Improving fan efficiency by reducing the tip gap.

[Lynxman]

I decided to test how much cooling improves by making a tighter shroud for my CPU fan. I have always been annoyed at the giant gap PC fans have and wonder why Noctua for example doesn't reduce this gap to increase efficiency. I 3D-printed a 0,5 mm ring that fit inside the original shroud and ran Prime 95. It only took a few minutes to draw the ring and print it. The results were interesting.

The test system is an i7-7700 system which has been "delidded" to improve thermal efficiency (8°C improvement from delidding). The cooler is a pretty low end Akasa ak-959cu.
 

The first reading of core temperature at full load without the ring was 74 - 74 - 71 - 71. Idle temps before the test were 35 - 37 - 35 - 35. 
I then inserted the ring carefully, it barely fit without the blades rubbing, so the gap was very small, and let the idle temperature stabilize again. Then I ran the same Prime95 sequence. 
The new temperatures were: 71 - 72 - 68 - 68. Idle temps before this test were 34 - 36 - 34 - 33. The idle temp  after the test was 33 - 36 - 33 - 33 again.

Room was stable at 26.3°C in both tests.
 

A 2.75°C drop is pretty big improvement from just reducing the gap between the fan blades and the shroud by ~0.5 mm. The efficiency of the fan is obviously improved. It's well known in aerodynamics that keeping the gap between a blade and the shroud in a ducted fan, compressor or turbine improves efficiency (I have worked as a helicopter technician for 18 years).

I ran the same tests over again to confirm and the result was the same: With stock fan shroud gap: 73 - 73 - 70 - 70 With the reduced fan shroud gap: 71 - 71 - 68 - 68. A 2°C improvement the second time. Needless to say, I glued the ring in place with some hot glue after that test. The ring is the grey part around the fan.

 

BTW. previously I have added a cold air intake to my CPU cooler that dropped the temperatures by 15°C. It was over 90°C even with a much lower ambient room temperature than the above tests. It even lowered the temperature of my GPU and system. I like working with what I got instead of buying a new cooler. Downloadable here for inspiration: https://www.thingiverse.com/thing:2638688

Here's a picture of the ring installed.

Here's an older picture of the stock fan, without the ring:

[KaminKevCrew]

I think the main reason that manufacturers don't do this has to do with tolerances in their manufacturing processes. The fans will have to be much more accurately made for the smaller fin tolerance to work properly without rubbing, so that might have something to do with it.

[Lynxman]

I think that too, but the gap is very big on most fans, and I think Noctua especially are equipped to improve this area of fan design.

[EnergyEclipse]

Sorry if I may sound ignorant, but isn't you have done is have just gotten a duct for the fan thus to make to always intake air from the outside of the case? I would imagine the reason you see a difference is because you are in taking air from the outside of the case, because before it could have been receiving some hot air from your open air graphics card?

 

 

[Lynxman]

No. The duct was used in both tests. I removed the duct picture now to avoid confusion.The difference is the thin grey ring around the fan blades. Without it the gap between the fan blades and the wall is greater.

[tomaatvk]

If I'm not mistaken, this has to do with how, when the fan ages, the blades warp ever so slightly increasing their radius. This in turn would make the gap you see as a problem shrink. As a fan manufacturer you wouldn't want your fans killing themselves by letting the blades scrape the housing, so you increase the gap a litle bit to account for the increase in blade size over the lifespan of the fan. This was even said in either an LTT or a gamers nexus video if I recall correctly.

[WereCat]

What effect does it have on noise and turbulence?

[manikyath]

my bet as to why manufacturers dont make that gap tighter is because it is already so narrow that "careless/forceful mounting" can warp the frame enough to make the blades touch.

 

besides, they'd have to factor in long-term stuff like maybe over the course of 10K hours the fan blades could slightly warp from centrifugal forces, to name one extremely odd example.

[Lynxman]

Reduced blade tip gap gives less turbulence and higher efficiency. No perceived noise difference, but it performs better.

I have never seen a fan blade stretch over time. I've used PC's since the 80's and many of my current fans have been running almost 24/7 since 2010. They just don't spin fast enough. A 120 mm fan spinning 2000 rpm will have roughly 223 g on each blade which makes a 1 g blade weigh 223 g. Insignificant unless it's heated to 105°C, which is about the glass transition temperature for ABS.

[KaminKevCrew]

1 minute ago, manikyath said:

my bet as to why manufacturers dont make that gap tighter is because it is already so narrow that "careless/forceful mounting" can warp the frame enough to make the blades touch.

 

besides, they'd have to factor in long-term stuff like maybe over the course of 10K hours the fan blades could slightly warp from centrifugal forces, to name one extremely odd example.

I agree with the sentiment that careless mounting can cause issues. Also, not all cases are made completely square. I'm sure the steel on some cases could be warped enough that the fan shroud would warp, causing it to rub.

 

 

As for the blades warping over time, I highly doubt that ever happens. PC case fans do not turn fast enough to have much centrifugal force on them. EDFs, as used in remote control jets have blades that are made of the same stuff in many cases, and those don't warp enough to rub on even tighter tolerances, despite the fact that they run at 20,000+ rpm easily, even when small debris can get sucked up and hit the blades. It's really not an issue here.

[KaminKevCrew]

2 minutes ago, Lynxman said:

Reduced blade tip gap gives less turbulence and higher efficiency. No perceived noise difference, but it performs better.

I have never seen a fan blade stretch over time. I've used PC's since the 80's and many of my current fans have been running almost 24/7 since 2010. They just don't spin fast enough. A 120 mm fan spinning 2000 rpm will have roughly 223 g on each blade which makes a 1 g blade weigh 223 g. Insignificant unless it's heated to 105°C, which is about the glass transition temperature for ABS.

I absolutely agree here - as I posted above, many remote control aircraft have fan blades made of similar material, and they have absolutely no issues with spinning many times faster than PC case fans under much more demanding use cases.

[Lynxman]

Well if one is careless enough to warp a fan during installation then it will be noticed and remedied just as easily as if one lets a cable rub against the fan blades, which is probably much more common. 

[manikyath]

2 minutes ago, Lynxman said:

hich makes a 1 g blade weigh 223 g. Insignificant

now keep that there for 10 years, on something like for example a venturi AF:

 

as opposed to.. you know, some sunun maglev.

[Lynxman]

Just now, KaminKevCrew said:

I absolutely agree here - as I posted above, many remote control aircraft have fan blades made of similar material, and they have absolutely no issues with spinning many times faster than PC case fans under much more demanding use cases.

Me too. I've had several EDFs.  Tip gap is a critical factor for EDF efficiency as well.

[Lynxman]

Those thin Venturi AF blades probably weigh 0,3 g each, and only spin at 1400 or so RPM.   I'll do the math...

 

EDIT:
32 g blade weight at 1400 rpm with the blade CG at 50 mm from the center and 0,3 g static blade weight. Absolutely negligible.

[manikyath]

1 minute ago, Lynxman said:

Well if one is careless enough to warp a fan during installation then it will be noticed and remedied just as easily as if one lets a cable rub against the fan blades, which is probably much more common. 

you REALLY dont need much force, especially on bigger fans, and when your tolerances are tending closer to zero, force required tends closer to zero.

 

there's a golden middle way somewhere in there, which i guess most manufacturers have found.

 

i'm sure that noctua could easily get their fans an extra edge in silence and efficiency, but i'm also sure that they dont want to degrade reliability.

[KaminKevCrew]

2 minutes ago, Lynxman said:

Me too. I've had several EDFs.  Tip gap is a critical factor for EDF efficiency as well.

It is indeed - It's really fun to look at full scale jet engines, because those things will have a tip gap so small you can barely stick a piece of paper through it, yet they're massive, massive things. Jets are really cool.

[manikyath]

3 minutes ago, Lynxman said:

Those thin Venturi AF blades probably weigh 0,3 g each, and only spin at 1400 or so RPM.   I'll do the math...

 

EDIT:
32 g blade weight at 1400 rpm with the blade CG at 50 mm from the center and 0,3 g static blade weight.

you can bend them out with one finger, without holding the fan in place to stop it from rotating away from your finger, *thats* how flimsy they are.

[Lynxman]

Just now, KaminKevCrew said:

It is indeed - It's really fun to look at full scale jet engines, because those things will have a tip gap so small you can barely stick a piece of paper through it, yet they're massive, massive things. Jets are really cool.

Yep. A large gap causes more air to flow back and can cause a stall. 

[Lynxman]

That Venturi fan looks like it already has a much better tip gap than most fans anyway, so it'd be hard to improve on unless you apply a tlayer of tape around or something.

They are flimsy because the forces are super low. 1,3 W fans are not able to warp modern plastics significantly.

Edited December 10, 2017 by Lynxman
Poor wording.

[KaminKevCrew]

2 minutes ago, manikyath said:

you can bend them out with one finger, without holding the fan in place to stop it from rotating away from your finger, *thats* how flimsy they are.

They might be flimsy when you push on them, but the amount of force on them when they're spinning still won't likely be very high. If there was enough force to warp the blades, you would definitely be able to hear it. Fan blades that bend while they're spinning do not make a pleasant sound. 

 

I also bet that when you put force on the blades, they snap right back into place. That's because plastic has some pretty good memory in it. It's not like it's going to stretch from use anyway - you'd notice something like that as white lines on the plastic itself, which you never see.

[manikyath]

Just now, KaminKevCrew said:

They might be flimsy when you push on them, but the amount of force on them when they're spinning still won't likely be very high. If there was enough force to warp the blades, you would definitely be able to hear it. Fan blades that bend while they're spinning do not make a pleasant sound. 

 

I also bet that when you put force on the blades, they snap right back into place. That's because plastic has some pretty good memory in it. It's not like it's going to stretch from use anyway - you'd notice something like that as white lines on the plastic itself, which you never see.

yes, and i've had several cases of said venturi AF "touching rim" in action due to less force than you'd imagine, coming to say, its in a dusty spot, the chance of dust or dirt getting in the way is probably a factor as well.

[Lynxman]

I think they are the exception when it comes to pc fans. The fans I'm talking about do not have such a small gap as those fans. Most fans have gaps that are several mm wide.

[Lynxman]

Example: http://noctua.at/media/catalog/product/cache/2/image/9df78eab33525d08d6e5fb8d27136e95/n/o/noctua_nf_f12_pwm_3.jpg

[manikyath]

2 minutes ago, Lynxman said:

I think they are the exception when it comes to pc fans. The fans I'm talking about do not have such a small gap as those fans. Most fans have gaps that are several mm wide.

happen to ever seen how much dust tends to get built up in a system that doesnt have someone caring about a 2.75°c difference sitting in front of it?

 

i've seen a "several mm wide" gap around fan blades get stuffed by dust in less than a year, having a <0.1mm gap in places like that is a jam waiting to happen.

 

besides, i hope you didnt print that shroud in PLA