AVS Forum banner
1 - 17 of 17 Posts

·
Registered
Joined
·
7 Posts
Discussion Starter · #1 ·
This is my first speaker build project. It is complete and working. However, in hindsight, I am now worried that I have made an error in my calculations. I am hoping that someone could check these calcs, and if I have stuffed up the enclosure size, to recommend possible rectification. What I am concerned about is that I have made the box too big and during heavy work the driver will exceed its Xmax of 19 mm.

The subwoofer is a Dayton UM15-22 using a SPA500DSP plate amp. I selected a sealed enclosure design, with the amp sitting outside the enclosure. Listening is about 50% music and 50% home theatre.

To start with, yes, I know that Dayton recommends a sealed enclosure cabinet of 3.1 cf (about 87 litres).

However, when I first put the numbers into WinISD – and assuming I have done this correctly – it comes out with a recommended enclosure size of 0.1952 m^3 (or 195 litres), at a Qtc of 0.704. As an aside, WinISD suggests that at 87 litres enclosure, the sub will have a Qtc of 0.884.

As a cross check, using the calculator on diyaudioandvideo, it comes out with a sealed box size of 171 litres, at a Qtc of 0.707.

Speakerboxlite suggested 127 litres.

A couple of other online calculators came in around the 180 litre mark (including two that appeared to be a copy of diyaudio).

After looking at the disagreeing numbers, I built a box with a target useful volume of 171 litres. Given the amount of bracing inside the box, its internal volume is probably closer to 155-160 litres. In addition, it has approximately 30 litres of fibreglass insulation.

From a sound perspective, I am happy with the result. I have been working through various types of music and it is interesting to note the differences in source material between my old subwoofer and the new subwoofer. The most pleasant surprise was Robbie Williams, Beyond The Sea from Finding Nemo soundtrack. That kick drum, while not loud, has amazing presence.

Anyway … I started to get concerned with my construction when playing Blade Runner 2049. While the sound was amazing and shaking the windows, the mount of driver excursion had me a little worried. By eye, I estimate that cone movement was around +12-14mm. And in the scene in Rogue One, when Darth Vader’s Devastator appears at the end of the move, the amp went “click” and the sound cut out for a sec. Whoops.

So, I went back into WinISD to look at the calculated cone excursion a bit more carefully. And this is where it starts to get interesting.
  • For a volume of 195 litres, WinISD calculates the cone excursion will reach 19mm at 20Hz and 210W, with 103db, and about 106db at 25Hz.
  • For a volume of 130 litres (my build of box 160 litres + 30 litres fibreglass), WinISD calculates the cone excursion will reach 19mm at 20Hz and 270W, with 103db, and about 106 db. (Note: I have assumed that insulation can simply be taken as a reduction is speaker volume.)
  • For a volume of 87 litres (Dayton’s suggested box siz), WinISD calculates the cone excursion will reach 19mm at 25Hz and 410W, with 106db.

As an aside, WinISD calculations suggest that while the driver is rated at 600W RMS, the box size would have to be around 60 litres (Qtc of 0.998) to keep cone excursion within limits (at an impressive 117db at 60Hz).

So, in hindsight, I can see where I went wrong. In chasing the best theoretical Qtc, I overlooked driver excursion limits.

Getting back to my build, I see a couple of possible actions.
  1. Do nothing. However, respect the limits of cone excursion at 240W. Particularly for material that potentially contains subsonic material
  2. Install more fibreglass. Adding 30 litres (box volume of 100 litre) increases the power limit to 350W before cone excursion limit is reached. More importantly, it does flatten out the sub-20Hz response curve so there is less chance of a large subsonic signal causing driver excursion.

So, my questions are;

A) Do my calculations seem correct?
B) Will adding more fibreglass solve this problem? (I really want to avoid doing stuff like adding sand or closing off part of the box.)
 

·
Registered
Joined
·
78 Posts
A) Yes

B) No, more fiberfill will not fix your problem. Your best option would be to build a smaller enclosure. If thats not an option you can use the dsp in your amp to reduce the ULF output, but unless you can set an actual voltage limit you'll still need to pay attention to your signal levels.
 

·
Registered
Joined
·
403 Posts
The easiest solution is to remove all the fiberglass fill, and line the walls with closed cell foam to reduce the enclosure volume. Add as many layers as you need to reduce the enclosure volume until have it where you happy with the results. If you find you reduced the enclosure volume to much, no problem . Just remove some of the closed-cell foam .
 

·
Registered
Joined
·
3,963 Posts
Put a HP filter in place. I personally wouldn’t make the box smaller, high Q boxes sound drab to me.
 

·
Registered
Joined
·
476 Posts
I don't really see a problem with an oversized sealed box. I would play with WinISD and find the power level where cone excursion is safe and then try not to exceed that in practice. The benefit of a lower Q box as you've seen is the flatter response. If you want a true flat response, you're going to end up EQing out the top end anyway, since it'll be much louder than the low end. So I would just set the gains such that you don't have excursion issues at 15Hz or so, then EQ the top end flat. If you feel like you don't have enough bass at that point, you need more wooferage.

Also don't forget that amps roll off on the low end. WinISD keeps power constant. So low end over-excursion might not actually happen.
 
  • Like
Reactions: InTheDeepEnd

·
Registered
Joined
·
3,062 Posts
A rare find that somebody made a box too large instead of too small.

From what I recall, the UM has a decent Xmech so going a bit past Xmax should not be a problem. Normally, I just let it roll 10% past Xmax in the simulations. Mainly because as you get near and go past Xmax the suspension typicallly "tightens up" as the spider and the roll on the cone become more reistant to movement, the part the simulators don't take into account. Some drivers have "progressive spiders" or the spider coil tresists movement the farther the travel becomes to protect itself. Quite common in car audio subs.

Very easy to make an enclosure smaller internally. As others have stated you can use closed cel foam to make it smaller. Heck, you can make a pair of 30 liter boxes, smal ones to fit inside and add one or two to figure out what you want. Take anything from a pile of kds wood blocks, some strong tape and put them inside. Get wild an tape a bunch of bricks together and go for it. Once you find the proper size you like, fabricate a piece to take up that space.

Normally, most people make them too smal which can be fixed but a pan to do. My plywood stretcher never seems to work right. :(

In summation, I'd go 10 to 15% past Xmax in the simulation at 500 watts and call it done. If your amp shut down, that usually means it has been overdriven, the amp itself--not the subwoofer itself. The UMs are commonly used by car audio folks that pound them and they hold up to abuse quite well.

Extra bonus, since you shut down the amp by overdriving it---good excuse to add more subs! ;)
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #8 ·
Thanks to everyone for the comments and the feedback. It has given me a few things to mull over. The scene in Rogue One that caused the amp to cut out did concern me. Yeah, I was playing it loud, but not that loud that I though I was near limits. I suspect that the sound engineers inserted a nice little sonic surprise and the amp just said “nope”. There is no way any of the music that I play would have that type of thump. So, the concern was more with playing movies that had exceptionally wide dynamic range.

RBhifi & 18Hurst, I do like the idea of inserting something like closed cel to reduce its volume. Maybe that can be a project for another day. That said, given the cost of MDF verses the foam, it might be easier just to build a new box.

Gorilla Killa, now that you have mentioned it, the Dayton SPA500DSP does have a subsonic filter. It has an adjustable cut off (lowest 25Hz) with either a -12 or -24 dB slope per octave. Applying that filter is probably the easiest way of allaying my fears for now. Looking at the WinISD curve, the cone excursion only becomes a problem (at 300W) for frequencies below 27Hz. In a few test frequency sweeps, I can barely hear below 23 Hz as it is (even though the cone is clearly moving). I might miss out on a few subsonic rumbles but applying the filter is more likely to preserve the driver than to see it disappear in a puff of smoke.

This was my first build from start to finish where I did all the calculations by myself. I have learnt a lot about loudspeaker design during the build and the trade-offs that a manufacturer must make. All in all, I am really happy with the finished product. I much prefer the sound from a sealed unit and the old ported commercial unit has been given a new home. Visitors are amazed at the sound. (However, the cat does not approve.) I also now have the speaker building bug.

Rectangle Slope Plot Font Line

Rectangle Font Parallel Slope Plot
 

·
Registered
Joined
·
642 Posts
I would keep it big.

If it shuts off or bottoms out, turn it down, or implement a high pass filter.

My favorite subs so far are in an alignment with a Qtc just over .5, and they're killing it.

A smaller box will limit excursion, by making excursion more difficult, which means it takes MORE power to get the SAME output below resonance, which means MORE coil heat, and MORE stress on the amp. You get a narrow band increase at the system resonance, which will move up in frequency in exchange. Not for me. I understand the rationale, and a Qtc of .707 looks nice lining up theoretical excursion limit lines (which are only theoretical) with theoretical power limits (which in reality vary a bit depending on the frequency and load, when you're talking about real amplifiers driving real speakers), so IMHO, F-smaller if you're already paying for the living space of a large enclosure, and you have EQ abilities. You can peak the response to emulate the sound of a higher Qtc alignment, if you like that (and there's nothing wrong if you do), and not burn up power needlessly fighting additional air stiffness to get your low low kink.

JMO. :)
 

·
Registered
Joined
·
1,131 Posts
What volume did you use for the driver itself ? I cannot find the drive volume number anywhere, I don't understand why manufactures don't include this number in their specs.
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #12 ·
What volume did you use for the driver itself ? I cannot find the drive volume number anywhere, I don't understand why manufactures don't include this number in their specs.
To be honest, I don't know. I just entered the driver dimensions into calculator and it worked out the rest. That said, looking at the raw numbers, I assume the calculator is estimating a volume around 30 litres.
 

·
Registered
Joined
·
1,131 Posts
I was under the impression WinISD doesnt take into account the driver's own volume and you have to add that to the box volume when making up the dimensions.
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #14 ·
I was under the impression WinISD doesnt take into account the driver's own volume and you have to add that to the box volume when making up the dimensions.
First up, an apology, the volume I mentioned above of 30 litres is wrong, by an order of a magnitude.

On the Dayton web site for a UM15-22, they have this figure.

Volume of Displacement (Vd)1
548 cm³
Which I assume is 5.48 litres.

In WinISD, in the driver editor, in the dimensions tab, plugging in the driver dimensions from the Dayton web site, gives 0.0056m^3, which I assume is 5.6 litres.
 

·
Registered
Joined
·
1,699 Posts
On the Dayton web site for a UM15-22, they have this figure.
Vd is the amount of air the cone can move and is calculated by multiplying Sd and Xmax.

I used .25 cubic feet for the space the UM15 would take up, but it's probably a little less than that. I would rather lean towards a larger box in my approximations, but in practice for simple sealed and ported boxes, it probably doesn't make much difference.
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #17 ·
Vd is the amount of air the cone can move and is calculated by multiplying Sd and Xmax.
Ahhh, OK, thanks. Sorry HFGuy, my bad.

Still, I assume that the various speaker designers account for speaker volume when they ask you for the speaker dimensions.
 
1 - 17 of 17 Posts
Top