Designing the O.B.E, a tapped horn project

[Twellmann]

Ok, here's the deal. I´m attempting to design/build a tapped horn for a couple of my friends.

It's job is to act more as a bass-enhancer, rather than an actual subwoofer, thereby coining the term Outdoors Bass Enhancer

Here is a list of design parameters and questions:

Driver: P.Audio HP-10W http://www.ljudia.no/Archive/Product...nformasjon.pdf
I have two of these to work with

Power: Amp6-Basic, this is a small chipamp based on the TA2020. It gives about 10W@8Ohm

Size: That, I'm not really sure about. They have a speaker now which is ~4Cu.ft/~100L. The O.B.E could be larger, but that is something we've not yet settled on.

It will mostly be used outdoors on the ground, so no cornerloading.

My idea was for it to cover ~50-100Hz, and obviously be as sensitive as possible. I don't mind the FR being peaky or rough since it will mostly be listened to by a bunch of somewhat intoxicated festival-goers.

So my next step is obviously messing around in Hornresp trying to find something that suits me(them). That leads me to a couple of questions:

Bending the damn thing. I've browsed through a lot posts with regards to bending a TH, and it seems that most people agree on one thing. That a little variation in the taper, in the bend, isn't that big of deal.

Two drivers in the same horn This is something I'm not really sure about. There are many THs with two drivers, most of them have one driver closer to the mouth, and the other one further in the horn. How do I calculate that? Obviously Hornresp can model the position of one driver, but not two. Do I simply place them as close together as possible and expect to get mean response, somewhere between the two?

Throat area My understanding is that in Hornresp throat area is called S2, and it is the area perpendicular to driver, right in front of the driver. That is at least how the diagram shows it. But some THs use an extra throat area, by making the hole in front of the driver smaller tha Sd. I can't find out which it is.


Attached are pics of Input Parameters and Response, and also the Hornresp file

Have a look while I continue messing around, any input is appreciated

When I begin construction I will post some pics

[littlemike]

Personally, I'd shorten L12 a bit, lengthen L23 a bit, and increase the throat and mouth size unless cabinet size becomes an issue. Remember, there are two drivers in the simulation, so you need to have enough room to fit them in.

With a little bit of work, you should be able to get over 100 dB at 2.00 v 2pi from 50 Hz to 200 Hz.

[Oklahoma Wolf]

Originally Posted by Twellmann

Two drivers in the same horn This is something I'm not really sure about. There are many THs with two drivers, most of them have one driver closer to the mouth, and the other one further in the horn. How do I calculate that?

The 17cm L34 value in your screenshot gives you the answer... use this value as the midpoint where the drivers mount. In my horn, which uses one woofer in front of the other, one is behind the L34 point and the other is in front. Just make sure L34 is long enough to fit the drivers in the horn.

+1 on shortening L12. Current value will make for a little bit of a messy impulse response.

[bjorno]

Originally Posted by Twellmann

..My idea was for it to cover ~50-100Hz, and obviously be as sensitive as possible..

Hi Twellmann,

See my test simulation for a 50 Hz TH; picture #1.

Qts= 0.35? Have you made real measurements of the T/S for the P.Audio drivers See the picture #2?

b

[Twellmann]

Originally Posted by littlemike

Personally, I'd shorten L12 a bit, lengthen L23 a bit, and increase the throat and mouth size unless cabinet size becomes an issue. Remember, there are two drivers in the simulation, so you need to have enough room to fit them in.

With a little bit of work, you should be able to get over 100 dB at 2.00 v 2pi from 50 Hz to 200 Hz.

Originally Posted by Oklahoma Wolf

The 17cm L34 value in your screenshot gives you the answer... use this value as the midpoint where the drivers mount. In my horn, which uses one woofer in front of the other, one is behind the L34 point and the other is in front. Just make sure L34 is long enough to fit the drivers in the horn.

+1 on shortening L12. Current value will make for a little bit of a messy impulse response.

Originally Posted by bjorno

Hi Twellmann,

See my test simulation for a 50 Hz TH; picture #1.

Qts= 0.35? Have you made real measurements of the T/S for the P.Audio drivers See the picture #2?

b

Thanks for the replies, after a bit of fiddling I got something very similar to bjorno's, the biggest difference being that I aimed for similar taper on all three parts. In my mind it would be easier to translate in to a bend in a box.

Room for drivers... well, right... I kinda didn't consider how much space the actual drivers would occupy, thanks for the reminder

Qts... I haven't measured the woofers, and I suppose it would the cleverest thing to do, all I need is a no-cost method.

One other thing, the damn thing is quickly getting huge

We actually have 4 of the drivers, and my idea was to make a small box with two of the drivers and some tweeters. And then make the sub to maximize bass-output. But maybe it would easier to just make one big ported box for the whole lot. We would lose flexibility, but also decrease difficulty (of the build)

[Twellmann]

I don't know what has happened with the T/S, but I've changed it to the correct parameters now, at least they are the as what's on paper

[Twellmann]

Update

I have been fiddling some more with hornresp, and I'm starting to understand how the different parameters affect each other.

The size of the box hasn't been determined yet, but that will probably come soon.

One thing I have been busy with is finding out how to bend the damn thing. I've found a method that is rather straight forward, at least when you only have a single taper (the same taper in all segments)


I attached some pictures to show my method, and I'd like some comments please.


The first picture is a 3d image that shows the horn being bend. It starts with a stumped triangle, that is cut into three pieces, that are rotated relative to each other.

The second picture shows how I find out where to divide the segments. It's simple geometry that helps to find the angles.

The last picture shows something like the first one, but this one shows how the segments end end up as rectangular box (almost).

This way of bending it is a compromise, but I dont't think it will make a big impact on the final sound.


If you have any questions about the pictures, please feel free to ask.

[littlemike]

I currently use a similar geometric approach. I am actually still refining it. To be honest, I have never folded a tapped horn using the same method, I've done things differently each time. Certainly, some approaches have worked better than others.

I do all my folding in Sketchup, I do the bulk of the math in a spreadsheet. When folding a tapped horn with my current method, I draw a square in the fold to "locate" the corner, then adjust the length and area through the corner based on geometric calculations that approximate the path length change through the bend based on the dimensions of this square.

I've not built one that I have folded up this way yet, but hope to get to that shortly. As such, I do not know if it will be any more accurate than what I have done in the past until I turn the pixels into sawdust and make some measurements.

Undoubtedly, I will learn something from the process, which will probably result in yet another change to my folding approach.