Spacing between radiator and fans

[NiHaoMike]

Has anyone quantified how the performance of a radiator+fans varies depending on the spacing between the two? (There will be a duct between the two so there won't be any flow bypassing the radiator.) I think increasing the spacing will improve flow in the areas of the radiator that would otherwise be covered by the "dead zones" of the fans, but beyond how much spacing does it get into diminishing returns? The radiator I'm using has a 127mm x 225mm finned area with a 20mm thickness, matched up to two 120mm fans.

It's actually for power electronics cooling rather than PC cooling (main difference is that power electronics will tolerate higher temperatures, 100C die temperature is well within ratings) and it's nonconductive mineral oil instead of water (water doesn't go well with 520V DC), but the general idea is the same.

 

[Blaylock]

I know Phobya makes a fan shroud for this exact reason. I don't remember what sizes they have but I'm sure they make a 120mm since that's the most common size. As far as diminishing returns, I would imagine this is absolutely the case since a fan's static pressure will drop dramatically the further you get from the source.

 

[EarthDog]

There's been a lot of studies on this from a ways back... skinee labs is one. But its usefulness depends on the fan amd radiator iirc...but generally beneficial.

 

[NiHaoMike]

I'll be 3D printing the adapter so any reasonable spacing will work. The question is at what point does increasing the spacing further just waste material for minimal additional gain?

 

[EarthDog]

That varies with the radiator and fan used. IIRC, the difference between say, 1 fan frame and two fan frame spacing the temperature difference wasn't much. You see more significant increases with fans that have larger hubs, but again, it's a couple of C at most.

 

[BugFreak]

If I remember right the biggest difference I saw was in noise. It's been a long time since I bothered with it but as I moved the fan away it made less noise.

 

[Blaylock]

To be brutally honest you will most likely need to do your own testing. I would start by creating five shrouds with thicknesses in 10mm increments up to 50mm. This would give you a maximum of 150mm of spacing and as little as 10mm. Just my 2 cents.

 

[EarthDog]

The testing I was referring to...

https://www.techpowerup.com/MartinsLiquidLab/Radiator-Fan-Orientation-And-Shroud-Testing-Review.html

TFC Shrouds and shrouds in general - Always provide a good performance benefit and reduce noise at the same time. Highly recommended in all radiator setups particularly for higher speed fans unless the space does not make them an option. You can see as much as 10-16% gains on higher speed 38mm fans and 6% gains with slow speed fans. Generally the higher the fan speed and the larger the fan hub, the more gain you will see with shrouds. Optimal depth appears to be somewhere around 30mm +-, this is likely to depend on hub diameter and RPM to some extent.

 

[Mr.Scott]

Did it way back when. 25mm is about the break point for diminishing returns.

 

[KurtBlanken]

Here's testing Martin did: https://martinsliquidlab.wordpress.com/2012/01/15/radiator-shroud-testing-v2/

As for how long of a shroud before diminishing returns, that really depends on your specific fans and radiator.

Another thing to consider is whether your fans have vanes. If they don't then adding some vanes to your shroud will improve performance more than extra ducting distance would.

 

[The_Jizzler]

I think as most setups went towards quieter and lower CFM fans, and the fact that legit PC water cooling rads became available, shrouds became a non issue. I used to use them when we all were using heater cores and screaming mega high speed/cfm 92 or 120 mm Delta or Sanyo fans. We just needed a way to adapt fans to the heater cores and homemade shrouds with those high speed fans was a way to increase airflow and adapt a 3-4" fan to a 6-10" heater core. Noise was an afterthought. After the market started catering more to enthusiasts and we got decent rads, people kept selling shrouds as I think that's what we were all used to.

Unless you got high rpm high CFM fans I don't think you are going to see any reasonable difference

 

[EarthDog]

The test above was with 'PC' rads. Here's the conclusion...

CONCLUSION
Shroud benefits are very much radiator specific. While the MCR120 shows only very minor benefits in pull condition which are nearly immeasurable, the RS120 shows very dramatic benefits of 1C in pull condition. My first round of shroud testing only tested on one radiator with two fans and found gains on both. That test appears to still be good, however this round in testing a different radiator it has become clear that the benefits can not be generalized across all radiators and that shrouds on some radiators in the right conditions can even hurt performance. The MCR120 was generally a better performer in push condition without shrouds.

About the only generalization is that push/pull is still your best bet if cost is not a problem and that will hold true for all radiators.

Bottom line:

Shroud benefits are Radiator Specific!!!
Are shrouds a performance benefit and should you use them for your build? MAYBEÂ…