My First Sub

[---k---]

Thomas,
I've been trying to read up on Bucking Magnets. A lot of differnt opinons out there. Do you beleive that glueing them to the back of the driver affects the drivers T/S parameters? I'm guessing not if you and Tibor used them. I asked Mike at SoundSplinter. He said he had never tried, and that the his ocncerns would be needed a really big bucking magnet and that the beveled backside of the driver would make adhearing it difficult.

Second, do they have to be glued to the back of the driver? I saw your post where you just put them on the shelf above the TV. This would seem to indicate that I could glue them to the box walls.

I moved the driver a bit around more today, and found when I rotated into the verticle position, I couldn't get anywhere near as close as horizontal. wierd. I will definitely need to do some shielding.

Any thoughts/tips?

--now off to sholve the snow. grrrr

[ThomasW]

Bucking magnets you'll use will have minimal effect on T/S parameters.

There's no such thing as a big bucking magnet. The largest I know of is 20oz. Several can be ganged together to make a larger one. (you might start by getting 4 total from PE)

Since they're small, and even though they're put on in opposition to the driver's magnet, they'll will be held in place by the stronger field of the driver's magnet. I use clear hot glue to make sure they don't fall off. This means they can easily be removed.

The Arvo isn't an example for you. They have no boxes and therefore can't be shielded with steel. The stack was placed on the mantle prior to placing them on the drivers. When I noticed the effect I just left them there. Sometimes it's better to be lucky than smart...

[---k---]

Bucking magnets are cheep. I ordered 8-15oz from PE. Should be here soon.

Thomas,
I know that you recommend two layers of 3/4" MDF. I was thinking about using 1-3/4" and 1-1/2" with the bracing being 1/2". I figure that I can save almost 30 pounds per sub this way. How bad of an idea is this? Can you rate it on a scale of 1-10?

[ThomasW]

That will be fine since you're using 2 layers of steel

[---k---]

I got the Bucking Magnets today.

First, I'm a little confused. When I would move the bucking magnet towards the rear of the driver, I would get resistance, but then it would grab and attach itself to the driver. If I flipped the driver over, it was very strongly attracted. This seemed odd.

I did a little testing. I paused the TV on a white screen and moved the RL-p15 towards the screen. I could tell some effect of 2 magnets, 4 about did the trick completely. Do you think hanging 4 pounds of bucking magnet off the back is too much?

Also, the driver has a rubber boot surrounding the driver magnet. I'm thinking about hot glueing some 16ga steel to this rubber boot for sheilding. Does this sound like a good idea? Maybe do this insted of in the box sides?

Thoughts?

[ThomasW]

Ryan,

I explained above about the magnets reversing, this is due to the much greater strength of the magnet on the driver, compared to that of the bucking magnets.

I don't think it's a good idea to use that many magnets, they could change the T/S parameter of the driver. That's why I suggested you to buy only 4, using 2/driver.

Trying to create a steel "can" around the magnet isn't all that effective (yes I've tried). If you want only use one layer of steel build the box and just line the inside.

[---k---]

Thanks Thomas,
I don't think what you said ealier sank in, because I didn't have the magnets and couldn't see what you were describing. I re-read it, and now understand.

About buying 8, I thought you said I could "start with 4"
I got 8 only because they were cheap, and I didn't want to have to re-order. If you think that 2 is the max w/o changing he T/S parameters, that is what I'll use.

If I line the inside of the box with steel, would it be a good idea then to cover it with egg create or something like that in addition to stuffing the box with fibreglass?

[ThomasW]

Regardless of whether there's steel inside, the box should have damping material.

[ThomasW]

All the latest off-topic have been removed. Lets say on topic and help Ryan complete his project.

[---k---]

Thomas,

Thanks for clearing the clutter.

My father in-law will be in town this weekend with his truck, so I intend to make a trip to Lowes to buy a few sheets of MDF and begin building.

Got a question, I read over at AVS that a subs magnetic field will be greater when the sub is in use. Is this true? I ask because I was moving the driver all over the place the other day, and was almost to the point of not bothering with the shielding. Was this a waste of time?

[cjd]

Try it and see. You will have the addition of the magnetic field created by the voice coil which acts against the static field of the magnet... Not sure if it'll change how well it works (or doesn't work).

You know, it's the perfect excuse to get a front projector... nice biiiig screen, and no monster glass box any more! Just, um... monster sub boxes.

C

[ThomasW]

Got a question, I read over at AVS that a subs magnetic field will be greater when the sub is in use. Is this true?

I've not heard of this, but I think it's better to error on the safe side. Use 1 or 2 bucking magnets and a single layer of steel. That way you'll have some margin to play with regarding placement and there won't be negative impact of the perfromance of the driver.

You can always add or subtract the magnets if they aren't needed. But it would be a drag if you left out the steel only to find out it was necessary.

[---k---]

Thomas,
(or anyone else)

I saw a post the other day on DIYAudio.com where someone had performed a finite element analysis on a sub box to determine the amount of panel flexing and the required bracing. I was fascinated by this. Being a structural engineer and having a couple of FEA programs sitting on my desk, I thought this would be a cool thing, and a way to get the weight down. I must say, I was skeptical at first at the amount of bracing and thickness you guys typically recommend. But after running my first model, a simple 2D analysis of a 20"x24"x3/4" MDD w/o bracing, I got a deflection of 0.379". WOW! This was using 21psi pressure on the plate, which is what I calculated the SS RL-p15 will create in a 90L enclosure.

So I'm wondering, what do you think would be the maximum deflection I would want? You got any feelings on whether 0.1" is okay?

I'll give you the results from some full 3D models after I finish them.

Oh, and I know that this is just theoretical, because the real world joints may not be as stiff as my model, but this is more educational than anything.

Thanks.

[Scott Simonian]

a simple 2D analysis of a 20"x24"x3/4" MDD w/o bracing, I got a deflection of 0.379". WOW! This was using 21psi pressure on the plate

Wow! I had no idea. This kind of information sure is cool to know. BTW, how are FEA measurments done?

[ThomasW]

This is a hot button topic since some computer analysis studies (Noah Katz?) show no problems with single layer 3/4" material (note MDF isn't great, the best results are with Baltic Birch ply). But Dan Wiggins posted (I can't remember where) about what a small amount of deflection did in terms of cancellation.

My motto is overbuild. Doubled 3/4" MDF is 8 times more resistant to flexing then a single 3/4" layer.

Any flex, (vibration) means the walls are acting like a speaker. So doubled 3/4" MDF or a single 3/4" BB ply with "window" bracing... this is window bracing with 1" MDF (not readily to the home builder)



The only issue with this construction method is that it's very difficult to line the box with steel.

Also if the box is lightweight (comparatively speaking), it will "walk" with high output levels.

[Dennis H]

This was using 21psi pressure on the plate, which is what I calculated the SS RL-p15 will create in a 90L enclosure.

Not quite. Atmospheric pressure is about 14.5 psi. That's the static pressure inside and outside the box. The SS driver displaces +/- 2L from it's rest position. So the change in pressure is +/- (2/90)*14.5 = +/- 0.32 psi.

[cjd]

For the Tumult I did for my dad, I used this brace:

Image not available

and double wall 3/4" birch ply.





Spiked, ~21x21x26" internal dimensions.

It doesn't flex, it doesn't walk, it just cranks out the bass.

I had to deliver it with out the driver installed.

C

[---k---]

Not quite. Atmospheric pressure is about 14.5 psi. That's the static pressure inside and outside the box. The SS driver displaces +/- 2L from it's rest position. So the change in pressure is +/- (2/90)*14.5 = +/- 0.32 psi.

Dennis,
I read that last night, slept on it, and still haven't figured out if you are correct.

When I calculated 21psi originally, it didn't pass the gut check, but the results were where I thought they should be. I took 90L and assumed that the pressure would be increased 92/90= 1.02, so the pressure would be 14.5*1.02*1.4(adiabatic heat of compression).

Just FYI, my models w/ 21psi showed:
20x24x3/4, no bracing - 0.379"
20x24x1", no bracing - 0.162"
20x24x1 1/4", no bracing - 0.84"
20x24x3/4, with two braces - 0.011"
20x24x1 1/4", with two braces - 0.003"

What you are saying makes sense. Looking at my calc, I think I failed to consider the pressure on the outside of the box. But the results from 0.32psi don't make sense. Applying 0.32 pressure would just linearly change the results above. So 0.379" would change to like 0.000577". That is way small.

I think I'm going to have to break down and look for my old fluids or thermal book to figure this out. Argggghhh.

[cbj]

While you guys are on the subject of bracing I'll go ahead and ask the noob question (and hope I don't tread on --k--'s thread too much). Is there a prefered type of bracing and why? I saw Thomas post pictures of window bracing and the other type of bracing I have seen a lot of is the type in the Adire plans where a bunch of circular holes are cut out. I realize they are trying to accomplish the same thing, but does either have an advantage in functionality or ease of construction?

[---k---]

This is a hot button topic since some computer analysis studies (Noah Katz?) show no problems with single layer 3/4" material (note MDF isn't great, the best results are with Baltic Birch ply). But Dan Wiggins posted (I can't remember where) about what a small amount of deflection did in terms of cancellation.

Thomas,
Yeah it was Noah. I was hesitant to bring this up, and more so to mention his name, because I knew this was a hot button topic, and I didn't want to rock the boat until I got my sub built.

But, as I did more models, got out my crack gage to remind myself how big 0.05" really is, and read a review of an SVS PB10, where the guy added clamps to brace it and improved the sound quality and decay, I had about convinced myself that you were correct. I was hoping I could bait you into giving my a deflection amount though.

I would have never dreamed that I would need that much bracing and wall thickness. I guess I still have a lot to learn from you guys who've been around a while. ;x(

Ofcourse, this assumes that my 21psi is correct. If Dennis is right, I may go back to calling you crazy and only add weight to keep it from walking, and vibrating the piss out of the speaker on top.

[ThomasW]

I realize they are trying to accomplish the same thing, but does either have an advantage in functionality or ease of construction?

I've never built a sub using the brace design like Adire or Chris' in the pic above.
I used a form of window bracing when I built the AS-15. There I made the outer layer and the braces were spaced by gluing in separate pieces of wood that also formed a complete second layer for the cabinet. That way I didn't have to deal with lining up dado cuts made in 1.5" thick material.

The AS-15 design was an extreme example of a no-holds-barred approach to building a sub. And as a result it weighs in ~300lbs.

Ryan,

You don't need an extreme from of bracing. If you go with doubled MDF and add a few simple side-to-side braces that will be plenty and make it much easier to place the steel then put in the bracing.

Noah has as far as I know has never built a sub. He spends quite a bit of time talking design/ theory and running sims, but that's not a substitute for hands on experience. Since math/physics isn't my strong suit (I'm industrially dyslexic), so I avoid interacting with him. But I've been building speakers for 40+yrs, so I have a reasonable idea what works and what doesn't.....

[Paul H]

Ryan,

Re-run your numbers, keeping in mind that the pressure outside the box is not zero - so you're calculating pressure differential between inside and outside - absolute pressure is of no consequence for these deflection calculations.

Paul


edit: typos

[Dennis H]

Ryan, I don't think anybody is crazy here.... well except maybe me.

Let's run the numbers and see if we can figure how much SPL the box is putting out.

.32 psi --> .0058" of flex (check your decimal)
Convert to mm --> .15 mm
5 panels moving .15 mm is like one panel moving .75 mm
The panel is flexing so let's say the "average" area moving is like a 15" driver.
The driver is moving 24 mm and the net panel movement is like another driver moving .75 mm.
.75/24 = 3% "distortion" if I can use the term loosely.
Relative SPL of the panels is 20*log(.75/24) = -30dB below what the cone is doing.

Now keep in mind that the panel is a flexing resonant structure so it will be making noises at different frequencies than what the cone is doing. There's a pretty good chance that the 3% "distortion" is audible. Maybe not terrible but audible to a trained ear.

Edit: oops 5 panels vibrating, not 7.

[ThomasW]

Ryan,

Another bracing option is a matrix. Here's a thread with some pics of that. The ply sections aren't glued in and he's yet to drill the holes between the sections.

https://www.htguide.com/forum/showthread.php?p=216351#post216351

To use this technique in your steel lined sub, you'd use liquid nails to attach the matrix to the cabinet shell

[---k---]

Ryan, I don't think anybody is crazy here.... well except maybe me.

Let's run the numbers and see if we can figure how much SPL the box is putting out.

.32 psi --> .0058" of flex (check your decimal)
Convert to mm --> .15 mm
5 panels moving .15 mm is like one panel moving .75 mm
The panel is flexing so let's say the "average" area moving is like a 15" driver.
The driver is moving 24 mm and the net panel movement is like another driver moving .75 mm.
.75/24 = 3% "distortion" if I can use the term loosely.
Relative SPL of the panels is 20*log(.75/24) = -30dB below what the cone is doing.

Now keep in mind that the panel is a flexing resonant structure so it will be making noises at different frequencies than what the cone is doing. There's a pretty good chance that the 3% "distortion" is audible. Maybe not terrible but audible to a trained ear.

Edit: oops 5 panels vibrating, not 7.

Dennis,
Very cool. Thanks for putting that into some real numbers.

Thomas,
You got to understand, so of these questions are just for my big egg head. I like to understand the mechanics of things. I'm not quite there yet, and I know it, so bear with me as I experiment and ask questions.
I was planning on doing sometype of window bracing. Probably one in the vertical direction and two in the horizontal above and below the driver - if I put the driver in the middle. I'm thinking maybe putting it in the bottom would be better.

[ThomasW]

You got to understand, so of these questions are just for my big egg head. I like to understand the mechanics of things. I'm not quite there yet, and I know it, so bear with me as I experiment and ask questions.

No problem, just let Dennis do the math instead of me.

When I need serious number crunching I hassle Jonmarsh to do it. That way I know there are no mistakes, or if there are, who's to blame ...

[noah katz]

Couple of things:

--k--, there are a couple of errors in your calculations.

It's the pressure change that flexes the walls not the absolute pressure; and the 1.4 is applied to the change, not the absolute.

Dennis,

".75/24 = 3% "distortion" if I can use the term loosely.
Relative SPL of the panels is 20*log(.75/24) = -30dB below what the cone is doing."

Distortion is not based on displacement, but SPL (right?), which you calculated as -30 dB.

So distortion is 0.1%.

Also, as I've said elsewhere, IMO it's not distortion.

If you assume the first panel reesonance is at least an octave above any significant excitation, which will depend on XO freq and slope, the panels are simply moving out of phase with the driver.

So the result is not distortion, but a miniscule lowering of level.

For the example here of -30 dB, that gives an output reduction of 0.004 dB.

Another thing, the panel deflections mentioned were max at the center of the unsupported area; it's a bell-shaped deflection curve which has a much lower average deflection over the whole panel area.

Regarding placing a speaker on top, if the points of contact are near the corners of the sub box, there will be essentially no vibration trasmitted.

[cbj]

Thomas thanks for the clarification on the bracing. I learn quite a bit reading threads like this one and appreciate the time and effort you and others put into this site. Back to reading mode . . .

[---k---]

Noah,

Wow, okay the math is starting to get out of control.

I'm not sure which of you is correct. I think I've gotten both of your opinions, and will keep exploring it. I don't think that there is much use in turning this into much more, because I don't think either one of you is going to change your minds. I may send you a pm, and get more of your thoughts, because I think there is a lot of merit in running the FEA and determining how much flex, ect. is going on.

If it could be kept civil, and some of the noise out, this actually may an interesting discussion to have over AVS and see if we can get Mark Seaton's and Tom Danley's response. I'm more than willing to provide the FEA models, if someone will give me the correct pressure to apply.

[noah katz]

I was planning on a thread at AVS when I got motivated enough to get my results together.

Even though I'm pretty sure these results show that people have been overbuilding sub boxes, doing it this way will have the unaesthetic result of very noticeable (by touch) panel motion.

We are extremely sensitive to vibration, and can easily feel motion that we cannot see.

[---k---]

Here is that SVS PB10 review:

http://www.soundenvironments.org/modules.php?name=Downloads&d_op=getit&lid=8

If you google PB10_Review.pdf, you will get links to all the forums that discuss it. Some interesting threads. I'm making my way through them. The Home Theater Forum has responses from Tom V. The AVS one is interesting as it was only a year ago and yet all the names are differnt.

But, lets not get sidetracked. Looking forward to the discussion on AVS about this.

[ThomasW]

Noah.

We look forward to your thread on AVS

Ryan,

Suffice to say you won't make a mistake by using thicker walls, bracing, and damping, which is what all the high-end speaker mfgr's use....

All REL designs use 30mm MDF

The Revel Performa B15 cabinet is constructed of 1-inch (25mm) thick MDF wall, optimally designed and internally braced with the aid of laser interferometry.

Vandersteen, the window bracing is cutaway in this pic



Aerial Acoustics



Wilson Audio note window bracing in isometric view for the XS

Image not available

[Dennis H]

Noah,

Actually, I calculated the "% distortion" first and converted to dB SPL. For SPL, like voltage, the formula is 20*log(P1/P2) so 3% displacement (pressure) is -30dB. I realize that none of these numbers are exact and "distortion" is used a bit loosely but it's just a crude whack at getting a ballpark feel for the SPL of the cabinet vibrations. The bottom line is that the radiating surface of the cabinet is so big that it doesn't take much displacement to get audible vibrations.

Certainly both views are valid and have their place -- those that feel that the vibrations are not that big a deal and those looking to remove even small problems with the sound. In the real world of listening experience, I think a 2' cube of 3/4" MDF is getting to the point where most experienced builders would opt for bracing and/or a second layer.

Edit: here's a pic from Siegfried Linkwitz showing % distortion and SPL. When in doubt, ask SL.

[noah katz]

Thomas,

That's all well and good, but they may be doing that for purely marketing and the aesthetic reasons I mentioned, as opposed to performance.

I admit it's a tough row to hoe for mfgr's, the salesman has only to say "Feel how much less this one vibrates, it's way better."

I'm only trying to help out DIY'ers who might want not want waste time, money, and effort.

Another thing. The 15 cu ft box I'm just finished has about 48 sq ft of surface area.

If I had doubled up to 1 1/2" thick, the internal volume would have been reduced by 3 cu ft.

IIRC WinISD says that would give about 1 dB less output for the power available, far overshadowing the insignificant output reduction from panel motion.

[Dennis H]

If I had doubled up to 1 1/2" thick, the internal volume would have been reduced by 3 cu ft.

But you could brace it with very little loss of volume.

[ThomasW]

Noah,

We look forward to you building 2 subs. One using your concepts, and another using generally accepted principles.

The data you publish from the both the listening and scientific testing will be an interesting read.

Now let's return this thread to the original topic, helping Ryan design and build his sub.

[noah katz]

"But you could brace it with very little loss of volume."

Just what I did.

To first order putting a crossbrace in the middle of a panel reduces bending deflections by 8X.

Sorry Thomas, I will now cease and desist, unless somebody else doesn't

[---k---]

I'm going to cease!

Well, my father-in-law and I went to Lowes and got some 3/4" MDF (they don't have 1/2"), and then to HD to get some 1/2". I must say I was a doubter when people were posting on AVS that HD's MDF is lesser quality, but it is. I should have just gone back to Lowes and got more 3/4", but I was there at that point. Whatever.

We got it home, and they left to return to MI. My wife was a trooper and helped me make about 80% of the cuts. The garage isn't heated. I had a space heater running before we started, but I only have one outlet in the garage and couldn't run both at the same time. grrrr. (summer time project -- rewire the garage and insulate) When we finished, we could see our breath!

I learned that even with a peice of 1x4 clamped down as a guide, I can't make a straight cut with my cheap circular saw to save my life! I also learned that I'm not smart enough to account for the width of a circular saw blade. I had intended that my sub would be 24.5" tall, so it would be half a width of MDF tall.. Well, I think one sub is going to be 24 7/16" and the other 24 9/16" tall. Hopefully I'll be able to use a router to trim everything up even.

I got to cut the peices for the tops and bracing left. I think I'll hold off on the bracing until the boxes are half built, so I can get more accurate measurements. Tomorrow, I'll probably attempt to laminate the sides, before assembly.

I'm intending to laminate the 3/4" mdf with the 1/2", then assemble the top,bottom, and sides. Install the bracing, and then the front and back. Is is easier to cut the hole for the driver before assembling the box or after?

[SteveCallas]

I may have missed it, but have you given up on the idea of using steel inbetween the layers to act as magnetic shielding?

[ThomasW]

Frozen garage, (been there done that, many times... :B )

Everyone makes noobie errors, 1/16th" off isn't worth giving a second thought.

Cut the hole after the box is assembled. That gives you a rigid surface to work with.

[---k---]

Steve,

I bought the steel today. I probably should have put the steel between the layers and built the box like Thomas had described, but I was concerned about the strength of the glue to make a composite panel. I also got the sheet metal screws, but if I'm going to the trouble to have a double layer of MDF, I want ALL the strength possible. For simplicity I cut the the sides the same size, so I think I'll just end up lining the box with the steel. If I don't line the box, I'll have deal with the gap at the edges.

Thomas,
Thanks for the encouragement. Like I've said before, I'm a great rough carpenter, but a lousy finish carpenter.
I'm not too worried about my cuts. I know I can fix a lot things with sanding, a router (which I've never used before), and epoxy filler.

I bought some Tightbond II. Do you think it would be a good idea to use Gorilla glue on the joints, since it is expansive and will fill any not so perfect cuts, or will the Tightbond be good enough?

[hamm]

I built a large box for my tempest - 23 liters tuned to around 16hz. 500 watt OAudio plate amp on it. I used a double thick baffle (1.5" MDF) for the sub, and internally cross braced it. Used dado cuts all around.

I used titebond II for the wood - i say wood glue is the best glue for gluing wood Use your finger to create a nice even-ness all the way throughout the joint. I then used a sealent in the cracks (inside the box), and glued down dampener with some 3M spray adhesive.

I keep gorilla glue around for non speaker related gluing.

[ThomasW]

Ryan,

Unlike many I'm not a fan of Gorilla Glue. It fills gaps with a foam matrix. I see no benefit to that.

I use Titebond II and caulk all the interior seams.

With MDF you can fill any surface flaws with Bondo.

[---k---]

Thanks. Tightbond II it is. Got the 4 pack of GE Silicone caulk, so were good there.

I've been wondering what was best to use to fill the countersunk holes and such. Now when you say Bondo, do you mean something like this http://www.lowes.com/lowes/lkn?actio...00075127-20052
which is more like a dry wall putty, or do you mean the real two-part epoxy?

[noah katz]

"I probably should have put the steel between the layers and built the box like Thomas had described, "

Actually that would give the least stiffness benefit, almost none in fact, because it puts the stiffest material on the neutral axis of bending.

Putting it on either inner or outer face increases the stiffness the most.

Congratulations, you're one of the few who's cheated Murphy - let's hope he's not in a vengeful mood