system resistance variation same DUT different directions

[cef3rd]

What could account for the difference in the system resistance curve for a device under test when the airflow direction is reversed say for a PSU from AC-inlet to connector vs connector to AC-inlet?

The air flows through the same components but expansion and contraction of air could differ with the two directions right depending on the layout, right? Are there any references on this topic?

 

[iainuts]

Hi cef3rd,
The Crane paper #410 is the bible on fluid flow and pressure drop. I'd suggest getting a copy, here:

http://www.cranevalve.com/tech.htm

As to why there could be differences, take a look at various restrictions in this paper. For example, pressure loss due to pipe entrance versus exit geometry.

For a "well rounded" rounded entrance, the restriction coefficient, K = 0.04

For a rounded exit, K = 1.0

So depending on if fluid is going into or out of this entrance/exit, the flow restriction will differ.

The reason for this has to do with how the flow stream of the fluid reacts to a restriction. Just downstream of the geometry itself, is a point where the fluid can actually be more restricted than at the geometry itself. For example, an orifice plate has a hole in it. The flow through that hole has to squeeze through this hole, but it actually must continue to contract or squeeze down past that hole. The point of maximum restriction (minimum flow area) is actually downstream of the actual hole. This point is known as the "vena contracta". So for various restrictions, the geometry itself causes restrictions (a vena contracta) at other locations, downstream of the actual geometry. How the fluid reacts to those geometry is dependant on how the fluid 'sees' the restriction. The example of a well rounded entrance versus a well rounded exit is another good example.

All of these restrictions are well defined by the Crane paper.

 

[ko99]

cef3D,

What is the air source (ie, power supply fan, system fan/blower, test chamber, etc) ?


ko (

http://www.ecooling.biz)

 

[cef3rd]

The air source is a test chamber (blower on a flow bench).

 

[ko99]

Reversing a fan can sometimes cause the results you mentioned. But with no fan I agree with iainuts that it's due to entrance/exit differences. Another reference on this:

Gordon Ellison, Thermal Computations for Electronic Equipment, Van Nostrand Reinhold or Krieger

One last thought: if the supply has a fan, even non-operating, it would certainly create more resistance in one direction...

ko (

http://www.ecooling.biz)