someone made a documentary-like video about water cooling

[izumi_konota]

I'm not sure if it's okay to post video of other people here, so let me know if I'm wrong.

 

This guy is crazy: https://www.bilibili.com/video/BV1HV4QegE5t

 

In this 2 hour video, they ran a marathon test of different water coolers. The test itself costs 3 months since Feb. And instead of just showing results, they then visited several AIO cooler manufactures including Asetek and Deepcool. They even managed to interview some OEM and Tier 1 suppliers to see what makes a water block broken, and those suppliers really shared their know-how about this industry.

 

Really impressive. Maybe someone want to translate it. Lots of detailed designing ideas in this video.

 

---

 

Let me have a quick summary of opinions from manufactures & suppliers:

• Asetek: (Personally I think this is really a official response, and not my favorite one)
  • Some chemicals including Cu, C, Si, S, O are related to the compound blocking water.
  • Debris seems not copper, and might be a case-by-case problem.
  • No detail response available.
  • Performance degrade have multiple root cause like fins covered by dust, or liquid vaporize. Still, case by case. A sudden performance drop is covered by RMA, but a long-term degradation is not.
  • As a OEM provider, they an provide 24-36 month RMA assurance for brandings. Brandings might have other promise to end users.
  • Critical environment, like this marathon, is not covered.
  • Water temperature shall be controlled under 60 degrees (or, for U.S guys, 140F).
  • Reaction between aluminium and copper is rare, even in lab.
  • Reason of using aluminium as radiator & fin, instead of copper is mainly about cost saving and copper won't give much performance. Copper have higher heat resistence, and not having much business value. Still, CQ curve (cost of quality curve).
• ProArtist / YaJun:
  • A factory tour. Fin making costs a lot. Some small companies might not clean their product enough, causing copper debris flowing.
  • Yajun actually put their tube into water cooling liquds, and some white floc came out from the tube. They think it's caused by unstable vulcanization of different parts of rubber. QA check is hard because there are only 3 providers in the whole industry, and they won't cooperate enough with their engineers, causing inconsistent materials.
  • Dusts in water block seems comes from a mixture of liquid and flux for soldering. It can't be observed unless destroy the radiator.
  • They can't do much to resolve performance degradations. They can only make it slower. But they think AIO coolers might be swapped after 2-3 years.
  • AIO cooler is similar to car cooling system, and facing same problems. A proper maintenance is necessary.
  • Aluminium & copper: same opinion.
• Titan: (Personally I think they are the least informational one)
  • Different performance between Intel and AMD is caused by their different designs. AMD's initial design is changed from Intel, so it's not good enough. Still improving on each batches but they are not advertising it.
  • Debriefs & flocs seems comes from plastic washers. Happens under extreme condition.
  • Performance degradations: it's normal. We can RMA if necessary.
  • Aluminium & copper: not only cost problems, but they still observe electrochemical reaction when using full-copper materials. Soldering flux is not preventable.
• Deepcool: (Dude this engineer is awesome, he is a scientist)
  • Deepcool have their own cooling liquid. They can adjust their recipe to ensure there are enough backup ions, so their liquid can keep alkaline during their assured duration. This can make reaction harder to happen, even if temperature is higher than room temp. But a lower temperature is also helpful to prevent oxidation. (P.S: Cooling liquids are made from Alcohols. After oxidation they will become some form of acid and ketone, and gradually eats your radiator by ketoaluminum reaction).
  • A good maintaince can basically prevent performance degradations, at least for their products.
  • Black materials in water block is not debriefs caused by aging. They are molecular migration between copper and washers. (Well I think this is another saying of aging).
  • Intensive fins is not always a better choice. This is a complex hydromechanics problems related with speed, reynolds number and solubility. They will use CFD workstations to perform design.
  • Supply chain management is important. They will perform strict test to ensure stability.
• VK:
  • Sample test shows the blue messy stuff is flux. Their thinner fin makes blockage more easier to affect performance but they still choose that. They prefer max performance, instead of long-term stable performance.
  • RMA is okay if user thinks performance droppers unnaturally.
  • Unclean radiator causes lots of problems.
  • Tubes have two types: EPDM or FEP. EPDM might have vulcanization problems but they are using FEP, which is more stable and won't dissolve.
  • Test each radiator on flux is not possible. They will wash it but it won't solve the problem.
  • Future solution is using smart sensors to have early problem detection.
• BingDian (OEM of THERMALRIGHT):
  • No more new tech news.
  • And yes, supply chain management is hard. All manufactures including Asetek was having same problem before.
• YiChen (OEM of SAMA):
  • No more new tech news.
  • Old storage product might cause different results.
• EK:
  • Debriefs might be left over of washing.
• Apaltek:
  • Electrochemical reaction between copper and aluminium only happens if you are using wrong cooling liquids.
  • Workers on the pipeline can also affect product quality.
• Anonymous manufactutre:
  • Cooling manufactures bluffs a lot. A 6 year promise is bullshit now because workload increases a lot.
  • Biggest improve of this industry is dropping ethylene glycol, and use propylene glycol. Other than that, is fans.
  • Max theoretically available RMA period is about 3-4 years. 6 year is gambling. Manufactures are betting user's will change their computer within 4 years.
  • Social network advertising is more important. No one cares about long-term quality because end user is hard to verify it.
  • Repairing rate is not important. AIO market is small, and rate of repairing have no statistical meaning.
• Mei Ji (supplier of radiator):
  • Yes, flux will left when soldering. This is a full handmade process.
  • They are designing a new radiator to prevent flux contacting water. Sample shows there are nearly 0 flux leftover.
  • AIO makers are making radiator themselves, but Mei Ji think they will do worse.
  • Washing is...well, not that helpful.
  • As supplier, they are making things according to their customer's requirements, which is AIO maker. They can have enough quality if AIO maker providers a exact requirement.
  • Some test samples are too old to be test.
  • Some small factory will buy products not passing QA.

There are some other info about future test and sample analysis.