Re-sizing a horn in hornresp?

[Samps]

When re-sizing a front loaded horn in hornresp, do you simply take the percentage of the original size and multiply S1-S5 times that, along with Vrc? 

 

So if I was re-sizing an existing horn that was designed to be 20 inches wide, and I wanted to make it 10 inches wide (sizes picked for simple math) do I simply multiple S1-S5 times .5 and Vrc times .5? 

 

That seems logical to me since the lengths of each segment remain the same while only the volume of each segment is reduced, along with the volume of the rear chamber.

 

Should I also change the Lrc or does that spec even matter?  Are there any other values that need to be adjusted? 

[SME]

The 'S' variables are areas, not lengths.  When you want to resize a horn, which dimensions are you talking about?  Are you planning to scale down the horn in all dimensions equally?  Or maybe only scale in width and depth?  Or something totally different?  The answer you are looking for depends on your answer to these questions?

[Samps]

I'm trying to take existing designs and testing them at different widths with different drivers.

[SME]

Basically each S is the cross-sectional area of the transition between each horn segment.  So if you are only changing width, then you need to know the orientation for these transition areas.  If the area is oriented perpendicular to the width, then the amount of area won't change at all.  If the area is perpendicular to the height or depth, then it will change proportionally, like you are doing above.  If none of the above is true, then you will need to do some trig calculations.

 

For the horn segment lengths, if the axis of the segment is perpendicular to the width, the segment length will change proportionally.  If the axis of the segment is perpendicular to the height or depth, the length will not change at all.  Otherwise, you'll need trig.

 

If all you change is the width, than the volumes should change proportionally like you are doing above.

 

Does this help?

[Samps]

Thanks. I think the math is pretty simple. If I'm keeping the design of the horn intact and only changing the width, all of the segments should decrease in volume by an equal amount. For example, if you took the lilmike F20 and doubled the width it would double the volume of each segment of the horn.

[lilmike]

Not so much an equal amount, it is an equal factor.

 

Area = width X height. When you change things as discussed, you're varying height, so if the new height is 1.2X the old one, areas and volumes will be 1.2X larger, lengths are unchanged.

 

This works fine within a range of widths for typical folds where there are one or more parabolic flares, all in the same axis. The range of widths has to do with the aspect ratio - too high of an aspect ratio will behave like a skinny slot port - it will get lossy. Nothing more than some bad results to base this on, but I draw the line at about 1.5 to 1.6 X the driver's width now.

 

I do it all the time when I'm designing things (more like - I did it all the time when I used to design things). I typically use an excel sheet and calculate a range of widths, then evaluate the results. This approach allows me to try different drivers and compression ratios, optimize panel utilization, and look at changes to reactance and loading.

 

Sure, it is a bit over-simplified, but then so is the approximation of the horn in HornResp. Certainly good enough to evaluate alternatives.

 

Lrc isn't much of an issue in sub horns. The volumes will scale like the areas, but the lengths aren't all that relevant (has to do with resonances and reflections IIRC).

 

Now - when it is a 3D fold instead of 2D, the math gets a LOT more complicated and simple scaling like that goes right out the window.

[Samps]

Thanks Mike! I'm not going to build a double wide F20, I was just using that for an example. I tried modeling the F20 fold with a 12" but you can't make it narrow enough to get the response in line. So I'm going to build another THT but wanted to check the different width options with the HF12 to find the sweet spot. My model is likely off a bit but it should be close enough for what I'm doing. I already know the design is decent. The full size F20 is too big and I'm set on a FLH.

[Kyle]

I typically don't model in horn resp but it might benefit you to first model out your subwoofer in 3D, make the adjustments there, and then use the tools to calculate various areas or dimensions and feed those back into the model.  Then if you go back and make more adjustments, start back at the model and pull out the new numbers. This gives you a 1-way flow of information that is easy to maintain. The other advantage is sometimes your locked into dimensions due to construction constraints so its best to model what you can really build. Might as well bake that in from the start. There is no point in modeling what you can't build...

[lilmike]

I typically don't model in horn resp but it might benefit you to first model out your subwoofer in 3D, make the adjustments there, and then use the tools to calculate various areas or dimensions and feed those back into the model.  Then if you go back and make more adjustments, start back at the model and pull out the new numbers. This gives you a 1-way flow of information that is easy to maintain. The other advantage is sometimes your locked into dimensions due to construction constraints so its best to model what you can really build. Might as well bake that in from the start. There is no point in modeling what you can't build...

+1

 

I draw first, then model. I typically have an idea about what it "needs" to look like, so that helps. I draw a 2D layout of the fold and panels in Sketchup, then I make a spreadsheet with the widths and their respective length from the throat. A plot of this helps identify any mistakes in the expansion rate. I don't even bother with a 3D render until I have a layout and model I like. For me, unless it is a super-complicated cabinet and the render is required to know what the heck is going on inside, the render is the last step, sometimes I build prototype(s) before I draw it all up. 

 

Lengths (L12, L23, L34, and L45) are determined based on where the flare rate changes and where the driver enters (in TH and OD), the distances are calculated using the corner approximation technique described by Soho54 as "Advanced Centerline". They're generally close enough. 

 

Widths are measured directly from the layout. Once I know widths, I calculate the areas by changing the height of the internal panels. This is done in a simple spreadsheet approach. Where appropriate, I also calculate volumes this way, based on the chamber area and height.

 

Still do most of my modeling in hornresp. Not doing much of that at the moment though.

[maxmercy]

Pretty sure the HF12 in a THT models  best around 15" wide....here it is compared to HF15 version 24.5" wide, both cabs modeled at Xmax:

 

 

JSS

[Samps]

Thanks guys. I'm not trying to design a new horn. Only taking a THT and trying to figure out how big to make it. I've built two THTs. A 17" wide with a HF12" and a LP at 18x24x72 with the DVC 15". The 17" both sounded better and had a flatter response. The DVC is not the same driver now that it used to be. So I deciddd to build another with the 12" which so far has been my favorite sub for music. I'm trying a 19.5" wide since that was as big as I could build out of two sheets. It may just have a big hump around 30hz but I'm not so sure that's a bad thing for music. Sort of a built in house curve that won't need any boost. I'll high pass it to get some extra spl at the expense of extension. This will be 100% music. It will be nearfield with a pair of Dayton PA460s helping out up front.

 

Next project will be a custom horn which I will apply some of this great info.