Radiical Computing only sells products that they have tested and used. This is the first review providing purchaser feedback. It is our intention to publish data in such a way that lack of technical knowledge will not impede the readers understanding. The test results page is not intended as a definitive answer to the many technical questions relating to hardware purchase. Please feel free to use the "contact us" button for further information and we will endeavour to answer your questions.

Please Note: Older reviews and test results have been moved to the Reviews Page

Silverprop EVO-SE Test Results

The reader of this review should be aware that Radiical Computing undertakes this analysis as an independent reviewer. It is important for the reader to understand that we have no fiscal or other relationship with Silverprop Australia Pty Ltd. Our objective is to present reviews of products for the water cooling community in an objective manner. We purchase all products for review from resellers in the same way an enduser would purchase. We do not purchase products that have been specially prepared to increase their performance. Reviews are presented in a "warts and all" fashion for all products we test. As manufacturers of high performance water blocks it is not in our best interest to make exaggerated or negative claims about the performance of other manufacturer's products. Our aim is to maintain credibility in the water cooling community so our findings are reported as objectively as possible. Reviews are written without the hype often associated with commercial reviewers.

Silverprop has been a pioneer of water cooling in Australia for a number of years. In product terms there is no doubt that Silverprop have led the way in product presentation from their early days. I remember my first Silverprop water block purchase. The care and attention to detail of the Silverprop team has always been second to none. It is difficult to imagine that any product coming from Silverprop would not live up to the presentation standards they have set themselves.

Picture courtesy of Silverprop

One very difficult area for the prospective purchaser of Silverprop products has been the lack of substantive reviews. Certainly the EVO-SE has been reviewed a number of times but has not been reviewed with the high level of frequency that is expected by the end user. Reviews have been highly technical and relate more to the flow characteristic than simple comparison with other water blocks. Compare the numerous reviews of say Asetek or Swiftech and you get the feeling that the same number of reviews should be present for Silverprop. This situation has led to considerable debate over comparatives in performance. Some time ago we here at Radiical compared directly three manufactured water blocks. These were the Silverprop SP, the LRWW block and our own Series One X-Block (For those interested, in this comparison, you will find the review named the The 3 Block Shootout here )

Since the time of that review the Silverprop EVO-SE has moved on its development. There have been a number of revisions of fin design although the external appearance has remained the same there has been a movement away from the copper topped version which was the EVO-SP to brass and aluminium top products. Separate water blocks must be purchased to accommodate the various CPU's from Intel and AMD. To date I am unaware if a model is available for the new AMD64 CPU. The current model EVO-SE internal structure is shown below.

Pictures courtesy of Silverprop

For those not familiar with impingement design the center slot shown in the picture above directs water onto the recessed fin area below it. The result is that the water is forced to become turbulent and interact aggressively with the walls of the fins. The outcome of the increased turbulence causes more water molecules to collide with the hot fin material. Heat is then carried away from the fins by the water flow and cooled in the radiator.

Water in Water Out

The big problem with impingement designs is the need to maintain flow rate. To do this you need to ensure that the pump that you use has sufficient capacity to maintain an adequate flow rate through the water block at the rated level for that water block. Most manufactures add this information about their product by way of a statement such as "suitable for use with Eheim 1250 pumps" or something similar.

The flow rate we established for the EVO-SE was arrived at by using the Eheim 1250 and a single BA radiator core. Tubing length was kept to a minimum. No other restrictions were present at that time. The flow rate was taken to be approximately 6L /min. This should be taken as a best case scenario for flow rate. Expect less flow rate than the figure presented if your cooling loop contains long tubing runs, bends or other impediments to flow. When we added a moderately restrictive GPU block this value fell to below 5.4L/min. In his article here Joe Citarella illustrates the flow characteristics and mentions the need for care in adding restrictive elements to the water cooling loop. The clear indication with all water blocks is to check the effect on flow rate with the whole loop in place so as to ensure you have enough flow rate to maintain the cooling cycle.

So we have established that we have a moderately restrictive water block and as such there will be a need to ensure that we match the pump to the block. What else do we need to know about this block before testing.

Check it Out

Its a fairly obvious trait of us humans to intricately examine every new purchase or acquisition. The inspection of this block turned up some interesting findings. It is clear from the molding marks that both top and baseplate of this block are cast items. The weight of the block at 410grams makes this block reasonably heavy. The central barb is a composite entry point and the contoured slot at its base is the integrated impingement mechanism. A disturbing finding in the base of the outlet barb is the intrusion of the sealing gasket. The concern here would be the possibility that the extruded gasket may block the flow of coolant out of the block.

The intrusion of the gasket material is not a new finding. The same intrusion was noted in the EVO-SP we tested earlier. In the case of the EVO-SP the block had to be opened and the offending gasket material removed before the block would function properly. A caution for users who do open their block this action voids your warranty from Silverprop.

If its not on then its not on

A major problem facing the user of this block is the fitting mechanism. It can be a nightmare for some to fit. How do I know this. If you have read the review mentioned previously called the The 3 Block Shootout you will already know the difficulty that I had initially in fitting this block. Fortunately I have now improved with practice and can fit the block competently with very few errors. Sounds confident but true. For those of you who find continuing difficulty I found that if you simply turn one of the clamping mechanisms around so that the triangular area faces the opposite way to a position of flat against the block you are more likely to achieve success. Simple really. Please attempt to follow the manual description to fit this block before attempting any other method of fitting.

To illustrate the method of placement of the clamping mechanism described above here is a picture of the positioning of the retention mechanism. You should note the right hand clip has been rotated to provide additional space for tubing and attachment to the socket mounts.

Please note that you may have to check that the inlet barb is located in the same position as mentioned in the manual supplied by Silverprop. Check also to make absolutely sure that the notch in the base of the water block corresponds to the socket holder. As usual no responsibility for any damage you do in following these instructions is accepted by us.

Lap Lap

I have yet to meet a machined water block base that is flat. I am sure they exist but as yet this joy has alluded me. Cast copper bases are usually far flatter than rolled copper simply because they do not have to respond structurally to the machining process. Check the base with some engineers blue on a bit of glass plate or something very flat and then lap the water block base flat and smooth. No point in complaining if you don't do this. I have never been a supporter of the "see your face in it" lapping method and I would mention there is anecdotal evidence to suggest that very polished surfaces can actually squeeze out the TIM material leading to lower performance results.

Party Time.

Procedure

Test Baseline temperatures

Overclock the CPU and measure observed changes in Idle temperature using MBM5

Load the CPU with work intensive tasks and measure again the load temperature observed. Because this CPU is capable of active throttling of tasks CPU utilization is set to 100% in Folding@Home. Concurrent use of 3dMark03 has also been added.

Graph the results to demonstrate the data.

Remounting

Due to the pressure of other projects the block will be re -mounted on three occasions only.

Test Bed

AMD 2100XP "B" Core AIUHB 0306 (Previously tested to reach 2573Mhz)

Motherboard Epox 8k9AI

Corsair XMX 3200C2 256 megs DDR Ram

ATi Radeon 9600 Pro

Eheim 1250 pump

Radiator Jaron BA with shroud

2 x B12025 12m 120 mm Fan

OS Windows 2000 service pack 3

Ambient was set at 25 degrees Celsius for all testing.

Ambient temperature monitored 12.5 mm from radiator intake

Results

Default 1730 MHz at 1.60 volts (133 x 13)

Ambient Temp; 25 degrees Celsius

Idle Temp at Default; 32 degrees Celsius

Load Temp at Default; 35 degrees Celsius

The following overclocks were conducted to test the blocks responsiveness to overclocking.

2167 MHz (13 x 166) at 1.65 volts Calculated wattage 82.69 Watts

2220 MHz (13 x 170) at 1.65 volts Calculated wattage 84.72 watts

2286 MHz (13 x 175) at 1.70 volts Calculated wattage 89.93 watts

2353 MHz (13 x 180) at 1.775 volts Calculated wattage 103.91 watts

2413 MHz (13 x 185) at 1.85 volts Calculated wattage 115.76 watts

The temperature results for the water block "at load" are given in graph form below.

The second graph illustrates the comparison between voltage and overclocked temperatures.

Readers of reviews on this web page have contacted me about providing reviews with a comparison of water block performance. To satisfy that need here is the comparison of the EVO-SE with our own CWX X-Block in a clock for clock direct display. Two areas which were commonly reached by both blocks are shown.

Water block design is complex and difficult thing to do. As a water block concept is being developed there is a point in time reached where decisions on the direction the block design will take that have to be made. Base thickness for example can be super lean, medium or thick. Water channels may be used or the block could be an open design. Direct impingement or crossflow. Will the designer opt for cup and injector nozzle utilization. The list goes on.

The EVO-SE designers have used several of the construction options given above. The EVO-SE uses direct impingement into a channeled base with a unique Plura Arc channel design. The Plura Arc channel is as it states a series of channels constructed so that the apex of the arcs reaches down to the centre of the block base. The base is therefore thinnest at its centre and becomes thicker as you move away from the centre. Heat generated at the CPU core travels through the thinnest base material directly at the point below the nozzle. Heat exchange reduces the further the heat travels from the centre.

The information about the heat transfer characteristic of this block goes to the point raised earlier about design decisions. The design limitations are created by the earliest decisions about shape configuration and dimensioning of the block structure. Balancing channel dimensions and flow rate are at the centre of those decisions. Base thickness and the size of the base absorption area are also critical to performance at higher operational voltage. The EVO-SE has the design limitation of a moderate flow rate.

Ambient Temperatures

Many commercial review sites review their products within a narrow range of temperature parameters. It is in the best interests of these sites to point to the favourable outcomes of over clocking and temperature results at much cooler temperatures than we experience here in Australia. How many times have you read an article which gives a glowing report to product "Z" at 18 degrees Celsius ambient. These reviews have little relevance when you are reading the article in a room where the ambient temperature is 35 degrees Celsius. What is particularly relevant in the cooler months of the year is simply not what happens when the ambient air temperature starts to rise in the warmer months.

It is highly likely with the EVO-SE that, in Australia's summer heat, you would need to address the increasing ambient temperature even when a lower nominal voltage over clock is used. What is needed is a block which approaches the smallest delta between air and water. I am left with some concerns about the EVO-SE ability to fully meet this requirement at higher ambient temperatures. Small changes in ambient produced quite marked changes in the blocks performance at 1.85 volts.

Hot and Hotter

There is no question that CPUs are getting a lot hotter. The CPU used in this review, however, is a low wattage model with a rating at default of 62.08 watts (Barton 2500+ being 68.30). In the data given above it can be quite clearly seen that a lot of heat is generated the more you move up the over clocking scale. At 2573 MHz at 1.95 volts this particular CPU will be developing 137.14 watts of heat output (calculation courtesy of http://www.benchtest.com/calc.html). At a point sometime before this heat output value is reached the EVO-SE has past its upper limits of performance and is failing to contain the heat output.

Conclusion

The EVO-SE design is ideal as a general purpose block design. At low to moderate voltages the block can pick up and dissipate most of the heat generated at the core. At higher voltages of 1.85 volts and above the block starts the process of being over run by the heat being generated. As a performance block there are several other water blocks which will outperform this block, including our own CWX model.