Alarmed by Modeling vs Real-world Results

[mosconiac]

I've seen some issues on the 'net where folks have built vented enclosures using the latest in software only to find that the results were not as expected. For example, software may tell you to cut your port at 12", but the real-world port ends up at 10" for the desired tuning frequency. Let's not forget that damping material will throw the tune off too.

This has me alarmed because I'm about to make MDF dust for a Shiva-X and I don't want to build something only to tweak it endlessly in the field until its right. I expect to experiment a little, but taking the driver out 10 times to tweak the box volume or port length or stuffing would become ridiculous.

WinISD says I should build a 5.4cf enclosure tuned to 18Hz with a 36 sq.in. port 36 inches long. Troels Gravesen's empirical data indicates that the actual port length can vary from 60-80% of predicted!!!

http://www.troelsgravesen.dk/vent_tuning.htm

In order to accomodate a 36" long port without resorting to u-turns, I need a longish enclosure which forces an end-to-end resonant frequency low enough that I'm worried it will get excited. Port resonance is 188Hz and the lowest box resonance is 144Hz! If I build the box long to fit what the computer says only to find that the real-world port could have allowed a much shorter box...well...

When I stuff the box:
1) what should I use (703? wool? foam? polyfill?)?
2) how much should I use?
3) should I build a 5.4cf box if WinISD wants to see 5.4cf?

When I fab the port:
1) should I make it 36" long like WinISD says?
2) should I start @ 29" long (80% of predicted)?

What have your experiences been with respect to real-world vs software???

[Mike P.]

Troels Gravesen's empirical data indicates that the actual port length can vary from 60-80% of predicted!!!

The results are from a 29.5 Liter box using mathematical equations from www.diysubwoofers.org. The bigger the box the more "forgiving" the port length can be. Assuming your box volume calculations are accurate, WinISD has shown to be within 1 or 2 hz from tuning frequency stated in bigger then 4 cu.ft. box volumes. Depending on how much stuffing you use, it will slightly lower the tuning frequency and port output.

[collo]

I assume you are using a flared port with the Shiva-X

WinISD uses the effective length of the port. Physical length will be slightly longer as shown here:

You should also try and have a clearance of one port diameter (unflared section) in all directions from the inside end of the port.

You generally do not stuff a ported box, but treat the walls that are furthest apart with something like closed-cell foam or heavy felt. If you do add stuffing, ensure you have an unobstructed path from the driver to the port.

[SteveCallas]

I wouldn't be alarmed - I've never seen a project have a port length off by 60-80%. I've seen tuning be off by 1-2hz, but that's not really a big deal.

[mosconiac]

OK, it looks like the experts have calmed my fears. How does this sound?

1) Build the enclosure slightly larger than 5.4cf (recommended by WinISD) so I can adjust it downward as needed.

2) Build a port (3/4" flare) with 36 cu.in., 36-3/4" long and shorten if necessary to achieve the tuning I want (21 Hz).

3) Line the furthest ends of the box with a few inches of OC 703 (no other fill) to tame the 144Hz box resonance.

[Mike P.]

Sounds like a plan.

[Jailer]

Don't be too alarmed with the results of the modeling software. It comes in close enough for it's intended purpose.

I modeled mine entirely with Winisd and it came in about 2 Hz higher than predicted on the tuning but it's good enough for me.

[Ricci]

All of the vented subs that I have measured have come in tuned very close to what is predicted. Within 1-2hz as others have noted. It's not a big enough difference compared to some other variables to worry about it.

[looneybomber]

Way OT here, but Ricci, your avatar shows your dual SDX-15 by Funkywaves, but what is that speaker next to it with what looks like the TD drivers by AE?

[Ricci]

looneybomber wrote:

Way OT here, but Ricci, your avatar shows your dual SDX-15 by Funkywaves, but what is that speaker next to it with what looks like the TD drivers by AE?

I wish they were TD's! and they might be in the future, but for right now they are Klipsch Epic CF-4's. They use dual neo 12's, a 2" compression driver in a tractrix horn and weigh 102lbs each. I drooled over them at my local hifi shop when I was 14 and just getting into audio. Picked up 2 pairs of them used over 10years later for right at $600pr. They aren't the most liquid and absolutely pure sounding speakers ever, but they sound good to me and will kick out the jams for HT or parties.

[DrWho]

A model is only as accurate as the assumptions behind the model.

If your design is such that the assumptions are still valid, then you should get extremely well correlated results between the model and the real world. You shouldn't, however, blindly plug numbers into a program until it shows a pretty line....that's a recipe for random results.

[Vasyachkin]

you worry too much

normal people just buy a generic box and are happy with tuning frequency two octaves off ...

ported boxes are compromised to begin with.