensitive ear can grasp the rustling of air and noise from the rotor (the noise is heard only if an Evercool fan is installed).

Spire FalconRock II and WhisperRock III

A year and a half ago we tested the Fanner 5F263 WhisperRock II cooler. Since then, the exterior of the company products has almost not changed: the box design, the aluminum fan with a protective grid are the same. But in the WhisperRock III, even the fastening clip hasn't changed.

The coolers are thoroughly packaged: there are plastic caps placed over the top and the bottom:

The lower caps are meant for preserving the thermal interface layer applied on the base. By the way, the processing quality of the base is good (8/10).

Note that the FalconRock II (right, on the photo) has a copper insert in the base which favors to faster heat distribution over the radiator.

The FalconRock II also has a more reliable fastening to all the teeth of the processor socket.

As I already said, the housing of the fans are made of aluminum. The FalconRock II fan's dimensions are standard: 80x80x25 mm; rotational speed 2300 RPM. The noise level (as per the manufacturer) does not exceed 25 dBa, with the air flow being 28 CFM. On the WhisperRock III, there is a fan of the same dimensions, but with different characteristics. In particular, the rotational speed = 2700 RPM, the noise level has gone up a bit to 26.5 dBA, and the air flow has reached 37.87 CFM.

The interim conclusion is: these are two standard coolers of slightly different characteristics, whose only shortcoming is too short power supply wires.

Spire MicroFlow II

Unlike the two previous coolers, this one has been officially approved by AMD, which means a full compliance with the specifications (sizes, mass, fastening scheme).

Therefore, the cooler offers an aluminum radiator with a copper insert in the base. On the base, there is layer of thermal interface (T725 - externally resembles pink chewing-gum :), and the processing quality can be rated at 9/10.

The fan is made of plastic, with the dimensions 70x70x15 mm, whose rotation speed is 4200 RPM and the air flow is 33.8 CFM.

During the operation, the cooler makes an immense noise (31.8 dBA, as per the manufacturer), so it can't be recommended to those who are critical about the noise. In all the other respects, the cooler has almost nothing to complain about: the fastening clip hooks to all the socket teeth, with a protective grid installed on the fan. The only complaint is about too short power supply wires.

By the way, the wires are wrapped in a special cover, which much improves the exterior.

The interim conclusion is: the cooler is sure to appeal to computer assembly operators for its small bulks, pre-applied paste and the protective grid.

Tests

Before testing the coolers I summed up all their specifications into a table (the higher the score, the better).

Most of the subjective parameters were measured in marks (noise, processing quality of the base etc.). That is, the higher the score, the better the parameter is (the quieter is the noise). We tested all the coolers with a Soltek Sl-75FRN2-L motherboard; as the monitoring utility we used Motherboard Monitor. We also used an AMD Athlon XP Barton processor that ran at 2GHz (10õ200) under the voltage 1.7V.

The processor was warmed up with the BurnK7 utility. Then, the tested system was assembled in an Inwin S506 housing, with the PSU brought outside. This PSU (PowerManFSP) has a fan of variable rotational speed (depending on the temperature), and its close proximity to the cooler might have affected the measurement results.

If the cooler had its own thermal interface, it was scraped off, and then the measurements were performed with KPT-8 silicon-organic paste. Here are the results:

The Titan CU9TB/SC showed frustratingly low performance level. Even though this cooler is the most expensive among those presented in this review. Its advantages are its exterior and the presence of a speed regulator. The latter allows minimizing the speed and attaining absolutely noiseless operation. However, with an overclocked processor and a housing of poor ventilation this trick won't work: the cooling efficiency drops considerably. Therefore, the cooler can't be recommended to overclockers and owners of powerful processors (XP3000+ and higher). But the fanciers of quiet operation, it may be of interest as an alternative to Igloo Silent Breeze I / II.

Conversely, Igloo 2510 Pro demonstrated excellent cooling level. But that result was achieved due high noise level (high fan speed) and strong pressing (even all the paste was squeezed out :). Anyway, its low price allows to forget about all the other shortcomings. Even more than that - the company offer the Igloo 2510 with a less noisier fan. However, to the PC enthusiast I would recommend the Pro model plus a speed regulator. This combination allows attaining a balance between the noise and the cooling efficiency.

When it comes to the Spire coolers, their major advantage is their low price that does not exceed $10. Despite the ornate names, all the coolers have approximately the same characteristics. Anyway, I liked the Spire MicroFlow II model best of all.

Read more on this topic: