Acoustic attenuation using sheet lead 2

[Warpspeed]

I have an electrically driven high speed centrifugal air blower mounted in an air tight box, and the aim is to make it as quiet as possible. The external size of the enclosure is roughly a three foot cube.

The box is made from one inch MDF sheet, bolted and sealed to a heavy welded steel frame, and the blower is spring mounted with a natural frequency of only 1.4Hz, so conducted noise is not the problem.

In open air the blower noise is 98dba, and enclosed is 76dba. The problem is panel vibration and the low inherent damping of the wooden panels. My next step is to try lining the box with lead sheet, but have no idea of how thick the lead should be. I am hoping for at least another 20db improvement, but 30db would be better.

Dissipating some of the sound inside the enclosure with a sound absorbent lining may also help, but I doubt I could ever get the desired results by this method alone. Any sort of internal lining would be limited to one inch thick.

Could someone offer practical advise on suitable thickness of lead sheet ?

 

[pennpoint]

Warpspeed
Check out this site;

http://www.soundcoat.com

Good Luck

pennpoint

 

[MikeyP]

Warpspeed,

General principles of enclosure design:

1a) Reduce the sound energy inside the box. There is then less noise inside to be transmitted outside. This means acoustic absorption. Most foam absorbers will be ineffective at low frequencies where your enclosure problems are worst. If your machine is running at a fixed speed most of the time, then you could look into tuned absoption panels which work on the Helmholtz resonator principle.

1b) Reduce the sound energy inside the box. Find out if it is possible to make the machine inherently quieter. e.g. balancing of the fans.

2) Increase the mass of the enclosure. Heavier is better, especially at low frequencies. This is because the majority of low frequency vibration in enclosure panels is "forced" vibration". Added damping will not be effective against this kind of vibration. With this in mind, the thickness of lead should be as thick as you can sensibly get. I don't think that the lead should be bonded to the enclosure but rather attached loosely.

3) Increase the damping of the enclosure to reduce resonant vibration. In your case, resonance will only be noticable at the higher frequencies (above the "critical frequency" of your MDF). However some constrined layer damping may help but it is likely to be a secondary effect.

4) Make sure the enclosure is airtight otherwise all your careful noise control measures will go to waste. Seal the joins in the enclosure with gaskets (PVA glue is probably as effective as anything and it flexes a little too).

5) An enclosure is only as good as its weakest point. Pay special attention to areas where ducts and cables enter the box. Make sure they are sealed effectively and flexibly.

Hope this is of use

M

--
Dr Michael F Platten

 

[Warpspeed]

pennpoint, I am located in Australia. Ordering the materials, shipping, and import duty is just going to be too much trouble, but thank you for supplying the information.

Dr Platten, wonderful advice, exactly what I was looking for.

1/ I have already tried to quieten the naked mechanism as much as possible by making the frame stiff, and with full dynamic balance of the high speed parts. The bare spring mounted motor blower assembly weighs 280Kg and there is no obvious detectable low frequency vibration or mechanical resonance. The nature of the noise is a broadband mid frequency roar caused by air movement in the blower casing, the blower tip speed being around 450 feet per second.

I understand what you are saying about frictional absorption inside the enclosure, and this will be my next step. The sound is echoing around in there, releasing energy into the flat panels on every reflection. Special acoustic foams are not readily available here, or at least not as far as I can discover. But conventional foam rubber is abundant. It is available in all sorts of densities up to almost solid rubber sheet. What should I be looking at here, for optimum absorption? At a guess most energy might be in the range from around a hundred to a couple of KHZ.

2/ The empty wooden enclosure weighs 310 Kg and the panels are sealed with silicone and bolted around the edges every five inches to a steel angle frame welded at the corners. It is certainly strong, and reasonably stiff. But tapping inside, like knocking on a door creates plenty of external sound. The wooden panels are very "live" with little internal damping within the wood. Also I suspect the natural resonance of the "knock" test unfortunately coincides pretty well with the bulk of the energy radiating from the blower.

3/ I had intended to secure the lead sheets around the edges with the same bolts, and separate the lead sheet from the panel in the centre by using glue and soft rubber pads. This should prevent the sheets from sagging, and maintain separation, but isolate any higher frequency vibration with a small air-gap. I will try the lead sheet idea second.

4/ The enclosure is certainly airtight as it is open to the suction side of the blower. The internal pressure operates around 2psi below atmospheric, and this is the reason for the heavy box construction. The front entry panel is bolted and gasketed and is also guaranteed completely airtight. External atmospheric pressure forces all the joints into compression, and this also helps. The whole thing has been leak tested and is completely sealed.

Noise does not seem to be coming from the entry and exhaust pipes, as the curved surfaces are pretty stiff. It is really only the flat panels that are vibrating in the centre. This is confirmed by moving around outside the enclose with a sound level meter only an inch away.

Once again, thank you for the excellent advice.

 

[MikeyP]

I think the foam rubber is something you will have to experiment with (especially seeing as you have a ready supply). To be effective, the foam must be of the "open cell" type. I would not expect much effect from foam below 500 Hz. Mineral wool is another readily available alternative. As these things are cheap, it is worthwhile trying them.

Re damping of the panels: There are very few modes in general at the low frequencies in these panels; There are even fewer which will radiate effectively. So I am still not convinced that added structural damping will be of much use (try search terms such as "critical frequency" or "coincident frequency" and "resonant transmission" and "mass law" for more info). However, your lead lining will add some anyway.

M

--
Dr Michael F Platten

 

[Warpspeed]

I will give it a try, but it will take a short time to organise this. Meanwhile, I will try to borrow a spectrum analyzer to see what I am actually dealing with, and better evaluate individual improvements.

Once again thank you for your help.

 

[vanstoja]

Fan/blower impeller blade passing (IBP)noise generation should not be taken lightly when installed in cabinets and close enclosures. If there are struts or diffuser vanes downstream of the impeller that are closer than one impeller blade chordlength (for axial flow fans)then high level airborne IBP noise can couple to cabinet walls. Also, distorted inlet flow to the fan/blower particularly when transporting streamwise vortices from flow blockage by upstream components is another source of IBP airborne noise that can couple to cabinet walls. Also of concern are air entry grills and louvers through one or more cabinet walls where gridwires or louvervanes can generate a structureborne source directly in the walls of the cabinet.
High-speed driver motors are particularly troublesome because IBP noise levels are particularly sensitaive to impeller speed. Brushless DC motors driving fans for cooling electronic equipment in cabinets run at speeds of 6000-8000 RPM and higher which has been found to be particularly bad when IBP sources are present. The high switching frequencies of these kinds of fan drivemotors also can contribute signifcantly to enclosed cabinet noise.
One possibility being explored for controlling cabinet wall vibrations in fan-cooled electronic component cabinets is active vibration control of wall vibration by piezoelectric "patches" which can be located at antinodes of the various plate modes of vibration. This necessarily entails modal testing of the cabinet walls to identify the critical modes and their modeshapes.

 

[GregLocock]

As a quick, and possibly illegal (stupid health and safety laws), hack, I'd try covering the box with 2 inches of rockwool and then covering that with 3mm thick roofing lead, beaten to the right shape and using self tappers to hold the panels of lead to each other. The objective should be to have the lead completely isolated from the underlying structure.

I doubt that this will give 20 dBA, since that is a rather extreme objective.

I do recommend you try and get a handle on the frequency, even just using a scope should get you close enough. 450 Hz is around gear whine frequency, that is why i recommend the construction above.

To get 30 dBA I think you'll be looking at either (1) some sort of tuned absorber or (2) concrete and bricks.

Cheers

Greg Locock