Radiator Fan Orientation and Shroud Testing


Welcome to my fan/shroud testing.  Over time I've heard and seen a few experiements around fan orientation and shrouds, some that were conflicting with others and before I spend too much time testing more radiators, I figured it was time to put my own questions to test and see what I can find out about things like fans pushing vs pulling and shroud depth optimization.

First and formost since most users usually run just one set of fans and you can place them on either side of the radiator, there is always the question of which side?  Do you install the fan on the side and "Push" air through the radiator, or do you put it on the back side and "Pull" air through like a vacuum?  Users with experience with fans and dust buildup quickly realise that the hub in the center of the fan leaves a dead spot where very little air flows.  You'll also notice that this dead spot is worse when a fan is pushing air vs. pulling air.  Then there are shrouds, many people have reported worthwhile improvements to installing shrouds on the radiator, but I'm not sure there has been any clarity as to where the shrouds help or what depth is optimal.

Before running these tests, I had always thought running fans in pull, and adding a shroud is best. Better yet would be two fans in pull.  I based all of this on having witnessed the flow through the radiator being much more distributed in pull (Less hub effect) as well as the thought that removing the fan heat from getting into the radiator is also good benefit to pull.  I had also heard from various forum sources that the best shroud depth would be several times the diameter of the hub to help straighten out the air flow. Well now that I had a radiator test bench setup, I figured it was time to put some of these ideas to some testing.

Before I go too far, I'd like to thanks the following sponsors for making this test possible with sample products to use.  

Radiator = XSPC RS120

A special thanks to Paul from XSPC for providing me with an RS120 single radiator.  The RS120 is an excellent radiator specifically designed for low speed fans (21mm thickness with approximately a 13.0 FPI fin density).  I particularly wanted to do this tesing on a single radiator because it enables me to utilize more fans and shrouds to test with, and I would expect to see similar performance gains on other radiators as well.   

Fans = Scythe Ultra Kaze

A special thanks to Linus from NCIX.com for providing me with the fans that were used in this testing.

They provided me with both the Yate Loon D12SL12 slow speed 47cfm 120x25mm fans, as well as this Scythe Ultra Kaze DFS123812-3000 high speed 136cfm 120x38mm fans.  Both fans present great value and a strong variety of cfm coverage for different setups.

Shrouds = The Feser Company (TFC) Shrouds

A special thanks to Martin from The Feser Company (TFC) for their new universal shround.  I'm really excited about this shroud because for the first time we have a universal and stackable shroud that looks great and fits on any 120mm fan radiator, these are very impressive and sharp looking shrouds.

Testing Method

I utilized the same testing methodology as I do for my other radiator testing, you can refer to those details here under "Testing Specifications".  The only thing I did differently is change the fans from my normal "Pull" to various push or pull or push/pull with different shroud depths.  The heat load was kept constant using a 120V 3amp variac and dialing in one of my 300 watt water heaters to the desired amount of wattage.  Fan RPM was held constant in all tests rather than voltage as I used my crystalfontz PWM controller which does a nice job at holding RPM very constant and easily controlled by a slider.

So in the ead I did the same logging of temperatures for extended periods of time to average out constant recordings typically over a 30minute period after at least a 30 minute warm up time frame (90minutes for the first warm up).  Then these tests were individually plotted to ensure stability in the delta temperature and averaged and summarized.

A typical test run for one data point.  Ambient temperature varies, but temperatures are continuously logged and averaged.

Test Results 38mm Single Fan, 2000RPM

Above is a look at the results from a direct delta vs shroud depth.  Very suprising to me, my standard method of using the fans in pull is not the best.  In addition the common thought that the deeper the shroud,  the better is also not true, both push and pull benefitted the most from a single depth 30mm shroud and push with one shroud yielded the best temperatures which was about 1.9C better than a fan in pull as I had typically been using.  It's very apparent to me, the hub dead spot affect is only one performance variable.  The closer proximity of the fan blades is also creating turbulence and additional efficiency closer to the radiator, and there is also likely some benefit in push due to air's ability to compress and increased density (improved thermal properties) that's playing some role in these results.

Now to take these results and compare relative to the worst case (Pull No Shroud) to get some percentages:

We gained over a single fan pull as shown above, nearly 18.5% using push with one TFC shroud vs pull no shroud.
Or for pull,  almost 16% using the shroud in pull vs. pull without shroud...very nice!!

It also appears the shroud depth of 30mm is nearly optimal in both push and pull, this was obviously designed at this depth for good reason.

Conclusion for Single 38mm Fan - 2000 RPM

The obvious winner here is to use a fan in a push condition if no shroud is used, which provides about a 9% gain, and better yet, a fan in push with one TFC shroud which is approximately 30mm in depth.  The fan in push with TFC shroud is producing nearly a 19% gain over a fan in pull, that's a significant gain!  I would highly recommend the use of shrouds where you have room to fit them.  It's a very simple modification that results in substantial radiator performance gains while reducing fan noise.

TFC Shroud = 10-16% More Performance & Less Fan Noise!!

Push/Pull Update Single 38mm Fan - 2000 RPM

I have also finshed a little more testing with push/pull configurations with two fans.  While I don't think very many people actually run fans on both sides, it's certainly an easy enough option.  In the end, the push/pull configurations was as expected better than either push or pull with a single fan.  In this scenario you gain the benefit of more fan pressure and result in an increased amount of flow rate.  I even purposely adjusted the RPM of the fans to match 2000 again even though running push/pull will have a tendency to increse the fans RPM slightly at the same voltage.  In addition I tested with shrouds again as well and found the single 30mm thickness to be ideal once again.  You don't see quite the shroud gains as the single fan scenarios, but still a worthwhile improvement considering you also get lower noise.  I found many folks have a hard time understand curve charts, so this time I broke it down into a basic bar chart with mini schematics of the fan setup to help clarify.  Somewhat suprisingly, you nearly get 2 fan performance with a single fan plus shroud.

Push/Push and Pull/Pull Update Single 38mm Fan - 2000 RPM

I had this whacky idea that maybe running two fans on top of each other on either side would be another viable option, so I tested that. Push/Push came out at 9.1C, which is barely any better than just one fan in push.  Pull/Pull came out at 8.9C which a little better, but still not worthwhile.  Bottom line, if you want to run two fans, run one on either side of the radiator in a push/pull configuration.

Push vs Pull vs Push/Shroud vs Pull/Shroud 25mm Fan - 1350 RPM

Now that we had the stronger and larger hub 38mm testing done, it was time to do a few quick tests to see if the trends continue with with slow speed fans.

In order to keep water deltas down a bit, I turned the heat down arbitrarily to 150 watts total and proceeded and complete 4 more tests.

Note these tests were run at a reduced heat load of 150 watts.

Push vs Pull has now changed places, but shrouds still continue to provide a good benefit, predominantly on the pull side similar to the larger gain in pull with the 38mm testing.

While I didn't complete the full series of tests, this was enough for me to see there is some emerging trends going on.  The push vs. pull question appears to really depend on the fan power.  I believe the gains in the 38mm showing more benefit in push is due to pressure and turbulence.  At these stronger speeds, the added pressure compressing the air along with higher turbulence across the radiator in push is providing the best benefit.  This appears to change however for the lower speed fan where pull actually does better.  I believe at this slower speed pressure and air flow is so slow that it's simply not turbulent enough to show gain in push and the radiator performs better with the more evenly distributed air flow over the radiator that's provided in a pull condition.

Bottom line

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.

Push Vs Pull - This depends on fan speed/power.  The high speed fans at 2000RPM with a 38mm fan thickness provided the best performance in a push condition.  The slow speed fans with 1350RPM with a 25mm fan thickness provided the best performance in a pull condition.  I would estimate that performance line is likely to cross in the 1500-1700RPM range where they are equal.  So.... slow speed = pull, high speed = push, medium speed = it doesn't really matter.

2 fans Push/Pull vs others - It was very clear that when using two fans per radiator section, that the radiator intself acts as a flow spreader and provided great benefit to straightening out the air for the second fan in a push/pull configuration.  Just like doubling up with a pump, there is a pressure benefit to doubling up on fans and this translates to about a 20-30% performance gain.  The added fan pressure simply means and added gain in air flow and corresponding performance.  None of the experimental pull/pull or push/push configurations provided any noteworthy benefit as it seams the air is simply too disturbed from the first fan to allow the second to perform properly when air is moving.  I would always recommend a push/pull configuration for two fans and add a shroud to both sides if space allows.

I hope this was useful information, again I'd like to thank my sponsors below.  With out their contribution in parts, this test couldn't happen...thanks!

Testing Sponsors

TFC Shroud Vendors