funny airflow effect from fan

Hi all

having recently changed a noisy fan (knackered bearings), I played a little with the old one on the test bench mainly to see whether it work any better in a particular orientation (horizontal/vertical, one side up or the other...)

while doing this I noticed a rather strange effect which happened if I impeded the airflow going into the fan

http://homepage.ntlworld.com/gbmain/ePGB/hj/odd_fan.jpg
...shows the simple setup, with the fan hooked up to a power supply and a piece of tissue arranged to dangle into the air coming out of the fan;

http://homepage.ntlworld.com/gbmain/ePGB/hj/odd_fan.avi (4.1MB)
...shows what happened.

In the movie clip:
- first I turn on the power, the fan starts up and the tissue clearly gets blown away from the fan;
- then, when I bring a piece of card up to approach the intake side of the fan (trying to hold it parallel to the fan's frame), the airflow goes into reverse

the fan is rotating in the same direction throughout BTW (before anyone asks!)

The effect seems to work well with the card spaced ~1cm from the fan's frame.

strange but true...

best wishes, HJ.

 

it made me remember, quite a while back, someone had posted about a fan, being unsure which direction the air was meant to go! at the time I just thought "for any sakes it's obvious" (anyway the things usually have arrows on them)

but then, I noticed this reversal by holding my hand up close to the fan - and it felt like it was blowing rather than sucking: that's when I got interested (and, couldn't help but wonder whether something of the sort had confused the issue many moons ago...)

moral of the story - never underestimate the question even when it seems stupid...

==

the noise from the fan changes a bit when the airflow reverses - goes into turbulence or something?

it wasn't really meant to be a serious hardware thread - just an amusing little oddity - conjuring trick, sort of... which is why I started off over in General Discussions.

it does make me wonder, though: what happens with a grille close by the fan (or a perforated metal plate like the blue side of the 'scope to the left of the picture)?

straightforward if it's an open mesh; but what about, say, 30% "hole" and 70% "solid "? Hmmm. And what about blocking the air going out of the fan?

very best, HJ.

 

Absolutely love your posts re: this subject. This should be required reading for all system builders. I tried to explain this in this thread and felt like I got beat up and sent home with two black eyes. What's that old saying about trying to argue with ignorance?

Obviously, I'm stretching a point but that avi clip you posted speaks volumes. I've been playing around a lot with mesh filters, screen filters and the like and been trying several different permutations - all on the intake side of the equation - some on the side and some on the front (front main as I also run HDD intake fans above the front main in a couple of cases). What you have just posted is going to take me to the next level and I'm going to try some yarn strands to measure things visually but will need to make a couple of offset grids to explore this further.

Please keep us updated re: anything related to this and I will try to reciprocate. (Am going out of town end of the week and it may be 2-3 weeks before I get back to this)

Again, thanks so much - absolutely fascinating information.

Regards,
Rockster

 

Remember the floating beach balls being held suspended in mid-air by a vacuum cleaner blowing on it? I suspect this is something similar and you will not get the same results if the fan is mounted. If you get some tracer smoke to watch air flow, you will see some odd behavior.

Since case fans are not typically run in the open, but rather mounted, this is not a realistic experiment.

 

Ah yes, floating beach balls and do you remember the bowling ball in an acrylic tube? Regardless, I think this experiment is quite telling.

Stuff some ribbon cables in front of or behind a fan and it isn't going to matter whether its a closed environment or not. Now don't get me wrong, no one has ever read a post where I've advocated running with open side, even on a temporary basis (except to isolate a noisy or failing component). However, some of the anomalies I've seen are tough to discount and I guess that's exactly what you are saying Bill. Its like using port tubes directly over a heatsink fan venting to the outside (like the old carburetor velocity stacks) - what a difference. Now, try the same thing with a shorter tube that leaves a gap - ouch.

There was a time when I thought more was better when it came to cooling fans - I'm embarrassed to admit how long ago that was. I'd also assume several members have played with positive and negative case pressures, card coolers and slot mounts to name a few. Dare I mention a TennMax lasagna fan? - yup, I'm using one on my hacked I-Opener.

I've pretty much settled down to being an advocate for "good, well balenced air flow" and that encompasses a lot more than just fans, but I find all of this to be fascinating. Look at what's being done with heat pipes these days - who would have imagined it 6,7 or 8 years ago.

Keep it coming HJ.

 

lol If anyone denies ever feeling the same, they're probably lying - or still feels that way.

Well, "from" the outside, to be precise!

Agreed - it is hard to get flow when there is nothing to channel the air. FTR, I have advocated running that way, but while blasting a desk fan inside to see if it stopped crashing after a bit.

Oh? What is it telling you? All it is telling is there is odd behavior with that fan when put in that scenario. Since those fans are not designed to be used in that scenario, it is an invalid test. It may be interesting, but I don't find it telling - not from any technical application perspective - except perhaps aeronautically, you put it in a stall.

From a heat displacement perspective, you want so many CFM. If the fan you use is not able to produce the CFM when drawing through a filter screen, you need to turn up the RPMs on that fan, or get another fan that moves more air.

 

Hi again

I've now tried a few more things - using a fan with decent bearings, and trying in a simple enclosure (I used an old single fan style PSU case)

what I found - although nowhere near as theatrical as that little movie clip! - has made me want to go at least a bit further with this.

please, don't expect anything more too quickly, as i) I'm busy with other stuff, and ii) simply haven't got the space available at the moment (there's a lot of "work" stuff on the other side of that Tektronix, it's only a small workbench & it's chocabloc).

Well now. And I had previously dismissed those flash cases with transparent sides as mere gimmicks to enhance the "showing off" value, but with no real practical purpose. Wrong, HJ!

posting this quickly mainly trying to catch Rockster (Hi Rockster ) before he travels. And I too must disappear for a few days imminently, family comes first.

best wishes, HJ.

 

Well, I tend to watch my monitors and don't want to be distracted by a flashy or noisy case. So, IMO, all flashing lights do is nothing for performance, consume more power, add some heat, and do nothing for performance - worth repeating. But a transparent side can be nice to take a peak inside every couple weeks to ensure there is no dust buildup.

And with all due respect, as for testing fans, you must have the necessary test mockups with precision controlled power supplies, expensive measuring equipment, skills and training to accurately simulate realistic electrical and operating conditions, and then consistently, accurately measure CFM, sound pressure levels at various distances, rotation speeds, and more.

And what would you compare it too?

I don't think any results you come up with will be pertinent to computers, per say, or case cooling. You may want to look at "lift" and see how things around a wing surface affect it, and then become a pilot.

 

Hi again

fear not Bill, rest assured that I'll only be using the most modern, state of the art measuring equipment available.

Happy April 1st, everybody

best wishes, HJ.

 

OMG - that's not an old Sony TC-366 slanted front reel-to-reel, is it?

 

well spotted Bill, very close: TC-377. Quarter track and a complete pig to work on inside. Not my favourite, I much prefer these (300k pic ~lifesize) 3rd hand to me in 1973, run hard since, still going strong. You can see how much he's been used by the way the paint's worn off! Folks nowadays simply don't understand the word "reliability ".

All seriously modern equipment at this end - no messing

best wishes, HJ.

 

Darn! I should have picked up on the reel locks instead of holders. But you can see why I was mistaken here: http://www.angelfire.com/electronic2/vintagetx/SonyTC377.html

Or fidelity, I am afraid. While I appreciate the absence of hiss, analog to digital, and back to analog again results in something more missing, that only a few recognize.

 

This is why forums are cool. I learn new stuff all time. Thanks Christer.

 

Well, I did spend 24 years in the Air Force - but it was on the ground, fixing tower radios. But I do understand lift.

 

Not sure what part of the rest of your post you are saying is the misconception.

I agree with what you are saying - but extended flaps have a greater purpose- they expand the surface area of the wing, which, when forward motion is added, increases lift. Not pretending to be an aeronautical engineer, it matters not to me whether that is air pushing down (with Newton's opposite effect being the plane is "lifted" upward) or if the plane is just pulled upward, the effect is the same - the wing will attempt to lift up the plane. The affect is easily seen on bigger planes where the wing tips can move (be pulled) up more than 6 feet during flight. Follow a convertible with its top up on the freeway and you can easily see it being sucked up too.

Greater surface areas provide greater lift at slower speeds - a good thing while landing, and a really good thing when taking off, and running out of runway. Yes, the flaps can be extended down or up to create drag - but increasing surface area I believe is the key.