Hey so after a couple of months learning hornresp... I've been slapping myself trying to learn how to actually fold said designs
I'm going to stick with FLHs for now. tapped horns in the future.
I've been trying to follow soho54s tutorial (Horn Folding in Sketch-Up), but obviously its a few years old and things are not exactly the same with horn response.
the part that has me particularly confused is the fact that i'm using the first column "length" of the horn export txt to draw out the straight, bottom portion of my horn model. Now... if I measure the centerline length (prefold), its longer than the model... Is this right? or am I missing something completely
Should I be using the column labelled side length for this?
I'm going to stick with FLHs for now. tapped horns in the future.
I've been trying to follow soho54s tutorial (Horn Folding in Sketch-Up), but obviously its a few years old and things are not exactly the same with horn response.
the part that has me particularly confused is the fact that i'm using the first column "length" of the horn export txt to draw out the straight, bottom portion of my horn model. Now... if I measure the centerline length (prefold), its longer than the model... Is this right? or am I missing something completely
Should I be using the column labelled side length for this?
I use the Length and Area columns for folding. (I take the Area of the segment end, divide by the enclosure depth and it gives the width between panels at the segment marker point.)
Centerline folding is not accurate. It's close but the advanced centerline method is better for a few reasons.
Here's a recent thread I started following a week long study of folding with the centerline vs advanced centerline methods. I started with plans for a popular diy flh design, reverse engineered back to Hornresp inputs using a detailed Sketchup drawing, than refolded it using the centerline and advanced centerline methods to see which was more accurate when compared to actual measurements of the physical product. I made some interesting observations. There should be some info in here that can be useful to you. There's huge amount of into there and plenty of pictures showing how to fold using both methods.
Horn Folding - a brief study of the centerline vs advanced centerline method
Centerline folding is not accurate. It's close but the advanced centerline method is better for a few reasons.
Here's a recent thread I started following a week long study of folding with the centerline vs advanced centerline methods. I started with plans for a popular diy flh design, reverse engineered back to Hornresp inputs using a detailed Sketchup drawing, than refolded it using the centerline and advanced centerline methods to see which was more accurate when compared to actual measurements of the physical product. I made some interesting observations. There should be some info in here that can be useful to you. There's huge amount of into there and plenty of pictures showing how to fold using both methods.
Horn Folding - a brief study of the centerline vs advanced centerline method
Last edited:
yes i've read that thread.
I understand advanced centerline vs centerline for measuring an already folded design.
unfortunately where I'm falling down is generating the straight drawing of a particular design... embarrassingly enough. my method was use the length column for drawing the straight portion of the horn, then draw a perpendicular line twice the length of of the width/2 column, and insert a centerline for going around bends.
What confuses me is that this results in a centerline measurement (of the unfolded horn) that is longer than the horn length as per design in horn response (because the centerline is not parallel to the straight side thus will be longer). post horn fold, both the advanced centerline, and centerline methods will measure a horn considerably longer than my hornresp sim 🙁 😕
I understand advanced centerline vs centerline for measuring an already folded design.
unfortunately where I'm falling down is generating the straight drawing of a particular design... embarrassingly enough. my method was use the length column for drawing the straight portion of the horn, then draw a perpendicular line twice the length of of the width/2 column, and insert a centerline for going around bends.
What confuses me is that this results in a centerline measurement (of the unfolded horn) that is longer than the horn length as per design in horn response (because the centerline is not parallel to the straight side thus will be longer). post horn fold, both the advanced centerline, and centerline methods will measure a horn considerably longer than my hornresp sim 🙁 😕
Ok, if I'm reading you right I know the problem. Please excuse the quality of my quick MS Paint illustration.
You are drawing the horn like the top pic, with a flat bottom. And then you are making a centerline through it like the red line. That red line is NOT the centerline length. The bottom flat line is the centerline length.
If you draw it out as a circular cross sectional area that expands in all directions like the bottom pic then the red line line down the center IS the centerline length.
Is that your problem or is it something else?
An externally hosted image should be here but it was not working when we last tested it.
You are drawing the horn like the top pic, with a flat bottom. And then you are making a centerline through it like the red line. That red line is NOT the centerline length. The bottom flat line is the centerline length.
If you draw it out as a circular cross sectional area that expands in all directions like the bottom pic then the red line line down the center IS the centerline length.
Is that your problem or is it something else?
Last edited:
yeah, you hit it on the head basically. so, basically, I'm doing it right and tripping myself thinking about it incorrectly.
If your drawn centerline is coming out longer than Hornresp says it should be then it's not right.
I think my previous post caused a bit of confusion so I deleted it. I think the text in the post with the pics caused some confusion too, so ignore that and let me start again. Ignore everything except the picture and the following explanation.
You can draw the horn as shown in the top picture or as shown in the bottom picture. But no matter how you draw it out, the red line length has to be equal to the centerline length as reported by Hornresp.
If you draw it out like the top picture and you draw the bottom (horizontal) line of the horn to the horn length as reported by Hornresp then the red line is going to be too long. So if drawn like the top picture, the bottom (horizontal) line is going to be shorter than the length as reported by Hornresp.
For some reason this is really hard to explain, I hope this makes it more clear.
You can draw the horn as shown in the top picture or as shown in the bottom picture. But no matter how you draw it out, the red line length has to be equal to the centerline length as reported by Hornresp.
If you draw it out like the top picture and you draw the bottom (horizontal) line of the horn to the horn length as reported by Hornresp then the red line is going to be too long. So if drawn like the top picture, the bottom (horizontal) line is going to be shorter than the length as reported by Hornresp.
For some reason this is really hard to explain, I hope this makes it more clear.
Last edited:
alright. so that answers my initial question exactly.
I've been drawing it like the top picture, asigning the bottom horizontal line the lenght as per horn respn, and then adding in the "red line" afterwards (through the center of my perpindicular slices) resulting in a "too long" red line (just as you pointed out). this is the method outlined in soho's tutorial... which is either no longer valid due to updates to hornresp's export... or I'm following instructions incorrectly.
this explains why its been next to impossible for me to fold ANYTHING.... they never seem to fit in the size enclosure I've assigned them lol...
I've been drawing it like the top picture, asigning the bottom horizontal line the lenght as per horn respn, and then adding in the "red line" afterwards (through the center of my perpindicular slices) resulting in a "too long" red line (just as you pointed out). this is the method outlined in soho's tutorial... which is either no longer valid due to updates to hornresp's export... or I'm following instructions incorrectly.
this explains why its been next to impossible for me to fold ANYTHING.... they never seem to fit in the size enclosure I've assigned them lol...
now the question becomes.... how the hell do I draw it like the top picture, and keep centerline length correct..... or rather... how the hell do I fold a horn if we're tossing that tutorial out.
Back to square one 🙁
Back to square one 🙁
I've never tried folding with Sketchup so take these comments for what they are worth.
I just took a look at the tutorial and it does look like he's using the wrong centerline length, as we just discussed. This can't give a perfectly accurate result but I don't think the centerline length will be more than a few percent off so at worst following the instructions to the letter would tune 1 or 2 hz too low in a worst case scenario. His method does a lot of processing through the corners and might lose a bit of length through the corners, I imagine he was trying to keep the results close to the advanced centerline method, so the overall horn length probably isn't really far off.
The reason you can't get the Hornresp export file into Sketchup properly is the export file is changed. It used to have a width column, now it's width/2. So if you really want to use the Sketchup folding method you can. There's two ways to do that.
1. Find a way to multiply all numbers in the width column by two and then it should work. This does nothing to address the centerline length issue but it will let you proceed with the process exactly as documented in the Sketchup tutorial.
2. Use the width/2 column as it is. Import the results into Sketchup. Now do it again and put the other half of the horn upside down and below the top half. Then it will look like the bottom picture in post 4 instead of the top picture. This will fix the centerline length issue and your finished product should have the correct length.
If you don't want to use Sketchup you can do the brute force method, same as I used to. Start with a blank piece of paper, draw the outside walls and start drawing the horn flare in, segment by segment, trial and error as you go around the bends.
Keep in mind that whether you use the Sketchup folding or the hand drawn method, the fold won't be completely accurate because of the issue described in this post - http://www.diyaudio.com/forums/subwoofers/175658-tham15-compact-15-tapped-horn-21.html#post2451737
It takes a lot of forethought and planning to get around this issue and fold so that the finished product accurately reflects the simulation.
I have a bunch of tricks that I use to make folding exponentially easier. If planned right from the beginning, the design can be simulated and folded at the same time. I can simulate and fold a horn in about an hour using these tricks. I prefer to take more time than that but it is possible. If you are interested I can talk about this more tomorrow.
There are also a whole bunch of different ways to fold horns. Some people use Sketchup in different ways than Soho54, some people use Excel, and there's even a spreadsheet that will auto-fold a horn for you. http://www.diyaudio.com/forums/subwoofers/171747-spreadsheet-folded-horn-layouts.html
You have options. Just pick a method and work it.
I just took a look at the tutorial and it does look like he's using the wrong centerline length, as we just discussed. This can't give a perfectly accurate result but I don't think the centerline length will be more than a few percent off so at worst following the instructions to the letter would tune 1 or 2 hz too low in a worst case scenario. His method does a lot of processing through the corners and might lose a bit of length through the corners, I imagine he was trying to keep the results close to the advanced centerline method, so the overall horn length probably isn't really far off.
The reason you can't get the Hornresp export file into Sketchup properly is the export file is changed. It used to have a width column, now it's width/2. So if you really want to use the Sketchup folding method you can. There's two ways to do that.
1. Find a way to multiply all numbers in the width column by two and then it should work. This does nothing to address the centerline length issue but it will let you proceed with the process exactly as documented in the Sketchup tutorial.
2. Use the width/2 column as it is. Import the results into Sketchup. Now do it again and put the other half of the horn upside down and below the top half. Then it will look like the bottom picture in post 4 instead of the top picture. This will fix the centerline length issue and your finished product should have the correct length.
If you don't want to use Sketchup you can do the brute force method, same as I used to. Start with a blank piece of paper, draw the outside walls and start drawing the horn flare in, segment by segment, trial and error as you go around the bends.
Keep in mind that whether you use the Sketchup folding or the hand drawn method, the fold won't be completely accurate because of the issue described in this post - http://www.diyaudio.com/forums/subwoofers/175658-tham15-compact-15-tapped-horn-21.html#post2451737
It takes a lot of forethought and planning to get around this issue and fold so that the finished product accurately reflects the simulation.
I have a bunch of tricks that I use to make folding exponentially easier. If planned right from the beginning, the design can be simulated and folded at the same time. I can simulate and fold a horn in about an hour using these tricks. I prefer to take more time than that but it is possible. If you are interested I can talk about this more tomorrow.
There are also a whole bunch of different ways to fold horns. Some people use Sketchup in different ways than Soho54, some people use Excel, and there's even a spreadsheet that will auto-fold a horn for you. http://www.diyaudio.com/forums/subwoofers/171747-spreadsheet-folded-horn-layouts.html
You have options. Just pick a method and work it.
Last edited:
I've been doing it via method 1.
for the particular horn I've been trying to fold, this results in a "drawn in" centerline that is about 10cm too long, about 4 to 5 percent over.
I'd like to do it via method 2 (to obtain an accurate centerline), but I dont understand how to terminate the horn at the mouth to keep the bottom and the mouth perpendicular to each other.
for the particular horn I've been trying to fold, this results in a "drawn in" centerline that is about 10cm too long, about 4 to 5 percent over.
I'd like to do it via method 2 (to obtain an accurate centerline), but I dont understand how to terminate the horn at the mouth to keep the bottom and the mouth perpendicular to each other.
If you want to use method 1 you are going to have to live with a centerline that's too long. I don't know any way around that other than by using method 2 or by manually resizing the length of each segment.
That's an issue I didn't think of. But you can just manually redraw it. Make sure the segment has the same cross sectional area at each end and the same centerline length.
I would suggest drawing a horn from scratch for experience. Either by hand or with Sketchup or any other method. The important part is to NOT start with a picture of a horn, start with a blank page and start drawing in segments one by one. If you do this at least once you it will give you experience in dealing with these issues.
The centerline vs advanced centerline issue also makes things confusing, I'd suggest always using the advanced centerline method.
Hey guys, I've also run into the issue regarding the "horizontal line" and the centerline lengths being different. If you draw it using the horizontal line as your length reference, but then proceed to slice and rotate your sections about the centerline, you can calculate the path length error that is introduced in a segment.
For a horn segment of width W, length L, and with start and end areas S1 and S2:
error = sqrt{ [ (S2 - S1)/2W ]^2 + L^2 } - L
In practice, at least for a bass horn segment with a small flare angle, this works out to being negligibly small. Much greater path length error is introduced by our approximations of path length through the bends.
Regards,
Matt
For a horn segment of width W, length L, and with start and end areas S1 and S2:
error = sqrt{ [ (S2 - S1)/2W ]^2 + L^2 } - L
In practice, at least for a bass horn segment with a small flare angle, this works out to being negligibly small. Much greater path length error is introduced by our approximations of path length through the bends.
Regards,
Matt
it does generally seem to be relativly small, except this particular design had a short, large flare at the termination of the horn. horizontal line measured 50cm, fake "centerline" measured 58cm
This is getting way more complex than it needs to be. This is literally all you need to know.
Any given segment must have the start and end cross sectional area that Hornresp says it should have. Then you draw a line down the center of that segment and that should be equal to the length that Hornresp reports for that segment.
The length of the horizontal line (the bottom boundary) makes no difference and it will be different depending on the geometric shape of the segment (how the segment is drawn). The ONLY length that matters is the centerline.
All of this becomes very clear when you draw up a horn fold from scratch. Using Sketchup to fold up a picture of a straight unfolded horn is going to cause problems if you haven't done it from scratch at least once.
There is no error or approximation if you use the advanced centerline folding method. So I'm not sure what you are saying here.
(To be more clear, the advanced centerline method might not be 100 percent accurate but it is by far the most accurate you can get without simulating the air in the bends as mass, resistance and compliance. That adds exponentially more complexity and the only way to do that is with Akabak, but then you introduce errors unless you can actually build conical segments because Akabak can't simulate PAR segments.)
Hi,FYI
Links to 'Calculations of Acoustical behaviour of Bends':
Acoustics of 90 degree sharp bends. Part I: Low-frequency acoustical response:
http://perso.univ-lemans.fr/~yauregan/publi/Bend_I.pdf
Acoustics of 90 degree sharp bends. Part II: Low-frequency aeroacoustical response:
http://perso.univ-lemans.fr/~yauregan/publi/Bend_II.pdf
Acoustic properties of miter bends. John W. Coltman:
https://ccrma.stanford.edu/marl/Coltman/documents/Coltman-1.44.pdf
b🙂
Links to 'Calculations of Acoustical behaviour of Bends':
Acoustics of 90 degree sharp bends. Part I: Low-frequency acoustical response:
http://perso.univ-lemans.fr/~yauregan/publi/Bend_I.pdf
Acoustics of 90 degree sharp bends. Part II: Low-frequency aeroacoustical response:
http://perso.univ-lemans.fr/~yauregan/publi/Bend_II.pdf
Acoustic properties of miter bends. John W. Coltman:
https://ccrma.stanford.edu/marl/Coltman/documents/Coltman-1.44.pdf
b🙂
Last edited:
- Status
- Not open for further replies.