Cooling GPUs Graphics Reviews Watercooling

Alphacool Eiswolf 2 in a test against a Nvidia GeForce RTX 2080 Ti with 300 and 380 watts – with a little initiative an (almost) perfect solution

Assembly and important notes

RTFM yes, but please think along with us and keep thinking. I have already contacted the manufacturer and discussed the discrepancies that have arisen in the following. Thus, everything is already on the to-do list and can be solved by anyone quite easily. First you have to completely disassemble and clean the card of your choice, which should not be a hurdle. All disassembled parts should of course be properly cleaned and bagged. Matter of honor, by the way.

The thermal pads belong on the water block and not on the PCB. The block should later be placed on the table during final assembly and the PCB should be placed on top. Then nothing falls off (and down) even when the board is turned upside down.

Once the board has been cleaned and the GPU has been re-applied with thermal paste, the assembly can be completed. And right now, my first throw-in. First you should screw the four screws directly to the GPU socket, not use the four holes and threads further out. Nvidia did the same with the original and for good reason with the Vapor-Chamber. The four threads further out must also be used to relieve the RAM from one-sided and asymmetrical pressure (Space Invaders!).

But here you have to rely on four more screws (from your own stock), which are (not yet) included and unfortunately missing. However, 8 screws are really the minimum for a proper contact of this very important area. I had tested with another two threaded sleeves, but could not find any improvements there. If you want to do it very precisely, you can also put the matching Ultrasoft pads in the coil recesses. In the maximum case, this brings up to 4 degrees less in the area around the voltage transformers. In my later measurements, however, I measured the original condition as the customer finds it.

If you look closely, you will also see that I have put acrylic washers under the spring screws. Firstly, this increases safety and secondly, and more importantly, the possible contact pressure. If you don’t believe it: test it. If you would like to test yourself and your results later on without any big aids, you can easily test the correct contact between the radiator floor and the GPU.

I mounted the water block several times and under different aspects. If you have done everything correctly and the contact fits, then no more than 15 to 17 degrees difference between water temperature and GPU diode (e.g. GPU-Z). So, for example, if you get extreme 83 °C for the GPU (temperature target), the water would be so hot at well over 65 °C that one would have to be afraid of bursting.

The water temperature can be measured with a simple temperature sensor, which can be how I screw in instead of the refill. All temperature differences of more than 17 °C are evidence of poor contact between the cooler plate and the chip.

If you have put everything together correctly (even as a newcomer you can easily do this in about 30 minutes), there is nothing to prevent the initial commissioning. Equipped like this, we can now start taking measurements. I measure the power consumption in the lab and set the two runs to 300 and 380 watts each as TBP for the entire card, which it then of course reaches permanently, but never exceeds. Measurements are made with the GPU diode (GPU-Z), a control sensor on the board together with an evaluation unit and several measuring points on the back of the board using a high-resolution thermal camera.

 

I set the fan speeds in descending order in the following steps: 2000 rpm (maximum value), 1800 rpm, 1600 rpm, 1400 rpm, 1200 rpm and 1000 rpm. For each rev range I take both a game (Witcher 3 UHD) and the stress test (Furmark). Each run lasts at least 30 minutes or until saturation is reached and the temperatures do not change any more.

The room is air-conditioned with 22 °C. However, for safety reasons, I only measure the lowest speed with the 300 watt variant, so that 22 measurements are made here. By the way, the plausibility measurement in idle resulted in approx. 26 °C for the GPU, no matter which fan setting was selected. The rest is a little different.

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Igor Wallossek

Editor-in-chief and name-giver of igor'sLAB as the content successor of Tom's Hardware Germany, whose license was returned in June 2019 in order to better meet the qualitative demands of web content and challenges of new media such as YouTube with its own channel.

Computer nerd since 1983, audio freak since 1979 and pretty much open to anything with a plug or battery for over 50 years.

Follow Igor:
YouTube Facebook Instagram Twitter

Werbung

Werbung