AVS Forum banner
1 - 20 of 26 Posts

·
Registered
Joined
·
8 Posts
Discussion Starter · #1 ·
In the past week or two I’ve been reading all the tapped horn threads, tried to learn how to use Hornresp and designing my own small sub.
All of this is very much inspired by the Insubnia and it has the same "small, low, sacrifice spl" idea behind it.
It's still not out of the question that I just build pair of Insubnias, BUT, for few reasons I’m trying to work out something else as well.

There might be, or dare I say it’s likely that there are, some mistakes and oversights in the build as I’m really out of my depth here.
I’ve only done one proper (set of 5) speakers from a kit and one ”superb” passive sub planned with WinISD. Building them is not an issue, but my lack of knowledge in audio could be.


Anyways, here’s what I got at the moment.

Driver: Tang Band W6-1139SIF



Net volume, according to Hornresp, is 48l. Size of the thing is quite a bit more though, as there’s so much stuff inside. Just as a comparison, more traditional enclosure of this size would be around 58l.
So, there’s around 10l of walls inside.
Walls are 12mm MDF. Probably a good idea to make the driver piece out of plywood.



I tried A FEW different ways to fold it. The way I ended up seemed to get me the least amout of wasted space and the form factor is what I’m after. I wouldn’t mind it having a bit less depth, but I don’t think that’s happening.

Top of the enclosure didn't end up perfectly square, so I added a small piece inside on top of the first fold. Not sure how critical that is, but better safe than sorry.
Rounded the corners a bit for aesthetics.

The driver is quite tight fit. Building it won’t be a problem, but I’m not sure how much space you’d need around it. From the side it looks a bit questionable, but here’s a how it sits in from the top:




There’s room on the sides, but is that enough?


Basics of Hornresp is fairly easy to understand and I found using enjoyable, but my lack of knowledge in audio in general doesn’t help. I have no idea what a lot of the things mean, what ranges are fine for what etc, so there can be some outrageous errors from me.
I’ll include the Hornresp file.

Here’s the parametres:




And the output:





So, what did I mess up? :D

No, but seriously, I really am here to learn, so if you see something, say something.
There’s so much I don’t know and what there is to know.

Every piece of information, links, hints and facepalms will be much appreciated!
 

Attachments

·
Registered
Joined
·
2,469 Posts
That is a lot of corners which will create air resistance that is not in the modeled response. I have not done a deep dive on the Ghorn, but I do remember that being an issue for that design.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #3 ·
Oh, right. I guess I could try to go up and down. I think the parametres have changed a bit (lower S2 etc) from when I did my earlier folds, so maybe it wouldn't make the enclosure too deep.

That would lower the amount of corners from 21 to 13.

What effects does high air resistance produce?

I'll try to look into that and I'll see if I can refold it later tonight.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #4 ·
After some fiddling with the folding, I ended up folding it the original way.

Now it's even better shape. The original parameters made it too deep folded this way, but now it works nicely.





Corners also went from 21 to 11, so that a win an well.

Parameters didn't change considerably. Some minor changes in L lengths.



I found some information on another forum about the air resistance. Not exactly what I was looking for, but I understood the basic idea with air resistance behind and front of the driver.



So, what's the next thing to fix? :D

One thing I'm mostly concerned about is the area around the driver. Is there any rule of thumb how much space you should have? Does the driver restrict too much?
 

·
Premium Member
Joined
·
5,461 Posts
^
this : the LOWARHORN
a refold of the GHorn by @LTD02 and @jpmst3(?)
iirc

25,5 x 30 (31.5?) x 72
it was about reducing the multitude of folds,
there is a Ghorn V2 that has reduced folding, also

refined for a UXL 18"
beastly TSTL

that other is from archived collections but the driver ? -DK

3rd pic is kinda neat
 

Attachments

·
Registered
Joined
·
3,076 Posts
I think the driver is blocking too much of the horn path. Looks like more than 50%. I'm not sure what the acceptable percentage is. Most horns are simply larger so the driver blocking the path isn't usually a big concern. You may be able to find some info on this by searching around. I'm sure it has been discussed many times.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #7 ·
^
the LOWARHORN
a refold of the GHorn
I went through some of the old threads and they did give me some ideas. I had the thought of ending the horn right at the floor like in their design, but initially it didn't work out.
Now that I look at it, I think it could be possible do it and still maintain the desired shape.
Maybe if I start the horn from top to bottom it would take out one whole fold away as well. Although it might become quite tall. Need to try this out.

Didn't find much on reducing the folds, I think it was more about changing the form.


I think the driver is blocking too much of the horn path. Looks like more than 50%. I'm not sure what the acceptable percentage is. Most horns are simply larger so the driver blocking the path isn't usually a big concern. You may be able to find some info on this by searching around. I'm sure it has been discussed many times.
Yeah, I have to see how much more space I can add to it before it starts to change the size and output too much. If I keep the start of the horn close to what it's now, I guess having the horn a bit deeper at the end wouldn't make it too much bigger.
I'll see what I can come up with later tonight.

I would imagine this has come up at some point, but I haven't seen it being discussed. I'll try to see if I find info on it.
 

·
Registered
Joined
·
2,395 Posts
After some fiddling with the folding, I ended up folding it the original way.

Now it's even better shape. The original parameters made it too deep folded this way, but now it works nicely.





Corners also went from 21 to 11, so that a win an well.

Parameters didn't change considerably. Some minor changes in L lengths.



I found some information on another forum about the air resistance. Not exactly what I was looking for, but I understood the basic idea with air resistance behind and front of the driver.



So, what's the next thing to fix? :D

One thing I'm mostly concerned about is the area around the driver. Is there any rule of thumb how much space you should have? Does the driver restrict too much?
Hey - I've seen something that looks sort of like that before.

What's left to fix? Driver access is a huge one. Space around the driver is another. Choking the mouth of the horn with a driver will result in a change in the response. How much? Who knows? That depends on the area of the horn and the area of the driver at that point, as well as the amplitude of the signal being reproduced.

How's the impulse response look? There's a lot more to a good horn design than a flat SPL plot.

Also - you're drawing parabolic flares (two parallel walls), might as well simulate with them. Change your HornResp model to PAR.Won't make much of a difference, but it will make a difference.
 

·
Registered
Joined
·
2,395 Posts
Oh - though I have never built a horn cabinet out of MDF, IMO using 12mm MDF will most likely be a bad idea. I've built plenty of speakers out of 19mm MDF over the years and I do not use it if I have a choice. It is heavy. It is weak. It does not glue all that well. It will split very easily when you fasten into the edges. Small brads (18 gauge) should be OK, but I wouldn't use anything larger. Issue with the 18-gauge brads is having them go where they're pointed when you pull the trigger, they often curve when driven. I have had 18-gauge brads curl right back out the top of MDF and Advantech OSB.

For something like you're working on, using a good 12mm ply will work OK, that's what I used with my small tapped horn builds.
 

·
Premium Member
Joined
·
25,257 Posts
Also - you're drawing parabolic flares (two parallel walls), might as well simulate with them.

Seems to me that flat panels give a linearly increasing area expansion.



MDF is also less stiff than ply.

That and more weight lowers the panel resonance frequencies.
 

·
Registered
Joined
·
2,395 Posts
Someone went through the math either here or over at DIYAudio long ago. Pretty sure it was soho54, but I am not gonna go search for it. I've done the math to prove it to myself and I model with a flare type that is appropriate to what I have drawn. In this case, I would choose parabolic.

A conical flare expands in two dimensions (X and Y) as Z increases. A parabolic flare expands in one dimension (either X or Y) as Z increases.

If the horn has 2 sides that are parallel and the other two expand at a constant linear rate, it is a parabolic flare. If no sides are parallel and each pair of sides expands at a constant linear rate, it is conical.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #14 ·
does one long fold work for that driver?
https://www.cowanaudio.com/th.html
Don't quote me on it, but for what I can tell, yes.
In one iteration I was working on, the L12 and L34 were very close to being the same length. Even in the current one they are somewhat close. It would, however, be very tall, towering over two metres.

For my application I try to keep it under 90cm long and 50cm deep. Although I'm beginning to question my reasoning for that as well. Now I just have to fold 429726 versions more and try to picture them in the room.

I'll give that page a read. Thanks!


Hey - I've seen something that looks sort of like that before.

What's left to fix? Driver access is a huge one. Space around the driver is another. Choking the mouth of the horn with a driver will result in a change in the response. How much? Who knows? That depends on the area of the horn and the area of the driver at that point, as well as the amplitude of the signal being reproduced.

How's the impulse response look? There's a lot more to a good horn design than a flat SPL plot.

Also - you're drawing parabolic flares (two parallel walls), might as well simulate with them. Change your HornResp model to PAR.Won't make much of a difference, but it will make a difference.
Hah, yeah, that one did end up looking like something got squished a bit. Not that that's too surprising as the driver is somewhat similar and if you start folding it like that...

I've sinse gone a bit different direction, not looping around the bottom. Now it ends right at the bottom left corner.
Added in some more clearance for the woofer and also now it doesn't look like a complete copy :p

Fiddling around with the parameters and getting a flat SPL seems fairly trivial if you have the knack and interest for such fiddling, but like I said in the starting post, my knowledge in audio past that point is near zero.
BUT here's for learning.

Good call on the parabolic. It does make a change (for the worse, of course). It's not very intuitive how it differs in practice, even if the reasoning seems solid.

I don't mind building it from ply, that's completely fine by me. I've just kinda used to the idea that (a lot of) speakers are built with MDF and with all the folds I was thinking there's so much gluing surface that it would work well for even a subwoofer, but I'm much more of a cabinet maker than a speaker builder, so plywood it is.




I've been folding and tweaking and folding again while scratching my head until I drew blood. I've already lost count on how many different ways I've tried, but here's the current revision.



There is a slight dead space, but I figured it's actually a nice way to route the wires to the bottom. I was going to add the terminal to the bottom anyways, so that it doesn't matter what orientation the speaker would end up in.
There's now more space around the driver.



Graph looks fairly good to me.



I didn't find the time to read up on the impulse response, so I'm clueless of what it means, but I found it in Hornresp, so here it is. I'll try to read about it tomorrow.
I'll attach the hornresp file while I'm at it.


With the CON to PAR change, the SPL plot did get less flat and it didn't want to go back being flat with some adjustments, so I guess I have to try how it fold conical as well.
So many variables. Well, now there's two good reasons to stay home :D

Thanks for all the replies.
 

Attachments

·
Premium Member
Joined
·
25,257 Posts
Someone went through the math either here or over at DIYAudio long ago...If the horn has 2 sides that are parallel and the other two expand at a constant linear rate, it is a parabolic flare.

If you plot the increase of area vs. length you'll get a straight upsloping line, not a parabola.

So whoever decided to call it a parabolic expansion isn't very good at math and/or English.

Perhaps it's that the rate of linear expansion is increased with successive folds, which you could call piecewise parabolic.
 

·
Registered
Joined
·
2,395 Posts
If you plot the increase of area vs. length you'll get a straight upsloping line, not a parabola.

So whoever decided to call it a parabolic expansion isn't very good at math and/or English.

Perhaps it's that the rate of linear expansion is increased with successive folds, which you could call piecewise parabolic.
100% agreed that the terminology is unclear. As I understand it - a horn with two parallel sides and a linear expansion in a single plane is referred to as a parabolic flare. This is consistent with what I have read and can find with a quick google search.

"A Par flare is used in the simulations because the cross-sectional area of an axisymmetric Par profile horn changes at a linear rate, the same as for a rectangular cross-section horn having two parallel straight sides and two sloping straight sides.

(The cross-sectional area of an axisymmetric Con profile horn changes at a non-linear rate)."

From McBean over at DIY
 

·
Premium Member
Joined
·
25,257 Posts
Citing references doesn't make it any more correct; it could hardly be more nonsensical than to say "A Par flare is used in the simulations because the cross-sectional area of an axisymmetric Par profile horn changes at a linear rate."

As a point of interest, a conical horn does have a parabolic expansion rate.


"A Par flare is used in the simulations because the cross-sectional area of an axisymmetric Par profile horn changes at a linear rate, the same as for a rectangular cross-section horn having two parallel straight sides and two sloping straight sides.

(The cross-sectional area of an axisymmetric Con profile horn changes at a non-linear rate)."

From McBean over at DIY
 

·
Registered
Joined
·
2,395 Posts
Sorry if my tone came across wrong. Last thing I am trying to do here is pick a fight. I'm no acoustic engineer, just an amateur that has learned a thing or two along the way.

I didn't invent the term, I just learned that a PAR expansion in Hornresp is represented by two parallel walls and two walls that diverge linearly. Since that is what I wind up drawing in nearly all of my subs, I simulate them as a PAR expansion.

When I read the HornResp help file and Kolbrek's papers, I saw definitions that matched my understanding, so I shared it here.

Perhaps something isn't being communicated effectively, posts on forums have a way of leaving the non-verbal parts out. Last thing I intended to do is pick a fight.
 

·
Registered
Joined
·
2,269 Posts
Citing references doesn't make it any more correct; it could hardly be more nonsensical than to say "A Par flare is used in the simulations because the cross-sectional area of an axisymmetric Par profile horn changes at a linear rate."

As a point of interest, a conical horn does have a parabolic expansion rate.
@lilmike

I’m bored/lazy today, so doing a bit of research on this one. I finally found a simple answer I can understand. Bonus points for not even needing to get into much calculation.

Posted by TimLewis on diyaudio.

TimLewis said:
'Par' flare = parabolic, the horn width is proportional to an expression involving the square root of x. The cross sectional area (pi * horn radius squared) is a linear equation involving x. It is a parabloid turned on its side.

'Con' flare = conical, the walls increase linearly, so the cross sectional area increases in proportion with x^2.

If you're building a bass horn with two parallel walls and straight expanding sections, then two walls increase linearly and two walls stay the same width. This means that the cross sectional area increases linearly so you want the Parabolic flare.

The easiest way to visualise it is the difference between a section of a cylinder (a slice of cake) and a solid angle section of a sphere.
So the definitions are based on a a horn with a circular cross-section. The linear increase of area results in a parabolic horn profile.

Chris
 
1 - 20 of 26 Posts
Top