Hardware What's the best supersize case fan?

Agreed. Computer cases have parts which cause a minute amount of air resistance, and the air incoming has to displace the air leaving the case. Then there is the bag issue. All of which are so minute as to vary the test results by probably less than 0.01%. Well within the commonly accepted margin of error for most applications.

PS: No, I don't have a better way of testing. Just throwing that out there.

 

I have a hard time agreeing with the technique used here.
What if you already have other fans for intake / outtake? Our cases aren't like big black plastic balloons that we wish to inflate.
On the contrary we want to enhance airflow through the case via intake+outtake.
To be fair to these fans you should also measure how long it would take to empty your plastic bag as well. But all in all your method of testing looks flawed for measuring airflow as you have rising air pressure countering the flow.
You should at least do a "fill bag"+"empty bag" and then average that.
I believe you cannot measure the airflow this way unless you calculate lost efficiency based on counterpressure from the confine into which you blow the air.

As I scour the net I find more and more reasons to believe the current method employed is incorrect. Especially in a scenario like the bit-tech one where the current method would probably have a lot smaller margin of error if you were testing high velocity fans like those you use for bathroom ventilation and such.

I think your own results said it best when one of the fans failed to inflate the bag.

 

You might actually be able to do a ghetto CFM measurement that is a bit more true if you have a cylinder of a certain diameter (larger than the fan blade area) and long enough for the turbulence to become fairly "calm". Then having an anemometer somewhere along the cylinder would probably give you a really good indication of airflow.
I'm not going to draw it in ASCII... you probably understand what I mean. Otherwise PM me and I'll sketch it up for you It would be a fairly "cheap" way of doing this.
Maybe it won't work.. I'm not an engineer.. . but it's an idea of how to do it without having to deal with counter pressure.
In short: If you know the volume of a cylinder and you can measure how fast this volume travels then you have your CFM...

In such case you might be interested in the TSI 9515 VelociCalc Air Velocity Meter or something like it.
I found a price for one here: http://www.raeco.com/acatalog/TSI_VelociCalc_AirVelocityMeters_Purchase.html
Not as ghetto as your current rig.. ;-)

 

Does it really matter if the cfm figures are completely accurate (compared to some perfect testing method) if the method provides results which are consistent and able to be compared?

 

Agreed

Besides that, I think that such an improvised approach is quite appropriate for a site dedicated to the modding community.

 

I would have loved to have seen how the silverstone air penetrator fan compares to the other fans, it seems like the best fan ever made from what i've read about them so far.

 

Bad link on page 6 - 'how much air it can push per Db'

Even if that bag test is not very accurate, it does provide a comparison between the fans.

But wouldn't it be a good idea to measure the speed of the air coming from the fans, as well as the amount of air. One may assume that the speed of the air would be the cfm divided by the area or something. But some bits of the fan provide a faster air column than others. A fast air column could penetrate the case further, providing superior cooling.
I must confess to have not thoroughly read through the article, so I may stand to be corrected, and what I suggest may have already been done.

 

oh, just noticed that this article is pretty ancient. Sorry :$. Ignore first part and most of the previous comment

 

Apologies for the whole necro thread thing but heres an idea.....

How about having the "black bag scenario" modded/taped onto the outtake of a tested/sealed pc?

I realise that it wouldn't be a fan test any more but i think it would provide a more "real world" test of what is going on and the throughput of the fans could be more accurately measured.

Nic