For all you fan curve experts out there.....

[noslo21]

I know how to measure a fan curve over a finite pressure range. I also, know how to measure the system response of the package that the fan will go in. As we all know, the theoretical operating point of the system would be intersection of the fan curve and the system curve. However, oftentimes there are obstructions in front of the fan when installed in the package. This directly affects the performance of the fan. Hence, the operating point has now moved. So determination of the "actual" fan curve can now be tricky condsidering the amount of variables. I know, when the fan is in its package operating at full power and the static pressure in a flowtube is zero, this point represents the "actual" operating point on the system curve. Now the question... what, if anything does testing this way with the fan mounted in its package on a flowbench testing over a finite range of static pressures tell me? The end point of such a curve would be the operating point... but can the rest of this data be used for anything? Or is it completely meaningless data?

 

[ko99]

noslo21,

In theory, you measure the system resistance curve so you can predict the operating point with different fans or fan speeds.

In practice, I don't see much value in measuring this curve if the fans aren't running. I can think of two reasons.

1. You can't isolate the fans from the system resistance, they are part of the system and have resistance even if the blades are freely rotating (blowers are worse).

2. The system resistance changes with air distribution. The uniform flow from a flowbench does not behave like the ring of high velocity swirling air from fans.

The method you described for determining the system operating point with the fans running is much more useful. You can also use this data to estimate the flow at different fan speeds, etc, with about as much (or more) accuracy than the other method.



ko (

http://www.ecooling.biz)

 

[noslo21]

ko99,

I completely agree with what it is your saying. I don't believe that I was very clear with what I'm asking. what I'm trying to figure out is if there is any merit to measuring a "fan curve" (just like one would do, to dertermine the fan performance) except by placing the fan in the package. Hence treating the whole assembly as one fan and measure that over a range or static pressures?

 

[ko99]

If I understand correctly, you would have to introduce arbitrary impedances (in addiion to the package) to create multiple points for a fan+package fan curve.

I can't think of any value for this, can you?

ko (

http://www.ecooling.biz)

 

[noslo21]

I can't think of any merit to this... However, there are people, who will remain annonymous... that would disagree. What they are doing is testing by the method described above, and then using this information to determine the flowrate in a package for a given static pressure. Or as they put it... "the flowrate at a given backpressure. The issue is that a specified amount of additional backpressure is speced into the package. We are supposed to determine the flowrate based off of this number. This posses a problem. I don't see how they can know how much "additional backpressure" needs to be added if they don't even know where our package operates on it's system rsistance curve. The performance of our package in the customer's system is a moving taget depending on how fast our fans run.