Acoustic of air cleaners with/without filter elements

[SomptingGuy]

Does anyone have any information (links, measurements, plots, anecodotes, or even uninformed hand-waving opinions) about the effects of air filter elements on the acoustics of automotive air cleaners?

Typical engineering practice is to assume the elements are acoustically transparent. My intuition says that they _probably_ give broadband attentuation and could possibly inhibit large-amplitude standing waves (in the way porous hose can). I don't see how they could shift resonant frequencies around though.

Can anyone shine any light on this?

 

[GregLocock]

Measured a few, never found a convincing effect at low multiples of firing frequency.

If you've got folded paper in there then you might get some interesting effects around lambda/4=thickness of filter pack, but it seems unlikely to be significant.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[SomptingGuy]

Thanks.

I've been given a couple of TL curves and asked to comment. There's a clear attenuation caused by the element, which increases with frequency. But what's odd is that with the element in place, the TL shape is similar, but the frequencies are all reduced (proportionately). It's almost like the speed of sound has dropped. As with all these situations, I didn't witness the testing and temperature variations from test to test are denied.

Assuming that the element doesn't have some heavy gas trapped in it, are there any other potential causes of frequency shifts?

 

[johnab]

similar question related to above post so would like to ask here,

(1) To optimize volumetric efficiency, if duct length need
to tune is any preference for dirty or clean side of
duct? (dirty side -before air cleanr & clean side
refer to after air cleaner connected to Th/body)

(2) Can we analyze clean/dirt side duct selection without
going for prototype/actual testing ? any software
suggestion

 

[GregLocock]

If the frequencies have all dropped then the mass of the filter element has coupled into the mass of the air. Well, that is a surprising result to me.

I haven't measured the TL of filters, only the noise.

johnab - you have to tune the clean air duct. the dirty side will be decoupled by the filter box's volume.

Ricardo Wave is the usual recommendation, I've never used it.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[ivymike]

it does sound surprising... could it be that the filter element was moving as a (somewhat heavy) diaphram on top of the pile of "clean" air, and the frequency response was being measured on the "dirty" side? (picture an eardrum, in reverse) Were all the frequencies really "shifted" proportionally, or did the lower frequencies shift more than the higher ones?

 

[GregLocock]

I'm kind of hoping Tom's busy working it out.

I don't like this theory - after all, microphone diaphragms, however heavy they are, do not change the frequency of the measured noise.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[KiwiMace]

I'm completely out of my league here and pose this more for my own understanding:

But in a microphone you measure the frequency of the diaphragm, this problem is transmitted noise. If we scale the diaphragm up to a wall, tranmitted noise is clearly restricted in frequency?

 

[GregLocock]

Sorry, I don't understand your question, but would add that I was wrong. If the microphone was sufficiently large , compared with the rest of the system, then it would couple into the modes, and it would affect their frequencies.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[SomptingGuy]

Greg,

That is a fascinating theory and I must admit I do find it plausible and attractive. The paper would have nearly zero stiffness and could consitute extra distributed mass.

ivymike,

A diaphragm would be a spring-mass-damper system with all mass at one point. I think the filter element would act as a distributed mass.

CinciMace,

The diaphragm is a microphone is designed to have its first mode well above any frequency you might want to measure with it.

johnab,

Greg is right - tune the lengths in the clean side to optimse volef. The dirty side only really has an effect on radiated orifice noise.

 

[SomptingGuy]

I don't quite know how google found this (I thought it couldn't search secure sites??) but it does seem relevant:

https://www.events.mondial.at/ei/rs.esp?id=172&scriptid=spcom3a&sid=577

The two phrases of interest are:

"Fibrous structure of the filter element is modeled as a micro-perforated panel using the flow resistivity and porosity"

and:

"enhancement of TL at troughs and shift of troughs to low frequencies"

So there is no distributed mass in the model, just extra damping. And I guess the extra damping could be what reduces the frequencies of the forced modes?

 

[SomptingGuy]

Well, I've been doing some simple 1D models using varying levels of absorptive material (to change flow resistivity and porosity) and cannot get the sorts of shifts I've seen in the plots. So it's back to the distributed mass idea for me. Interesting stuff.

 

[ivymike]

what on earth is the interpretation of a distributed (as opposed to point) mass arranged perpendicular to flow in a 1-D flow system?

 

[SomptingGuy]

There is no flow in this sort of TL test, so if the filter is made from folded paper, it's motion may match that of the air around it (nodes and antinodes). Just hand-waving though.

 

[GregLocock]

That's how I'd think of it, it's just a patch of super dense air.

So, do the measured results show a change in damping as well as a shift in frequencies?

That paper sounds exactly what you are looking for, here's another one (can't find it on the web)

The insertion loss of air - filters in HVAC systems R. Dragonetti, C. Ianniello and R. Romano

Internoise 2004

Looks at the 1/3 octave IL of the rather heavier elements they use - typically 3 dB . It also includes a very useful equation which I am not going to attempt to reproduce here.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[SomptingGuy]

Damping - don't know. I just have paper copies of TL. And my company stopped getting Internoise proceedings after 2002! Maybe I'll ask our library to get me a copy from the British Library. Think it's worth it?

 

[GregLocock]

It's good on the 1/3 octave broadband effects, it doesn't really discuss narrowband stuff. For a car, no, I don't think the odd 0.3 dB in broadband attenuation is really going to matter much compared with the effect at resonances.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.