Hornresp tutorial – Transmission line design
How to design a Transmission line using Hornresp
While this app is great for designing horns, in this Hornresp tutorial I’m going to show you how to properly model the response of a transmission line enclosure and actually achieve a good result. Many choose a T-line design for a full range speaker. However, I’m going to build a subwoofer using the Dayton Audio Ultimax 8″. You might be familiar with my first attempt in building a T-line, and you should probably take a look at that first. Now, I’m using Hornresp to properly model the frequency response of the enclosure. Also, we are going to modify the first enclosure to get the results we are after. As the first one was pretty much a failure.
This is the written version of the YouTube Hornresp tutorial. I suggest you watch the video just above. If not, continue reading.
Introduction
If you didn’t have the patience of reading the first T-line article, here’s a rundown. We got an 8″ subwoofer and we designed a transmission line enclosure for it. The area of line needs to be equal or larger than the effective cone area of the speaker. In our case the area is a bit larger than that. Not the best choice (the box could of been smaller), but it works. Next, we need to find the length of the line. Here are the steps :
- Find the resonating frequency of the speaker : 31.6 Hz
- Find the wave length of that frequency. Divide the speed of sound by that frequency. 343 / 31.6 = 10.85 m
- Divide that number by 4. So, 10.85 / 4 = 2.71 m. This is the quarter wave length.
In conclusion, the line needs to be 271 cm long. Of course, since we are using modeling software you can play with this length. But make this your starting point. Or, maybe, just keep it that way.
Hornresp tutorial
To get this app just visit http://hornresp.net and download it from there. The interface starts off very abruptly, because it contains some acronyms and some numbers. As a result, you already realized that this app has a steep learning curve. I divided the starting screen into 4 sections and I’ll discuss 2 of them right of the bat, as they are easy enough.
Before you do anything, just press the Add button to start the project. Up top, in section 1 just double click the Ang box. When I say box, I’m referring to that where it says “0.5 x Pi”. Just double click it and select half space. It doesn’t really make any difference what you select. The overall response curve will look the same regardless what you select. But anyway, that’s what I do. You can ignore the rest of the boxes.
Down below, at section 4, just select every box and inset 0 (zero) inside every box.
And now you are ready to begin. Hornresp tutorial going smoothly for now, but it will get more complicated.
Section 3
In this section you enter the parameters of your speaker. As a result, make sure you have the spec sheet of the driver handy.
Just double click the Sd box and a window will pop up. Simply copy all of the parameters from the spec sheet and click OK.
Section 4
Now we are getting to the hard part. First of all, bear in mind that this app was mainly designed for horns. However, it can be used for transmission lines as well, as they are close relatives. Secondly, I need to explain some things about horns.
Take a look at the illustration above. A horn has a throat and a mouth. In a horn design the mouth is larger than the throat. They both have different areas (or surfaces) which is marked by S1 and S2 in Hornresp. So S1 is the area of the throat and S2 is the area of the mouth. The neck is the length between the throat and mouth and you can input the length in the next box, right where it says “Con”. Now, let’s analyze the next image :
The neck of the horn can have different profiles. The growth of the horn can follow different mathematical functions : conical, exponential, parabolic etc. You can select this by double clicking the text “Con”. By continuously double clicking it, you can swap between the different horn profiles.
You can input multiple horn sections (up to 4). So, if you you horn starts conical then goes exponential. You input S1-S2-Con and then S2-S3-Exp. With the appropriate lengths, of course. Since this is a Hornresp tutorial, you got to learn about horns first. However, T-line guide is just around the corner.
Transmission line design
Now that we learned how a horn works. Let’s accommodate this app for transmission line design. Basically, if you convert the horn into a transmission line, the area of the mouth is equal that of the throat.
Like I said before, we chose an area which quite larger than that of the Sd of the driver. It’s a 19.2 x 19.2 cm square. If we multiply the numbers we get an area of 368.64 cm2. As a result, we put this number into S1 box, but also in S2. And the length is 271 cm. As for the growth of the “horn”, it doesn’t matter what function you choose. Since S1=S2, it will be a straight line between the mouth and throat. The logical thing to choose is conical, but it doesn’t matter. However, we will input a second section as well. Why? You will see in a minute.
So, as you see from the image above, we entered another section S2-S3, with the same area, of course. I split the length between the 2 sections. 135 cm for first section and 136 cm for the 2nd section. In total, it’s 271. If you look where the speaker parameters are, you will see an “ND’. This means single drive. If you double click it, it will transform in a red “OD”. This means offset driver. To have an offset driver you need 2 sections. The driver will be placed at the end of the first section and the beginning of the 2nd section. As a result the driver is not at the beginning of the line anymore.
Why do we offset the driver. Because it will result in a frequency response that otherwise would be impossible with the driver at the beginning of the line.
Line adjustments
We are pretty much done with the opening screen. To continue our Hornresp tutorial please go to Tools-> Loudspeaker Wizard. If it asks something about masking resonances, just say yes.
The first thing you want to do is click the lower-left drop down and select Power. I like to uncheck “Show Baseline”, but it’s up to you. The sliders we are going to fiddle here is the ones from the bottom :
- L12 Con – Length of the 1st section
- L23 Con – Length of the 2nd section
If you make L12 shorter, then L23 will be longer to maintain the 271 cm long total length of the line. However, if you Make L23 longer, to total length of the line will be longer as well. Remember that the driver is placed between these 2 sections. If you are curious on how the response would look like if the driver wasn’t offset, just slide L12 all the way to the left, basically having just one long 2nd section. Which basically is the same thing. If you have done that, we have a very large peak at 90 Hz and huge cancellation between 100 and 200 Hz. What we want to do is position the driver in such a way, that the cancellation moves to the right a bit.
As you can see, if we position the driver at 93 cm away from the beginning of the line, that cancellation moves to the 200 Hz region. Which is great, because this is subwoofer and we are not interested in those frequencies anyway. Now we just have to deal with those peaks.
Filling
Do deal with those peaks, we need to use damping material. For subwoofers we need to use high density stuff. Not that egg-crate sponge absorbent pads. Those will do nothing. In conclusion, use rockwool or mineral wool. It’s used for roof top insulation and can be found at any home depot.
How to design loudspeakers - video courses
In the window we are right now (loudspeaker wizard) just click the middle drop down list and select Filling. Now, the sliders need a bit of explaining. Fr1 is the density of the material. Since we are using rockwool, this is very dense and take the slider all the way to the right for 1000 units (rockwool has several thousands but past 1000 it doesn’t really matter). Do the same for how many segments you have. Presuming you will use the same material on each segment.
The slider Tal1 tells the app how much of that segment is filled. If the slider is in the middle. The segment is fully filled. If you slide it to the left. It will show a percentage, let’s say 60%. That means the segment is filled 60% starting from left to right. And respectively, if you slide it to the right, it shows much the segment is filled starting from right to left.
As you can see, I filled the first part of the line by 40% and just 3% of the end of line. This is because the app considers that the last 3% will be fully filled. However, I’m going to damp a longer part of the exit, but just line the walls with damping material (not fully fill). That’s why I chosen only 3%. This is the final response and that concludes our Hornresp tutorial. Have fun designing T-lines.
Build my own project
If you want to copy my own project and build the Dayton Audio T-line sub, you can follow these precise instructions.
This is how the enclosure will look like, with exterior dimensions. It’s pretty big. I used 18 mm MDF. Here are the parts list. These are Affiliate paid links:
- Dayton Audio Ultimax 8″ (Sound Imports) (Parts Express) (Amazon)
- Dayton Audio Binding posts (Sound Imports) (Parts Express) (Amazon)
- Rockwool (Amazon)
- Self-adhesive damping material(Sound Imports) (Parts Express alternative) (Amazon alternative)
Panel Dimensions
Here are the dimensions of the panels in mm :
- Front upper : 648 x 210 x 18 mm
- Back upper : 648 x 210 x 18 mm
- Front lower : 648 x 228 x 18 mm
- Back lower : 648 x 228 x 18 mm
- Sides (x4) : 434 x 192 x 18 mm
- Top : 648 x 434 x 18 mm
- Middle : 648 x 434 x 18 mm
- Bottom : 648 x 434 x 18 mm
- Inner panel long (x2) : 434 x 192 x 18 mm
- Inner panel short (x2) : 242 x 192 x 18 mm
These images should be sufficient for you to figure out which panel goes where and how to cut it.
Final thoughts
So, we concluded our Hornresp tutorial and finished our transmission line subwoofer. Let’s look at the measured frequency response and compare it to the one modeled by Hornresp.
This is the anechoic frequency response. Also, bear in mind that it’s from 10 to 200 Hz. It’s a subwoofer, the rest doesn’t matter. And look how good it looks. 33 Hz – 110 Hz super flat. It’s impressive to see an 8″ subwoofer play that low. Furthermore, since this is transmission line, the power needed to drive this subwoofer would be lower than expected (since the driver is playing in basically free air). Finally, if I judge the sound signature, it does sound like something I never heard before in a sub. It’s simply blends really well. Very transparent. It’s like you don’t have a sub at all. However, if you shut it off you feel that something’s missing. T-lines are something else. I haven’t tried them before, but now I’m glad I did.
14 comments
I love your posts and loudspeakers design courses. They have very practical information that is easy to understand for the uninitiated.
I am planning to do mid tops and I would like to know how to calculate in Horn Resp and the steps to take into account.
Hi Marius,
Just a couple of comments relating to your excellent transmission line design tutorial:
1. The response will indeed be affected by the solid radiation angle (Ang) setting used. How much of an effect depends on the system being simulated.
2. A more accurate result can be obtained by using an appropriate non-zero path length value. (A zero value assumes that the direct radiator and port outputs are co-located).
Kind regards,
David
This post has been awesome, but I cant seem to wrap my head around the Lining portion. The 40% in the first section; does that mean only the first 40% of the walls are lined, or did you stuff layer after layer of rockwool into the line until you reached the 40%?
The same question could be asked about that second section.
Sincerely,
Jonny
It means 40% of the first section is completely filled. For the 2nd section I selected 2-3 percent completely filled. When you select a percentage of stuffing, that’s what it means, how much of that segment you will completely fill. And since I only lined the walls of the whole last part, more than a quarter of the 2nd segment, I chose only 2-3% percent in hornresp. I don’t know if that’s true, but it’s something I tried to equate : 3% fully filled = 30-40% of lined walls.
A great beginners guide and easy to understand- thank you very much! A few questions: If you have a dual 2 ohm voicecoil car sub like me, which I wire in series to give a nominal 4 ohms, and the TS Specs give the Re as 1.6 ohm, do I use 1.6 or should I double it and use 3.2 ohms in Hornresp and other modelling software? There is a huge difference in enclosure size needed if I use 1.6 vs 3.2. Also, how do you decide at which frequency to tune your TL? Do you use the Fs of the driver as the tuning frequency, in my case 36Hz, or can I tune it lower to say 25Hz? If I tune it to 25Hz does that mean the fr will be flat to 25Hz and then taper off? Thanks.
If the sheet gives the specs with the coils in parallel, then in series: Re is four times larger (6.4 Ohm), Le is 4x larger and Bl is 2x larger. You should tune the line to quarter wavelength of Fs. But you can tune it lower if you want. That’s why you have Hornresp, to check how will the response will look like.
Very good explanation for those like me who had no idea how to start. Do you recommend some brands of subwoofer in particular that give good results to reach those 10hz that you see in the graph? or avoid some?
Thank you.
The main thing you should look for is the Qts value. The speaker should have a Qts higher than 0.7
Aloha 🙂
This has been a very-very interesting and inspiring read and your end result for the subwoofer use is to me astonishing and groundbreaking.
Question: I’ve downloaded the hornresp program and my 8″ drivers have almost the exact same parameters, so your offset driver recipe is the way to go.
You chose a square shaped neck, which was easily calculated, but how would the program come about with a pentagon shaped neck ?
I can’t find any reference to this shape in the help file and I can’t see where to type in the parameters for such a shape.
The pentagon shape has an inbuild capability of eliminating standing waves, hence why Bang and Olufsen chose it for their Beolab/Beovox Penta speaker series.
How would a pentagon shaped transmission line behave in terms of reflections and frequency amplifying/dampening capabilities, like the ones you see as spikes/valleys on your measurement results ?
Hey Marius, first of all: Thanks a lot! As I started building DIY speakers two years ago, your tutorials and yt channel helped me not only to get the basic concepts but you ignited my enthusiams for this hobby. So, one more time: THANKS!
But what I like to point out in regards to this t-line damping material thingy is: Isn’t Rockwool a bit dangerous? I think it might blow a lot of tiny particals in your living room while using those speakers.
Afaik it is highly recommended to wear a mask when working with rockwool, so I wonder if that might be a concern when using it as damping material in a speaker with a port.
Yeah, you’re right. That might be a health hazard. I’m not familiar with construction materials nowadays. Aren’t these materials safer now? Made from friendlier materials?
Hey, thanks for the answer. Not sure if it makes sense to puplish this post but I like to try to answer your question.
I did some quick research and it seems like there was at least some improvements on the hazourdness of rockwool in the past 20 years. But most source I found still pointed out that it is still an issue. As far as I understood, rockwool produced prior to year 2000 was suspected to be cancer causing. So hopefully thats not an issue anymore but still rockwool can release “respirable” particles which can reach and reside in the lung.
I wonder if this can be addressed by adding a thin layer of filter material. Of cause it would need to filter quite tiny particles similar too a ffp2 mask (which is what manufactures suggest to use when working with rockwool).
I’m just not sure how that would influence the accoustical features of the damping, thats most likely something you can judge way better than me.
Hi Marius,
many thanks for your great work about this very complex but very interesting matter.
Just one question: should it be possibile/reasonable design a TQWT using a fullrange speaker, e.g. Fostex FE103nv2 (if I remember its name correctly), using Hornresp?
Many thanks in advance.
Kind Regards,
Daniele Colombi
Hello
Your post is excellent, thanks a lot.
As I plan to build a two ways TL, my mid-bass driver has to work till 2000 Hz. How to avoir the hole in thĂ© response at roughly 200 Hz. I’ve been Reading it’s possible to position a small TL adjusted for 200 Hz nearby the mouth of the main TL. Is it possible to simulate that with Hornresp or an other application ?
Appologizes for my poor english.
And thanks again for your post
Hervé