News - Asetek granted GPU liquid cooling patent

[HEXUS]

An Asetek designed cooler is used in the AMD Radeon R9 295X2 reference design.

Read more.

[Tabbykatze]

Wow, how did they get granted such a broad patent. That will cripple any other cooling unit creators with royalties or patent infringements!

Unless there's something radical in the design and documentation I'm missing, I'm not sure how they manage to successfully patent the whole water cooling concept!

[cjs150]

Yet another example of how inept the US patenr office is.

Patents are not permitted when they are simply duplicating prior art. Micro-channels designs are at least 10 years old, I own several

[LeetyMcLeet]

Yet another example of how inept the US patenr office is.

Patents are not permitted when they are simply duplicating prior art. Micro-channels designs are at least 10 years old, I own several

Couldn't agree more. Kind of reminds me of that 'other company' being awarded generic, broad patents for something not dissimilar to other products already on the market. They're about as straight as a round-a-bout the US patent office.

[Tpyo]

Surely (and more so given the comments above) this could easily be overturned with examples of prior art?

[Spreadie]

I can't help feeling they've filed this just to have another stick to beat CoolIt with.

Pretty distasteful stuff.

[letsrace567]

It looks like there's a decent difference between this and conventional water blocks actually. there are 2 different "bodies" noted in the documentation and in the picture. It looks like the "thermally conducting body" connects directly to the card, presumably contacting the GPU, RAM, and VRM components, and then there is a separate "fluid circulation body" that attaches to the thermally conducting body to provide a means of removing the heat.

Going by Fig. 1, I'm not convinced that the liquid actually passes between the two bodies, but that the "fluid conducting body" is a more universal waterblock/pump combo that then attaches to the "thermally conducting body" that would be tailored to each card, thereby adapting the waterblock/pump combo to many different types of cards. Like a waterblock attached to a custom heatsink. Here's a quote from the patent doc: "The thermal interposer may also include a cold plate assembly removably coupled to a second surface of the planar body opposite the first surface"

It looks like there's nothing to worry about for makers of custom waterblocks, such as EK and XSPC, as this design is actually much different from the more traditional design, where the liquid actually circulates through the block attached directly to the card.

[edzieba]

Looks like the patent is actually on a cold-plate that sits between a universal waterblock and the card, with a heat-pipe embedded into the plate. I've seen existing 'universal' plate-adapters for GPU watercooling with AIO blocks, but I've never seen one of those plates incorporate a heatpipe.

[cjs150]

Having now read the patent I am even more confused as to how this was granted a patent. It is either a copy of the old switech universal designs or, as letsrace57 says, the liquid does not circulate through the central thermal conducting body in which case performance is likely to be vastly inferior to existing designs.

[EN1R0PY]

So... how much do you reckon the guy at the patent office got in his pocket that day?

I might have a go at patenting a device for covering the entry to a room with at least on user operated retaining clasp...

[EN1R0PY]

I'm pretty sure that I've seen designs similar, and at any rate they've just removed a very broad and generic design option from all other manufacturers which is hardly a good thing for the consumer.

If they had invented more than a sandwich they could have the patent, but that's all they've done here, the earl will be so disappointed!