knightfly wrote:Carl, believe me I feel your pain, and have felt it for the last 20 years as I've studied everything I could find after building a room while thinking I knew what to do - that room is inside a barn about 100 feet from my house, and I was amazed when my wife could tell from inside the house which song I was playing on the piano in the "studio" -
I've been trying to find the time between answering posts and the interminable daily crap we all put up with to do just what you ask - so far, I've got a long ways still to go.
In looking back, there are a few posts where I've delineated quite a few of the "rules" of sound proofing as I believe them to be - Here is a rough gathering of those comments, totally un-edited as yet. If there are any conflicting comments they will be gone when I'm finished with the FAQ - meantime, as usual I'm further behind than I'd like to be so this is all there is for now -
First of all, here are some definitions, then some rules of wall construction as I see them, with reasons if I'm lucky enough to remember them - afterward, I'll tell you what I think you should do where, and why -
The most effectuve use of materials can be had using a mass-spring-mass type of wall. The mass is whatever layers of paneling are attached to either side of the frame.
A leaf is all the layers of mass on ONE side of the frame, whether it's one layer or four.
An Air space is any non-solid, including insulation. The wider the studs in the frame, the deeper the air space.
Air is considered a "spring", because it is relatively stiff and will transmit sound from one leaf to the other.
1. Never put RC(Resilient Channel) on both leaves of a wall - it allows too much movement on the part of one of the leaves, decreasing TL through the wall.
2. Two different thicknesses of wallboard beats two of the same size for mid/upper frequencies, two thicker panels does better at low frequencies.
3. Low frequency attenuation is improved by NOT having RC, by increasing mass, increasing air gap between leaves, and increasing stiffness.
4. Fastening a second layer of wallboard to a first layer, when the first is mounted on RC or STEEL studs, should be done with screws through the first layer and into the RC or (steel)stud ONLY - if these screws touch WOOD studs the isolation of the wall will be reduced by as much as 9-10 dB. The STEEL studs are flexible enough not to cause a loss in performance.
I'm updating the following paragraph with the one marked 5A, below it - newer more complete information and practicality are the main reasons - I've left the old one in, so differences are obvious.
5. Fastening a second layer of wallboard to a first layer, when the first is mounted on WOOD studs, should only be done using 4" wide stripes of slightly thinned joint compound over the stud area, then fastening with Type G laminatings screws approximately 2-3" away from the studs and parallel to them, on approximately 12" centers. This allows the two layers to act as separate panels, which maintains each panel's coincidence dip separate from the other. It also keeps any fasteners from penetrating BOTH layers of wallboard AND entering the WOOD studs. NOT doing this will lose about 9-10 dB of isolation due to flanking through the fasteners and into the studs.
5A - replaces 5, above -
Fastening a second layer of wallboard to a first layer, when the first layer is mounted on WOOD studs: First, if building a SINGLE frame with wallboard on both sides, it's STRONGLY recommended to use Resilient Channel on ONE of the sides of the frame – otherwise, you lose several dB of isolation at mid and higher frequencies (above 100 hZ typically) – follow the rules for mounting wallboard to RC on this side – no screws in line with studs so you don't accidentally "short out" the RC.
Fastening a second layer when there is NOT RC under, the most practical way is to mount the first layer using half as many screws as normal, same with any following layers EXCEPT the outer layer – this one should have full fastening schedule used, and with long enough screws to penetrate all layers and solidly fasten to the studs.
In the case of double frame walls with wallboard only on the outside of each frame, use half schedule fastening on all but final layers on each side.
Full schedule screws for wallboard, according to USG – 8" on center around the edges of panels – 12" on center in the field (middle of panels) – half-schedule screws would be twice as far apart for all but the final layer.
In all cases, ¼" spacers should be used under panels until fastened, then pull the spacers and caulk before repeating for the next layer.
For a standard single stud frame with no RC using 2x4 wood studs, two layers of 5/8 sheet rock on each side will give an STC rating of around 48 dB – the same wall using ½" will also rate about 48 dB, but will actually perform worse in lower frequencies by 2-3 dB because of the extra mass of the 5/8" panels.
A double framed wall with 2 layers of 5/8 sheet rock on only the outer sides of each frame, with a total air/insulation gap of 8" (1" between frames) will test out at about STC 66 – the same wall using ½" sheet rock will be about 2 dB worse STC (64) and 3 dB worse at the low end as compared to the wall using 5/8 panels.
6. Insulation should be placed against inside wall panels to increase damping of the panel - This helps TL at low mid/high bass regions, as well as decreasing the effect of the Coincidence Dip which occurs in drywall at between 2 kHz and 4 kHz. Optimum density for in-wall, rigid insulation is around 2.5 to 3 PCF.
7. RC should never extend far enough for its end to contact the intersecting wall - cutting it about 1" short works well. Wallboards should also not directly touch the intersecting wall, but should have a gap of approximately 1/8" between them. This gap must be caulked with enough acoustic sealant to ensure complete seal.
I don't think I would personally feel too comfortable in a room with 3 layers of drywall overhead mounted on ANYBODY's RC - I would personally tend more toward the heavier leaf of a 2-leaf wall or ceiling being mounted solid to the studs, and just putting two layers of 5/8 firecode sheet rock on the RC. It kind of depends on how much isolation you really need, and whether you can MATCH it with your glass, doors, HVAC, flanking noise, etc - if not, it's a moot point at best.
In any sound barrier, the primary task is getting two leaves of mass separated by one air space, with ZERO hard coupling between the two leaves. It helps if each leaf has different characteristics, since any frequency that excites BOTH leaves will pass thru with the least attenuation.
As always with RC, if you goof EVEN ONCE and put a screw thru the wallboard and into a wood stud, you will lower isolation by enough to almost totally NEGATE the use of RC at all. I've read of walls being tested and dis-assembled, only to find that as many as TWENTY screws were "shorting out" the RC - a good argument for marking stud/joist locations before putting up wallboard.
For insulation, knowing what I know now I'd use 2-3" 703 in between the frames, and fiberglas batts (possibly the newer "high density") slightly compressed against the panels, between the studs just like standard house construction, but with some compression (use next size thicker batts than stud cavity depth) -
Alternately, if you can't find or afford 703 the standard house fiberglas batts are almost indistinguishable inside a wall, and may actually improve low frequency performance slightly.
With ANY sound wall construction, you want ZERO penetrations if possible, and definitely NO BACK TO BACK ones. Best electrical is surface mount, second best is offset boxes (by at least one cavity, preferably two) and total caulking of boxes.
If using surface mount wiremold, I'd put it just below any wall treatment so it almost disappears. I don't recall what the minimum height for outlets is to meet code, but you should conform to that whatever it is.
On the subject of glass between rooms, John will probably dis-agree strongly with this, but from 3-4 different sources I've seen, maintaining a wide air gap is at least as important as minimising flutter. If you could widen the gap between your walls at the window area, you could do both, but I wouldn't set up less than a minimum air gap in the window I described above that was less than 8 inches. If you figure that a slant of 1 inch in 10 will get you a 6 degree splay, that means that the width of a 2x4 (3.5") would be just right for a 36" glass. If you maintain a minimum 8" distance between glasses at the narrowest point, your wall (including wallboard and channel) would be a total of about 18" thick. (8" space, two 3.5" frames, 2.5" total wallboard, + at least 1/2" channel = 18" - more if you were to use the RSIC-1 clips and "hat" channel instead of RC.
First, for some background (lazy typist here) go to this thread
http://www.johnlsayers.com/phpBB2/viewt ... 86&start=0
and scroll down to the STC chart John posted on Mar 01 - look at the various wall constructions closely, then check the comments I made on Mar 17 as to WHY this is so.
Here are some general guidelines for soundproofing -
1. NEVER waste time/money/materials on construction methods that aren't proven to work, unless you have no life and just like to hurt yourself...
2. The best performance in a sound wall for the money, unless you can afford 3 foot thick concrete, will ALWAYS be a double-leaf wall with only ONE air space. Adding extra walls will usually make things WORSE, not better. (A "leaf" is however many layers of solid material that are layered together on one side of the frame. A standard, interior house partition with drywall on each side of 2x4's is a "double-leaf" wall.
3. This double leaf wall should almost NEVER have exactly the same construction on both sides - a wall transmits sound (bad) more efficiently at its RESONANT frequency. If both leaves of your wall are identical, then so will their resonant frequencies be identical. If that frequency is, say, 44 hZ, then that frequency will pass right through the wall with MUCH less attenuation than other frequencies. Not good.
4. For best performance, there should be no point where the inner and outer leaves of a wall have direct, hard contact with each other. This will allow sounds that impinge on the one leaf to directly vibrate the second leaf, and pfpfpfttt!!! Right on through... The way around this is separate frames, whether wood or steel studs. Drywall on the inside of the inner frame, and on the outside of the outer frame.
5. Things that improve Transmission Loss are: Mass of leaves (more is better, but DIFFERENT is better yet) - Distance between leaves (more is better - at normal wall thicknesses, even an extra inch can help) - Insulation - so far, it appears that 2.5 to 3 Pounds per Cubic Foot is the best overall density for in-wall insulation. This is why USG settled on 2.5 PCF for their Sound Attenuating Fire Blankets - see this link
http://www.usg.com/Design_Solutions/2_3 ... onperf.asp
That’s it for now, I hope this helps condense things a bit… Steve