Danger Den NVIDIA 4101 H20 Water Cooling Kit Review

Water Cooling by KeithSuppe @ 2007-12-19

Danger Den is certainly a pioneer in the H20 world, and while they have grown their products, they still retain that attention to detail. They offer so many products at their site it is often better to choose a carefully selected kit. Today we test a true performer, Danger Den NIVIDIA 4101 Water Cooling Kit.

Introduction the Waterblock

Introducion

Madshrimps (c)


At the heart of any water-cooling kit is the water-block. And at the heart of the water-block itself is the area centered above the processor it cools. This real-esate has a long list of monikers, for our purposes we will stick with impingement zone. In a design where the inlet is located above the processor core it is known as direct impingement, due to the water stream impinging upon it. How this relationship between water-stream and water block interact determines many factors, most importantly how well the processor is cooled. In the close-up above we see the unique impingement zone of DD's Copper-TDX. It is this block Danger Den chose to build their NVIDIA 4101 water-cooling kit around. I first reviewed the TDX back in 2004. The original design has been so effective it has remained unchanged until just recently. While DD will keep the original design, they have released a version specifically for multi-core (Quad core) processors. Danger Den's new MC-TDX built on the direct impingement extruded pin design can be seen below.

Madshrimps (c)


Madshrimps (c) Madshrimps (c)


The photo and thumbnails above exemplify an extruded-pin impingement area much larger than the original TDX. This may complicate matters since the 1/2" inlet, as seen in the left thumbnail, only impacts a 1/2" diameter area of those pins. The right thumbnail above shows Danger Den has recessed those pins directly below the inlet. I am not sure of the reasoning behind this unless the pins are shorter because the area is sunk deeper into the base itself. Photos taken from DD's site fail to reveal any details about these pins. Back to the TDX, pictured below with its universal mounting bracket which fits just about every AMD and Intel socket and requiring no disassembly in between.

Madshrimps (c)


While the base plate wasn't finished to a mirror image it was flat which I measure using a metal L-square. A mirrored finish has less to do with flatness then it does gaps and striations which are also critical for thermal paste application. In the end if your water block isn't making proper contact with the IHS it matters little how well the base is finished.

Madshrimps (c)


Removing the universal mounting plate reveals a Lucite top with 1/2" barbs. The top plate sits atop TDX's side walls and positions the inlet directly above the impingement zone. The outlet is off-set where the shape of the block narrows substantially to aid in evacuating water.

Madshrimps (c)


How a water-block is constructed will either support or defeat its intended design. Water flowing down through the accelerator nozzle is initially under high pressure as it’s forced through the nozzle opening. As the fluid impacts into the impingement area with its pseudo "cups" the resulting turbulence increases heat transfer to more water-molecules. The area of the block housing the impingement zone and accelerator nozzle is wider then the outlet area of the block. The wider area of the block on either side of the impingement zone is at a lower pressure then the impingement zone itself. Thus water impacting into the small walls which form pseudo cups and the larger walls supporting the accelerator nozzle all work together to evacuate water now carrying heat to the opposite end of the block and the outlet.

Madshrimps (c)


Concerning impingement, TDX is unique as the only water block to offer Accelerator Nozzles. The varying shapes of these nozzles alter the water-stream as it impacts the impingement zone. Below we see the TDX with its accelerator nozzle installed.

Madshrimps (c)


Madshrimps (c) Madshrimps (c) Madshrimps (c)


Next we take a look at the parts Danger Den chose to mate with their TDX ->

System Components

DD nVidia 4101 kit

Madshrimps (c)


Other components Danger Den chose for their NVIDIA 4101 water-cooling kit include the following:
  • TDX Universal Block
  • Danger Den D5 Fixed Speed LAING Pump
  • Single 5 1/4 Bay reservoir
  • 1/2" ID Tygon Tubing
  • Black Ice Xtreme Radiator
  • MCT Non-Conductive Coolant
  • Yate Loon Medium 120mm Fan (we received Sunon KD1212, Danger Den has since found a superior model)
  • Miscellaneous Hardware

    Madshrimps (c)


    12V-D5 which is sourced from LAING Thermotech as the D5-38 Vario pump. Just below chart taken from LAING Thermotech.

    Madshrimps (c)


    Madshrimps (c)


    One particular component in any water-cooling system which will have a large impact on temperature is the radiator. This is where thermal energy carried from the water block is dissipated into the atmosphere. Danger Den chose the and how effective the radiator is at its job Black Ice Xtreme for this purpose.

    Madshrimps (c)


    It may be debatable if you’re looking to make an immediate improvement in your system which will have the greater effect, the radiator fan or radiator itself. From personal experience I found experimenting with fans is not only less costly, it has a greater impact on temp. Danger Den chose the Sunon KD1212PTB 1-6A, DC 12V, 4.8W operating at 90CFM and a 44dBA SPL.

    Madshrimps (c)


    Danger Den supplied a 32oz bottle of Midwest Cooling Technologies mct-5 coolant with this kit. I've been using the product on and off for years and find it does offer an inclusive fluid as opposed to mixing distilled water, an anti-algae product and Water Wetter.

    Madshrimps (c)


    I've taken the liberty of re-printing some of MCT's mct-5 product claims:
  • *Non-Conductive - Conductivity tested at 4e10 Ohms/meter
  • Freezing point is -5° Fahrenheit (-20° Celsius)
  • Prevents algae growth
  • Pre-mixed and ready to use
  • Contains lubricating properties helping pump life
  • Prevents galvanic corrosion
  • Safe on plastic, rubber, seals copper, brass, aluminum, steel, and gaskets
  • Incorporates anti-foaming properties

    Everything you would hope a pre-mixed cooling fluid would offer. Insofar as electrical conductivity MCT states:

    * MCT -40 and MCT-5 are not likely to cause damage to operating pc boards powered at 12 volts or less. It is not to be exposed to power supply input voltage or components in the power supply unit...

    Danger Den has been selling MCT for sometime and I reviewed the product twice, once after a long term leak where the fluid had been dripping into my PCIe slot and onto the card. My video card actually survived although the product literally baked on and began to corrode a few SMD connects.

    Onto the installation and fill/bleed tests ->
  • Installation

    Filling and Testing:

    Danger Den supplies a bay reservoir with the nVidia 4101, however; when I initially set-up the system I used an Alphacool reservoir and used the Kandalf just to break-in the kit while photographing Mozart Tx for its review here. Mozart acts as the ideal work-station with its open area and large doors removed (temp wise). The only complaint I've ever had about Bay Reservoirs is that you must slide the reservoir into the front of the case and then attach the tubing. There may be no problem during fill/bleed testing outside the case, however; during the re-connect you have to be careful. As you can see towelling paper is placed beneath each connector in the event things go awry.

    Madshrimps (c)


    The following PDF from OverClock.co.UK entitled Danger Den H20 kit installation instructions is a thorough guide on the subject. Filling and bleeding your system out-side the case as seen above should be a given. Once assembled on the Mozart Tx the pump and radiator were mounted in the chassis top section.

    Madshrimps (c)


    Danger Den has been using open ended bolts for years. I chose to replace these with a Phillips head bolt making it much easier to slide the bolt through the underside of the motherboard (use a plastic washer) instead of placing a washer and nut on the underside of the mainboard and then another washer and nut on the topside. Besides the difficulty installing and securing two nuts simultaneously an open ended bolt isn't as secure. Altering the mounting hardware in this scenario has no ill effects on mounting pressure; in fact it made for a more stable mount. Be sure to alternate when tightening and tensioning springs to evenly distribute applied thermal paste. Below mounting hardware installed on the Socket-AM2 Asus M2N32-SLI Deluxe, with the Opteron 1212.

    Madshrimps (c)


    The kit was also tested on a Socket-775 Gigabyte P35C-DS3R running the Q6600. Installing TDX onto the socket is where you have to think about orienting the impingement zone so it’s centered over the core. As an example of what can go wrong below I applied a dollop (or large drop) of Tuniq TX-2 (which just won our TIM round-up) onto the center of a Q6600 and simply mounted the block in what looked like the correct orientation. Temps were rather high so I removed the block later in the day to find the offset in the photo and thumbnails below.

    Madshrimps (c)


    Madshrimps (c) Madshrimps (c)


    Finally the TDX mounted correctly on the Gigabyte P35-DS3R Socket-775. The block is oriented so the impingement area is located directly over the center of the IHS. This is essential regardless of the processor.

    Madshrimps (c)


    Next test methods, results and final words ->

    Test Results / Conclusion

    Tests / Conclusion


    When chip-makers began using more accurate internal thermal diodes I began to challenge certain methodologies where a thermistor or thermocouple was used to determine processor temps. Chipmakers were relying on internal thermal diodes not only for thermal throttling and/or speed-step technologies but for shutdown as well. Who is more invested then the chipmaker themselves at monitoring and protecting their own product. Early on problems arose, as the internal thermal diode was literally in the best position to measure temperature, accurate software was and still is hard to find. There are just a few programs able to read diodes accurately, one such program is Core Temp available as freeware. I will continue to use software such as Core Temp because the simple fact is: as heat is dissipated more effectively the internal temp of the processor should drop. Measuring that interface between IHS and heatsink base is fine if that’s exactly what you want to measure. If, however; you want to determine CPU operating temp which is critical to performance and longevity, monitoring internal thermal diode is key.

    AMD Socket AM2 /Intel LGA-775 Test Systems:


    Madshrimps (c)Madshrimps (c)
    Processors 1. AMD AM2 Opteron 1212
    2. Intel Q6600 Socket-775
    Mainboards 1. Asus M2N32-SLI Deluxe WiFi
    2. Gigabyte P35C-DS3R
    Memory 1. GEIL DDR2-800MHz (2x1GB)
    2. Super Talent DDR2-1000MHz (2x1GB)
    Graphics BFG 8800GTX
    Power Supply NZXT Precise 1200W
    Cooling 1. Stock Air-coolers Socket-AM2/Socket-775
    2. Swiftech Apex H20-220
    3. Danger Den nVidia 4101
    Storage Seagate Barracuda 80GB 7200.10 SATA2
    Optical 1. Plextor PX755-SA DVD/RW (SATA)
    Open Chassis Mozart Tx
    Operating System Windows XP


    Testing:

  • Sound
  • Swiftech's Apex Ultra H20-220 uses a Dual 120mm radiator and the kit provides 5V, 7V and 12V connector options. The fans, Delta WFB1212 is rated at 72CFM @ 12V 34dBA. The philosophy behind Swiftech's design is low noise / high performance. Danger Den's design seems focuses more on performance rather then silence, however; this is where a conundrum exists. When I received the NVIDIA 4101 kit it came with a Sunon KD12121PTB rated at 90CFM @ 12V 44dBA. This fan was one reason it performed as well as it did, although it was rather loud. The kit now comes with a Yate Loon "Medium" rated at 70CFM and 33dBA.

  • Danger Den nVidia 4101 sound level Sunon = 46dBA
  • Danger Den nVidia 4101 sound level Yate Loon = 37dBA
  • Swiftech Apex Ultra H20-220 sound level = 39dBA


  • Temps
  • Thermal paste used throughout was Arctic Silver Ceramique following their recommended application methods. Concerning flow rate and fan speeds both kits use the D5-38 Vario which ran at 12V and all fans were run at their max 12V setting. All tests were conducted in an open-chassis eliminating the temperature differential between internal case temp and ambient temp. Ambient temp was maintained at constant 20C ~ 21C. Processor core temps were measured using Core Temp 0.95 monitoring the internal thermal diodes. To simulate LOAD the CPU S&M was run at 100% for at least one hour.

    Madshrimps (c)


    Epilogue:

    Danger Den now supplies the Yate Loon medium fan rated at only 70CFM and 33dBA which is why I ran another series of tests using a Delta from the Swiftech kit. At 72CFM and 34dBA rating it’s almost identical to the Yate Loon. One of the least expensive and most effective methods of altering your water-cooling system is to make fan changes. Increasing CFM will get you closer to ambient, albeit at the cost of silence.

    Conclusive Thoughts

    I was truly amazed how well DD's NVIDIA 4101 performed considering the competition. Both kits are based on 1/2" ID and both utilize the LAING D5-38 Vario pump, however; Swiftech spec's a dual radiator in their MCR220. Given the results above this leads me to conclude the catalyst for performance resides in the five year old design on which the Copper TDX is based. This begs the question, will the MC-TDX out-perform its elder sibling? Regardless we're discussing the nVidia 4101 water-cooling kit which proves to be a formidable contender.

    PRO
    Excellent performer (even better with Sunon 90CFM fan)
    TDX Copper block cools quad core and retain accelerator option
    Single Radiator is compact, versatile mounting

    CON
    Noise level (w/Sunon fan)
    Mounting hardware could be better


    Danger Den provides a truly high-end water-cooling kit in the 4101. Its one of the better products I've tested and perhaps the best 1/2" kit on the market based on a single radiator. Insofar as my sample having the 90CFM Sunon, substituting the Delta which is almost identical to the Yate Loon Danger Den now supplies, did raise temps a few degrees, but it reduces noise. Would I purchase this kit? Yes, highly recommended!

    Recommended for



    I would like to thank Danger Den for submitting their kit for testing, until next time!
      翻译: