Partition Walls

Partition walls, Ceilings, Windows, and Doors construction and design to minimize sound transmission

Partition Walls

Postby Scott R. Foster » Mon Feb 09, 2004 4:17 am

Partition Walls

Isolation is the act of separating yourself from a sound source [or the rest of the world from such], and this is accomplished in studios, audiphile listening rooms and home theaters by building partition walls designed to have high sound attenuation characteristics. The most important factors in increasing the isolation caused by a partition are:

1. Make it air tight.  Partitions with holes will leak sound just as surely as boat hulls with holes will leak water. Include electrical outlets etc. in your search for holes to plug. You can remove this concern arising from wiring by running through the slab for connections to other rooms, and then using surface mount conduit for runs internal to the room - if you don’t mind the way it looks – otherwise use "putty packs" on every box that penetrates the wall and understand that these holes will diminish performance;

2. Make the wall heavy. Even an air tight partition can still “leak” sound by being induced to vibrate itself and thereby inducing vibration in the air of the room on the other side.  The more massive (heavier) a partition, the more energy required to induce vibration, and therefore the heavier the wall, the less energy it will let leak through;

3. Disconnect the two sides of a wall from each other. No matter how heavy, every material has resonances [put your ear to a railroad rail and then have someone a 100 yards away smack it with a hammer – OTOH maybe not eh?]. Isolation can be increased by physically disconnecting the two sides of a framed partition, so the framework of the structure, when vibrated on one side, does not conduct the vibration efficiently to the other side through the connectedness of the two sides. Incorporating this concept into a partition is called “mechanical decoupling”;

4. Dampen the cavity resonances. Even if the two sides of a wall are utterly mechanically decoupled, as in a framed wall with two completely separate stud frames, the air mass in the wall cavity will act like a spring connecting the two sides. This represents a Mass-Spring-Mass system - abbreviated as MSM. The air is a key part the spring in framed wall MSM system and it can be damped by introducing absorbent into the cavity. Cheap roll insulation works just fine for this job;

5. Make the two sides of the wall from different materials (or different amounts of the same material) so that the resonances of the two sides do not match. Sounds not attenuated by one panel may well be attenuated by the other if resonances do not match.  Do not however reduce the mass of the wall system in order to achieve this result.. if you can add a layer of drywall to make the two sides different, this is a good idea... but never take a layer off, or make it thinner just to get a difference in the mass of the two sides.

6. Making the connection between the two sides of a framed wall assembly less stiff, with a looser "springiness" between the two sides, enhances transmission loss [TL].  The lower the resonant frequency of the "spring" created by the framing and the air cavity then the better the walls works at stopping the transmission of low frequency noise. There are a numbers of types of wall parts designed to make the connection of the wall panels to the wall frame "looser".  Also, a dampening / constraining layer added to a wall panel can limit the vibration of the panel and enhances TL - and there are products designed to do this which work well. Fundamentally a wall with weaker, more wiggly, framing will make for a better isolation partition and a leaf made of a panel sandwich which is internally damped will not conduct vibration to the wall system as efficiently.

Isolation partitions are commonly rated in North America by a standard called STC, which stands for "Sound Transmission Coefficient".  While of some usefulness, this standard does not take into account very low and very high frequencies (it was designed around office noise, not musical noise). Nonetheless, STC is a rating commonly seen; just remember when you come across STC numbers, the extended low and high frequency content of musical noise will probably leak through a partition to a greater degree than its STC rating would suggest. Always look for testing of wall assemblies under consideration for you project that show performance per freqeuncy band - and focus on the low frequency end of the spectrum as this is the weak link for isolation musical noise.

Do not turn a two leaf wall onto a triple leaf wall as this will actually reduce isolation at some frequencies… do not put internal partitions in your wall design. DO NOT BUILD A 3 LEAF WALL!

triple leaf data

One very effective partition design for musical isolation is a concrete block wall with the open cells of the block filled with grout to yoield a solid block of materal – very heavy = high TL [transmission loss].  Improve such a wall by painting or plastering both sides to fill the pores in faces of the block.

A number of improved noise isolation framed wall designs incorporate either two separate frames, or single top and bottom plate with double studs (one for each side of the wall).

Another very practical design element is RC (“resilient channel”), which is a folded metal strip which is designed to be run perpendicular to the studs of a wall and then have drywall panels attached.  RC adds a mechanical decoupling element to even single studded wall frames by acting as an adjunct to the springiness of the framing members and the air cavity between the panels which comprise the two leaves of the wall. One edge of the RC is mounted to the framing, and then the drywall is attached to the other flange of the RC strip. Caution though – adding RC improperly and you can drive screws through the strip and into the underlying framing - which can diminish its usefulness. In order to maximize the benefits of RC use the largest mass load on the RC and largest RC member spacing consistent with safety and practicality (generally this means install the RC 24” o.c. with at least two layers of 5/8ths drywall on top). If RC is dropped altogether, consider what means may be required to maintain mid frequency performance in your wall system (add a layer of drywall!). Mechanical isolators like the RSIC clip are another type of device used for this same general purpose of decoupling the two sides of a wall.

When choosing insulation material for a wall cavity, denser is better (to a degree) and full is good (don’t overfill so as to press on the back of the wall panels), but this is a diminishing return and spending a lot of money on filler is probably a lesser return than adding mass to the wall itself, or enhancing decoupling. 703 is a very good filling for wall cavities – generally the higher the thermal rating for insulation (R value) the better you might expect it to work as acoustic cavity filler, but standard batts of roll insulation products are a good balance of cost versus performance and spending money on anything more "sophisticated" is probably a waste.

In the commonly used two leaf panelized isolation wall systems (a basic drywall–stud-drywall design for instance) with an insulated cavity, isolation occurs across a mass - spring - mass system. The framing and the air in the insulated cavity is the spring and the drywall panels on each side of the studs are the mass components. To achieve very good low frequency isolation you need massive panels and a dampened spring (or a spring with a resonance frequency far below the sound spectrum of concern).   Walls designs with wider cavities work better than "skinny" walls... the air component of the "spring" in a wall system with a wide cavity is "looser" - so that it vibrates at a lower frequency.

Adding a damping/constraining layer to a drywall sandwich on one side of a wall can help tame resonances - this concept is based on the idea of building a visco-elastic [bouncy but squishy - never dries out and gets hard] layer between the panels on one side of a wall - that constrains the vibrations.  Green Glue is a product which is specifically designed to achieve this goal and is a known good approach with proven economical benefits when used on massive drywall constructions looking to achieve large amount of noise transmission loss [thick multi layer drywall partitions show a good bang for the buck ratio when a layer of drywall is replaced with a Green Glue layer].  Though even something as simple as a 1/4" layer of asphaltic roofing material can provide similar benefits [same idea - just not a scientifically optimized product].

Keep in mind that for a partition wall with a door or a window - it is pointless to build a wall which greatly exceeds the isolative characteristics of the window / door.  Focus on getting the window door isolation as high as possible, then build a wall able to maintain that level. Also you must recognize the existence of flanking paths such as HVAC systems and ceilings.

Bottom line?

Mass is the King.  For all practical framed wall designs you will rarely go wrong foregoing other techniques for another layer of drywall... or other means of adding mass.  For most applications drywall [gypsum board] is the cheapest easiest way to add mass... but if your project allows, its hard to beat the brute force performance of a properly built [airtight] masonry wall [bricks and mortar are heavy and heavy things hate to wiggle].

Look here for more detail:

The effect on TL by adding additional layers (Graphs)

Once you have settled on a wall design with ample mass for your application take advantage of opportunities to dampen any wall cavities with inexpensive insulation to improve low frequency performance. To achieve the highest level of performance you will also have to consider ways to mechanically decouple and/or otherwise diminish the ability of opposite sides of the partition to conduct vibration one to the other - in this vein there are several specialty materials which can assist. Items like the RSIC clip and materials like GreenGlue are known to be effective if and when the project calls for achieving the highest possible isolation from the partition.

More on these issues and some good graphics are located here:

Here is an informative post by Brian Dayton of The Green Glue Company in which he shares general wisdom derived from his experiences conducting extensive scientifically valid testing of various wall designs.

the worlds cheapest sound isolation tips

which lays out these rules:

1.  use wider stud spacing as code allows.  Its generally always better to space studs 24"OC than 16"OC.  at the files IR818 and IR761 and some in IR693 have more than enough data to cement this in as a firm rule.  However, be wary of using spacing wider than the local authorities would allow.

2.  use normal boring, off-the-shelf, home depot low-density fiberglass insulation.    By my interpretation, the data from IR761 suggests that this material is preferable to higher density more expensive types.  It does not show higher tranmission loss at middle and high frequencies (indeed, it shows lower TL in those bands, generally), but it does seem to have a BETTER effect at lowering the frequency fo the mass-spring resonance.  

3.  don't use exotic esoteric, really expensive sealants.  Data presented by USG shows that when substituting their own sealant for a putty like sealant (heavier, higher damping), the only real penalty was at the coincidence dip, and the overall utility of the wall didn't decline.  If you wish to utilize a really flexible type of sealant, try Tremco Acoustic sealant.  This is a butyl, solvent-based material with extremely good properties.  But generally boring old latex type caulks are fine.  Believe it or not, alot of off-the-shelf latex caulks are more flexible than some marketed acoustic sealants.  

But do use alot of cualk.  Never rely on a single layer of caulk to seal the wall.  multiple layers (under top/bottom plates, each layer of drywall) back each other up in the instance that one isn't perfect, and this is worth alot.

4.  Don't feel compelled at all to float your floor over a concrete slab if you are on a budget.  I know that statement may spark alot of controversy, and someone might call me names, and without a doubt concrete slabs can be sources of flanking noise.  But the weight of the slab ensures that a very good level of performance can be attained before flanking -via- a slab becomes the problem.

Also, you risk making things worse when floating a floor.  It is no small challenge to get the MSM low enough in frequency to ensure that your efforts are positive, and the expense and/or labor won't ever be small to truly successfully float a floor.  

I know slabs can be problems, but for AIRBORNE (as opposed to impact sounds, where they are a real problem) noise, they aren't going to cause your whole project to be bad unless they are really cracked or something along those lines, allowing an air path under the walls.  

5.  Cut things.  When possible, just cutting structural connections can be a great help with flanking noise (mostly via floors).  Don't cut anything that would cause danger to yourself via water/mold or structural weakening, and always ask the local building inspector if at all in doubt.

6.   Cover your entire door with sound absorbing material.  Some foam or 703 or whatever.  Leave a space around the door knob if you need to, but cover the entire door.  (don't bother with this on walls)

7.  if your wall is decoupled (staggered studs, resilient channel, double studs, sound clips, etc.) then your wall will benefit from a deeper air cavity (such as using 2x6 studs -vs- 2x4 studs).  But if your wall is not decoupled (single wood stud or single rigid steel stud, and to an extent single flexible steel stud), depth is just wasted space and $$$.  IR818 from the NRC has sufficient data to justify this statement.

8.  $$$/lb.  Thats the most important factor when selecting what type of panel to use.  Plywood?  Drywall?  Cement board?  The best choice is basically always drywall because its least expensive per pound, and drywall is basically always the best choice or basically all walls or ceilings.  don't make a drywall floor or door.  

There is no magical nothing to wood or cement or whatever relative to gypsum.  None at all.

Is 5/8" plus 1/2" best?  is 3x1/2"?  whichever is cheapest in $$/lb

9.  don't risk weird decoupling schemes that aren't tested.  Things like little homemade brackets here or strips of foam there may - in at least some cases in my tseting experience - work against you as much as they help.  I know its easy to get caught up with the details of "caulk-vs-mudding my inner seams" or "do i need some foam here", but its really simpler than all that most of the time.

10.   When selecting a method of decuopling, alot of the work we have done has shed some light on how the different methods behave, and i summarize here.  This gets a bit long.

The data from IR761 would seem to show that thin-gauge (25 gauge) steel studs and resilient channel over wood or rigid steel studs perform largely the same.  And when looking at STC ratings, most of the historical data tends to favor either of those methods over staggered studs.  

However, the situation is this when scrutinized over a wider range of wall weights and frequencies:

-the properties of resilient channel in the market today are all over the place.  All over the place.  Widely different designs and even different gauges of metal are found, and the original USG channel appears to remain the best.  At minimum, you should ensure that your RC is 25 gauge.  You can do this with a ruler and some reference samples of steel from somewhere - just make sure the thickness is the same.  Or buy a micrometer somewhere.

all resilient, however, has an adverse effect on the location of the MSM when utilized because it adds a spring in parallel to the spring of the air cavity, raising MSM to a frequency above that predicted by a mass-air-mass calculation.

-steel stud walls, even thin gauge ones.  DO NOT CREATE a true mass-spring system.  some of the behavior of a wood stud system remains, and you cannot drive the MSM down and create a very low frequency decoupling point with any practical level of mass.  Both panel stiffness and a mass-air-mass factor into the location of the MSM in this system, and as weight goes up the stiffness factor becomes dominat and MSM stubbornly refuses to go down in frequency.  The single thin-gauge steel stud wall is an interesting thing.

-staggered stud walls DO create a true MSM with MSM falling quite close to the theoretical mass-air-mass resonance, BUT, they also suffer from mechanical resonance issues at slightly higher low frequencies that keep performance less htan ideal.

-double studs are just the best choice

But, the combination of these is more than any one together, as it were.  The use of thin gauge steel studs + resilient channel yields two springs in series (the sum of those being in parallel to the air spring), and allows you to create a true MSM that you can drive down in frequency with mass, and which will occur at a lower frequency than that of an RC wall with rigid studs.  Something else to think about is that if you short-circuit your RC on steel studs, the steel studs are still flexible and the results aren't likely to be as disastrous as they would be with wood studs.

So, considering the whole frequency range, the preferred decoupling methods are

1.  double studs
2.  sound clips (really, their advantages exceed the rest, but cost more)   OR thin steel studs plus RC
3.  stagg studs  (if using Green Glue these are as good as the sound clips)
4.  RC
5.  thin steel studs

That should be a good decision making guide.  Once decoupled, adding mass to a wall does MORE than it does on a non-decoupled wall as it both adds mass and lowers MSM.

11. the use of 1x4 wood furring strips or simple metal furring strips (even if not resilient) does help your cause.  They should be spaced 24" OC.  They sort of make a 16" OC wall into a 24" OC wall

12.  send me some sludgy authentic european beer.  I heard that makes your wall 5 dB better for every 6 pack you send.
Content posted by me is copyright 2004, 2005, 2006  Brian Ravnaas, but may be reproduced without permission for any non-commercial purpose so long as the intent is preserved.  NRC Canada data is copyright them and used with permission,
Scott R. Foster
Posts: 3854
Joined: Mon Feb 02, 2004 12:41 pm
Location: Jacksonville, FL USA


Who is online

Users browsing this forum: No registered users and 1 guest