SubMaximus -- A Large Front-Loaded Horn for UXL18 and Stereo Integrity HT 18" - AVS Forum
Forum Jump: 
 14Likes
Reply
 
Thread Tools
post #1 of 485 Old 08-07-2013, 09:45 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
Big changes made. Information from a variety of posts is now summarized here.

well, submaximus has taken a turn...hopefully for the better. after farting around with this for several hours, a pretty simple plan emerged. i was shooting for something that would be simple and doable, yet be big and have big performance. it is still quite large at 2x4x4.75 feet, but not too many internal panels and 3 of the 6 panels are at simple 90 degree angles to the walls. bracing of course would be required and is not shown. the corner deflections would be a good idea for the first few turns from the mouth into the horn, but it is tough to say how much difference they will make after the first few.

there are at least 2 folks who have expressed interest in this one, which is why i am bringing it back from the dead.

this horn was optimized for the uxl18 driver, so it probably works equally well with the lms ultra (but we can check that), and it works pretty good for the stereo integrity ht18. i haven't checked the um18, but that driver probably requires something a little different. 2200 watts takes the driver right to 30mm at the first excursion peak above the low corner. for the si18ht, 1100 watts takes it to 23mm at the same place. destiny must have had a hand in this one. :-)

it is an exponential horn with a straight last section (no expansion). this is one way to bring up the bottom end by a few db vs. a full exponential horn.

excursion minimum occurs at around 18hz and the -3db off the corner point is about 15hz, so i'm not quite sure what to call it...maybe a 16hz horn?

one interesting thing about this design is it holds response pretty well down to around 30hz, which is where many rooms tends to start benefitting from frequency dependent room gain. so in an average room the rolloff down to 16hz or so might actually be pretty flat. won't know until it is built and measured.

consider this a "first draft" and not a final plan. if anybody can double check it as much as possible, i'd appreciate that.

for those who aren't familiar with how the hornresp models translate to actual performance, the peaks in the response tend to not be quite as pronounced in the actual results as they are in the model.

so without further delay...re-introducing SUBMAXIMUS!

response, 2200 watts into 4 ohms, 2 pi space 1m (smushed):


fold/plan:
the idea with the fold plan is to "section" off the expansion into 4 flare rates. with 4 rates calculated to provide an exponential flare rate, changing the hornresp model to "par" will provide the same results as "exp". of course, we are one leg short in hornresp, so i'll have to manually calculate the appropriate expansion for leg 1 (purple). otherwise, we should be on target. 7.4 cm appears to be the cross section with an exponential flare rate at that point. L23 is a combination of the purple and the blue.


hornresp numbers:


response 1w1m 2 pi space:


response with 1100 watts into 4 ohm SI18HT driver 2 pi space 1m:


excursion uxl, 2200 watts into 4 ohms:


excursion si 8ht, 1100 watts into 4 ohms:












here is the submaximus horn with uxl driver vs. the uxl driver in a sealed enclosure of about 6 cubic feet.



at 16hz, the low corner of the horn, the horn has about 15db more sensitivity than the sealed enclosure. at the upper end of the bass, the gain is about 8db sensitivity.

here is the sealed enclosure max spl at 30mm and/or 2200 watts.



at 16hz, the max spl is about 107db. submaximus with uxl peaks at about 124db at the same frequency.








compression ratio is 3.4:1. my guess is that this is fine, but would like input there.

also, actual width is about 25.5 inches depending on material thickness. the 4x4.75 are exact outter dimensions. the 2 feet width is internal. bracing is not shown, but is REQUIRED!!








final scale drawing:


internal measurements and color coded boards to see how the butt joints work:








one method for arranging the panels on 4x8 stock. extra for bracing would be required. the corner deflectors are not included in the cut plan.




bracing is not shown either. the particular method for bracing doesn't matter as much as that it is well braced. this guy is going to generate a lot of pressure, so brace it very well!

the final model came out about 1db lower in sensitivity with 0.75" plywood. not different enough to repost all the plots. if really well braced, thinner stock can be used, but the cut plan is for 0.75", so keep that in mind. a little adjustment here and there would be require if using half inch for example.

the side panels completely enclose the other panels. in other words, start with a side panel on the floor. draw the pattern on it. place the panels on the side. when all panels and braces are in place, the other side with be placed on top to close it up.

there are only three panels on an angle. all other panels are right angles.

(c) 2014 LTD02. All Rights Reserved.

if you want to use this plan for a commercial design, please drop me a note first. otherwise...have at it!





comparison of full martysub with 18ho driver 1100 watts (red), submaximus loaded with 18ho driver 1100 watts (black), and submaximus loaded with uxl18 driver 2200 watts (blue). the effect of the stronger motor on the uxl can be seen in not just providing a higher total response, but also a different shape. the 18ho loaded submaximus looks pretty good though too... :-)





Post mortem analysis, conclusions, final comments:


SubMaximus -- A Large Front-Loaded Horn for UXL18 and Stereo Integrity HT 18"










OLD & OUTDATED

Original post for posterity:

As the title says, this is a first rough draft of a large front-loaded horn for the Stereo Integrity HT 18" driver. It seems to be a pretty darn good value, so I was curious to see what it could do in a large horn.

This one is 4 feet tall, 5 feet deep, and 2 feet wide (similar in size to the one that Myggpower's buddy in Germany built a while back second pic). It appears to work well. With 1100 watts power, excursion is in check everywhere except just below tuning.

The low corner is around 16hz or so. Sensitivity is pretty good at 95db 1w-1m-2pi space at 32hz and still at 90db 1w-1m-2pi space at 16hz.

Bracing isn't shown. Interior fold isn't too complicated either.

Frequency response would need a little e.q. to take care of a couple of upper bass peaks, but no big deal there.

With typical room gain kicking in around 30hz, this one should be pretty flat down to 15hz or so where a highpass filter would be advised around 15hz or so (there is a short note on how to do this with the iNuke DSP or the DCX2496).

Two of them should be able to provide upwards of 130db bass down to 16hz or so in even a largish room. Not bad for two drivers that cost a couple hundred bucks each, a cheap Behringer amp, and some wood.

Anyway, here are some notes. This is a first draft and as always there may be errors, so comments/criticisms are welcome.

I guess that I have to give it a name...so this one shall be called SubMaximus!














Listen. It's All Good.

Last edited by LTD02; Yesterday at 05:16 PM.
LTD02 is offline  
Sponsored Links
Advertisement
 
post #2 of 485 Old 08-07-2013, 10:13 PM
AVS Special Member
 
lilmike's Avatar
 
Join Date: Dec 2009
Location: Pacific Northwest
Posts: 2,167
Mentioned: 0 Post(s)
Tagged: 1 Thread(s)
Quoted: 4 Post(s)
Liked: 78
40 cubes...

That's large, no doubt.

Why the long L12?
lilmike is offline  
post #3 of 485 Old 08-07-2013, 10:50 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
it was actually to make the driver-to-mouth distance a little shorter (i had started with the driver all the way back, but the horn was 1-2 feet too long), to increase the area behind the driver by a smidgen, to keep the build fairly simple (instead of making another tiny panel right behind the driver), and iirc the frequency response actually improved by a hair (which was surprising).

edit: i just checked and in this case the long 1-2 actually smooths out the upper end response by 1/2 db here and there, but the main reason was so not to have to put another panel right behind the driver as that would be a minor complicating factor.

Listen. It's All Good.
LTD02 is offline  
post #4 of 485 Old 08-07-2013, 10:53 PM
AVS Special Member
 
lilmike's Avatar
 
Join Date: Dec 2009
Location: Pacific Northwest
Posts: 2,167
Mentioned: 0 Post(s)
Tagged: 1 Thread(s)
Quoted: 4 Post(s)
Liked: 78
Just curious was all.

The impulse and acoustic impedance plots look OK?
lilmike is offline  
post #5 of 485 Old 08-07-2013, 11:18 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
i have no idea how to interpret their practical significance. for example, once the thing is high-passed, low-passed, and a few eq filters are applied, i'm not sure the raw impulse response any longer holds much value, but i could be way off on that. same thing with the acoustical impedance, i'm not sure what the practical ramifications are of the data. i don't know if this is good, bad, or if it matters at all. you?


Listen. It's All Good.
LTD02 is offline  
post #6 of 485 Old 08-08-2013, 12:36 AM
Bass Enabler
 
Scott Simonian's Avatar
 
Join Date: Oct 2001
Location: Clovis, CA
Posts: 13,224
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 160 Post(s)
Liked: 641
Pretty sweet design, John! I like. Looks like a fun project too. Real nice extension on these guys but they are hhheeeeyyyyuuuuuggggeeeeee! tongue.gif

SubMaximus indeed. smile.gif

My Dual 18" LLT subs 120dB down to 10hz

 

Plan9Reloaded Co-host

Listen to the Plan9Reloaded Gaming and Technology Podcast (may contain NSFW language)

https://soundcloud.com/plan9reloaded/sets/podcast - direct pod link

http://plan9reloaded.com/site/ - main website

Scott Simonian is offline  
post #7 of 485 Old 08-08-2013, 04:15 AM
AVS Special Member
 
Martycool007's Avatar
 
Join Date: Jun 2011
Posts: 3,409
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 106 Post(s)
Liked: 144
Thanks for posting this LTD02! Can you tell me how low this sub will go, without factoring in room gain? I have always steered away from horn subs in the past due to the fact that most of them do not go as low as a typical sealed or ported enclosure, although, I have not had the chance to listen to a horn sub with a very capable 18" driver either!

What amp or how much power would this horn sub need to use to be effective with both maximum extension and output? I know that horn Subs are generally much more efficient than a sealed or ported cabinet.

How would this horn sub, with the SI18, compare to a sealed or vented enclosure with the same driver?
Martycool007 is online now  
post #8 of 485 Old 08-08-2013, 08:39 AM
Member
 
BP1Fanatic's Avatar
 
Join Date: Jan 2010
Location: Columbus, Oh
Posts: 35
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by Martycool007 View Post

Can you tell me how low this sub will go, without factoring in room gain?

"The low corner is around 16hz or so. Sensitivity is pretty good at 95db 1w-1m-2pi space at 32hz and still at 90db 1w-1m-2pi space at 16hz."

What amp or how much power would this horn sub need to use to be effective with both maximum extension and output?

"With 1100 watts power, excursion is in check everywhere except just below tuning."
62.05 volts = 1100 watts at 4 ohms.

How would this horn sub, with the SI18, compare to a sealed or vented enclosure with the same driver?

Use LTD02's HR inputs minus the S1-S5 and L12-L45. For Sealed, increase Vrc until you can get a F3 of 16hz. For vented, you will have to play with Ap1 and Lpt along with Vrc until you get to F3. Also, change the red OD to Nd.
BP1Fanatic is offline  
post #9 of 485 Old 08-08-2013, 08:52 AM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
thanks scott.

......................

marty,

the frequency responses that i posted are for 2pi space, which is essentially how they will perform outside in the middle of a parking lot (i.e. no room gain).

i would call the low corner on this one around 17hz or so. if you limit power to 3db less than what is shown, you can probably get away without using a highpass filter, but i would error on the safe side and use a high pass filter. and remember, it isn't just how low a sub will play, it is also how loud at that point which is important. this one will do about 122db at 17hz in 2pi space, which is equivalent to about SIX of the same drivers (with six times as much power) in a sealed enclosure of the same size.

as for an amp, the inuke3000dsp will put out around 2200 watts into 4 ohms bridged, which would be fine (using the limiter to about 1100 watts, 62 volts or so). or one channel of an inuke6000dsp could offer the same. if you really wanted to stretch it, you could get the dual 4 ohm coil drivers, wire them in parallel for 2 ohms and run one on each channel of an inuke3000dsp, but not overtaxing the amps is always better.

Listen. It's All Good.
LTD02 is offline  
post #10 of 485 Old 08-08-2013, 09:14 AM
Advanced Member
 
MemX's Avatar
 
Join Date: Oct 2012
Posts: 650
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 45
Quote:
Originally Posted by LTD02 View Post

thanks scott.

......................

marty,

the frequency responses that i posted are for 2pi space, which is essentially how they will perform outside in the middle of a parking lot (i.e. no room gain).

i would call the low corner on this one around 17hz or so. if you limit power to 3db less than what is shown, you can probably get away without using a highpass filter, but i would error on the safe side and use a high pass filter. and remember, it isn't just how low a sub will play, it is also how loud at that point which is important. this one will do about 122db at 17hz in 2pi space, which is equivalent to about SIX of the same drivers (with six times as much power) in a sealed enclosure of the same size.

as for an amp, the inuke3000dsp will put out around 2200 watts into 4 ohms bridged, which would be fine (using the limiter to about 1100 watts, 62 volts or so). or one channel of an inuke6000dsp could offer the same. if you really wanted to stretch it, you could get the dual 4 ohm coil drivers, wire them in parallel for 2 ohms and run one on each channel of an inuke3000dsp, but not overtaxing the amps is always better.
That is quite impressive!

My only fear with ported is bottoming the drivers out, even with all the electronic gizmos that can limit power etc!
MemX is offline  
post #11 of 485 Old 08-08-2013, 11:11 AM
AVS Special Member
 
lilmike's Avatar
 
Join Date: Dec 2009
Location: Pacific Northwest
Posts: 2,167
Mentioned: 0 Post(s)
Tagged: 1 Thread(s)
Quoted: 4 Post(s)
Liked: 78
Quote:
Originally Posted by LTD02 View Post

i have no idea how to interpret their practical significance. for example, once the thing is high-passed, low-passed, and a few eq filters are applied, i'm not sure the raw impulse response any longer holds much value, but i could be way off on that. same thing with the acoustical impedance, i'm not sure what the practical ramifications are of the data. i don't know if this is good, bad, or if it matters at all. you?


Oh, it matters... All of the graphs in HornResp matter, otherwise McBean wouldn't have bothered to program the math to create them. There are certainly degrees of concern though.

On the Acoustic Impedance plot - you're OK, I'd try to tame that first spike a touch personally. So long as the plot doesn't look like a phone pole in a flat field, you're on the right track, but generally speaking, the more even in height the peaks are, the better. Way back when, soho54 tried to explain the meaning of this plot to me, and despite his best efforts, as well as the efforts of others along the way, I'm not sure I really get it. When it comes to design, the thing I try to do is match the opposing peak heights, as well as make them all as short as I can, while maintaining the response I want in a volume I can live with. Easy...right?

As far as impulse? It's basically a graphical representation of the damping of the horn. In my experience, when the plot is a "cleaner" impulse, the horn sounds better. Any anomalies won't be evident in the frequency domain, but the ringing should be evident in plots that show a time-domain response. Of course - it all sorta goes out the window once the horn is folded anyhow, then throw in the typical acoustic processing....and park it in a room. Might as well try to get the horn as right as you can though.
lilmike is offline  
post #12 of 485 Old 08-08-2013, 12:17 PM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 935
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 16 Post(s)
Liked: 37
Quote:
Originally Posted by lilmike View Post

Oh, it matters... All of the graphs in HornResp matter, otherwise McBean wouldn't have bothered to program the math to create them. There are certainly degrees of concern though.

On the Acoustic Impedance plot - you're OK, I'd try to tame that first spike a touch personally. So long as the plot doesn't look like a phone pole in a flat field, you're on the right track, but generally speaking, the more even in height the peaks are, the better. Way back when, soho54 tried to explain the meaning of this plot to me, and despite his best efforts, as well as the efforts of others along the way, I'm not sure I really get it. When it comes to design, the thing I try to do is match the opposing peak heights, as well as make them all as short as I can, while maintaining the response I want in a volume I can live with. Easy...right?

How can you recommend what the plot should look like if you don't understand what the plot is showing? Not trying to be mean or anything....just found it to be a really odd recommendation.

My understanding is that if you really want the holy grail of acoustic impedance, then you want the acoustic impedance to essentially match the electrical impedance of the driver - which also means you really don't want any peaks anywhere. The peaks represent reflections happening due to the acoustic mismatch at the mouth (since the mouth is undersized). For a bass horn, it's kinda hard to avoid the reality of these reflections/resonances because the wavelengths are so huge and dictate an insanely large mouth - which nobody has the room for.

Also keep in mind that Hornresp adds a multiplier to the acoustic impedance chart - and that multiplier changes with minor changes to the horn....so you need to be real careful to convert to absolute values when making comparisons.
Quote:
Originally Posted by lilmike View Post

As far as impulse? It's basically a graphical representation of the damping of the horn. In my experience, when the plot is a "cleaner" impulse, the horn sounds better. Any anomalies won't be evident in the frequency domain, but the ringing should be evident in plots that show a time-domain response. Of course - it all sorta goes out the window once the horn is folded anyhow, then throw in the typical acoustic processing....and park it in a room. Might as well try to get the horn as right as you can though.

The impulse response shown in Hornresp can be obtained by convolving the frequency response with an ideal impulse ....which is to say that the frequency and time domains shown in Hornresp are intrinsically related. I bring it up because as you understand this relationship, you can start seeing the effects of the time domain while looking at the frequency domain (and vice versa) - and I think there are some insights to be gained there when designing a horn. And if you want to simplify things a bit, the tangent angle of the mouth flare offers some great insights depending on the expansion rate of the horn...

For instance, that high frequency ripple in this first simulation is showing the ringing present in the impulse, and the acoustical impedance plots are showing you why the amplitudes are different at different frequencies - etc...

LTD...what are your goals for this design? And are you willing to play a few acoustical tricks to smooth out the response at 80Hz and above?
lilmike likes this.

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #13 of 485 Old 08-08-2013, 03:13 PM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,330
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1226
Quote:
Originally Posted by LTD02 View Post

i'm not sure what the practical ramifications are of the data. i don't know if this is good, bad, or if it matters at all. you?
Use the System Design tool with the driver specs to come up with the 'perfect' design. Based on what that tells you get as close as possible to those plots within the confines of what your allotted cabinet size will allow.
Mfusick and MemX like this.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is online now  
post #14 of 485 Old 08-08-2013, 04:09 PM
AVS Special Member
 
lilmike's Avatar
 
Join Date: Dec 2009
Location: Pacific Northwest
Posts: 2,167
Mentioned: 0 Post(s)
Tagged: 1 Thread(s)
Quoted: 4 Post(s)
Liked: 78
Quote:
Originally Posted by MBentz View Post

How can you recommend what the plot should look like if you don't understand what the plot is showing? Not trying to be mean or anything....just found it to be a really odd recommendation.\

Cause that was what I was told to do...

Seriously - I know enough about the subject to know I don't know enough to write a treatise on it on an open Internet forum, so I kept it short.
Quote:
Originally Posted by MBentz View Post

My understanding is that if you really want the holy grail of acoustic impedance, then you want the acoustic impedance to essentially match the electrical impedance of the driver - which also means you really don't want any peaks anywhere. The peaks represent reflections happening due to the acoustic mismatch at the mouth (since the mouth is undersized). For a bass horn, it's kinda hard to avoid the reality of these reflections/resonances because the wavelengths are so huge and dictate an insanely large mouth - which nobody has the room for.

Thanks for the added information.
Quote:
Originally Posted by MBentz View Post

Also keep in mind that Hornresp adds a multiplier to the acoustic impedance chart - and that multiplier changes with minor changes to the horn....so you need to be real careful to convert to absolute values when making comparisons.

I'm very aware of this, and I glossed right over it, thanks for pointing it out.
Quote:
Originally Posted by MBentz View Post

The impulse response shown in Hornresp can be obtained by convolving the frequency response with an ideal impulse ....which is to say that the frequency and time domains shown in Hornresp are intrinsically related. I bring it up because as you understand this relationship, you can start seeing the effects of the time domain while looking at the frequency domain (and vice versa) - and I think there are some insights to be gained there when designing a horn. And if you want to simplify things a bit, the tangent angle of the mouth flare offers some great insights depending on the expansion rate of the horn...

For instance, that high frequency ripple in this first simulation is showing the ringing present in the impulse, and the acoustical impedance plots are showing you why the amplitudes are different at different frequencies - etc...

Thanks for the insights here too. As I said earlier, I know enough to know I don't know, so I don't want to mislead people.
Quote:
Originally Posted by MBentz View Post

LTD...what are your goals for this design? And are you willing to play a few acoustical tricks to smooth out the response at 80Hz and above?
lilmike is offline  
post #15 of 485 Old 08-08-2013, 08:23 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
i'm having a difficult time understanding the practical value of acoustic impedance data. other than being another way to view efficiency, what is it telling us?

a tht-like horn loaded with the dvc driver has a similarly high peak (about 12 acoustic ohms). is that good or bad and why?


Listen. It's All Good.
LTD02 is offline  
post #16 of 485 Old 08-09-2013, 05:10 AM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,330
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1226
Quote:
Originally Posted by LTD02 View Post

a tht-like horn loaded with the dvc driver has a similarly high peak (about 12 acoustic ohms). is that good or bad and why?

It's good. Horns work by presenting the driver with a higher impedance load than free air. Compare the impedance of a horn to the same driver in a sealed cab.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is online now  
post #17 of 485 Old 08-09-2013, 09:36 AM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
Bill, that is a plot of the acoustical impedance. I believe what you are referring to is the electrical impedance, where the peaks are correlated with sensitivity maxima.

The higher the 'acoustical' impedance, the 'lower' the efficiency of the system. That is what Mike was talking about with respect to them being near zero in a horn that is sufficiently large to maximize electrical to acoustical energy transfer. Bass horns are compromised, so that acoustic impedance heads away from zero because the energy transfer is increasingly inefficient the smaller the horn and the lower the corner frequency. What is unclear is how to know at what point a horn is good, average, or poor on that dimension or if it is simply another way of saying that if you want a bass horn that extends quite deep, it isn't going to be super efficient.

One additional question that I have for the experts is what is determining the locations of the resonances in a bass horn. For a sealed driver, I understand that there is a single resonant frequency and it is at that point that electrical impedance peaks, efficiency is maximized, and current through the driver is minimized.

For a horn, there are several electrical impedance peaks, where efficiency is maximized, and current through the driver is minimized, but I don't seem to be able to correlate them to multiples of a wavelength relative to the length of the horn, which is what I thought would be determining their frequencies.

The same question applies, I suppose, to why horns have several points where driver excursion dips to almost zero. That implies several resonant locations, but again, how to calculate them and where they will occur? They don't seem to be at logical distances relative to horn length and they don't seem to be multiples of each other (which is what I would have expected).

Listen. It's All Good.
LTD02 is offline  
post #18 of 485 Old 08-09-2013, 10:34 AM
AVS Special Member
 
Bill Fitzmaurice's Avatar
 
Join Date: Nov 2009
Location: New Hampshire
Posts: 9,330
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 1226
Quote:
Originally Posted by LTD02 View Post

Bill, that is a plot of the acoustical impedance.
I know. The alteration in the electrical impedance is partially attributed to the acoustic impedance. Do as I said, model a horn, then change all the horn parameters to zero, leaving you with a model of the same driver in a sealed box. Use the Compare Previous tool to compare the two acoustical and electrical impedance plots, as well as response. They're all tied together.
Quote:
The higher the 'acoustical' impedance, the 'lower' the efficiency of the system. That is what Mike was talking about with respect to them being near zero in a horn that is sufficiently large to maximize electrical to acoustical energy transfer
It's the other way around. Compare the acoustical impedance of a perfect horn to one with the mouth reduced in area by 75%. You'll see that the smaller horn has higher impedance peaks, but the peaks are much narrower. The peaks of the full sized horn are lower, but wider. The maximum impedance of the smaller horn is higher, but not the broadband average. The greater uniformity of the impedance with the larger horn translates to greater uniformity of response as well.
As for what acoustical impedance represents, it's the load the driver operates into. Just as a high impedance electrical source does not work well into a low impedance load neither does a driver. Ideally an 8 ohm driver would like to see at least an 80 ohm air load. The fact that sees a far lower load impedance than it's electrical impedance is why the transfer efficiency from electrical to acoustical watts is so poor.

Bill Fitzmaurice Loudspeaker Design

The Laws of Physics aren't swayed by opinion.
Bill Fitzmaurice is online now  
post #19 of 485 Old 08-09-2013, 10:48 AM
Senior Member
 
Stumbo's Avatar
 
Join Date: Aug 2003
Location: Bakersfield, CA
Posts: 479
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 40
It's been like 2 days, why has nobody built one yet?! And where's the low profile version?

Less data, more wank!
Stumbo is offline  
post #20 of 485 Old 08-09-2013, 11:55 AM
Advanced Member
 
MemX's Avatar
 
Join Date: Oct 2012
Posts: 650
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 45
Quote:
Originally Posted by Stumbo View Post

It's been like 2 days, why has nobody built one yet?! And where's the low profile version?

Less data, more wank!
Er... how would more of the latter get anything built? I would have thought it would have had the opposite effect...


wink.gif

LOL
MemX is offline  
post #21 of 485 Old 08-09-2013, 12:38 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
"It's the other way around. Compare the acoustical impedance of a perfect horn to one with the mouth reduced in area by 75%. You'll see that the smaller horn has higher impedance peaks, but the peaks are much narrower."

It's the other way around for the acoustic impedance peaks. In an ideal horn, the acoustic impedance peaks are near zero just like Mike said.

"As for what acoustical impedance represents, it's the load the driver operates into. Just as a high impedance electrical source does not work well into a low impedance load neither does a driver. Ideally an 8 ohm driver would like to see at least an 80 ohm air load. The fact that sees a far lower load impedance than it's electrical impedance is why the transfer efficiency from electrical to acoustical watts is so poor."

I think you have it backwards. When the acoustic impedance is low, energy transfer is high, efficiency is maximized. This is why the peaks get smaller as you go up in frequency for a horn with a rising response. As you approach 106db 1w1m, the acoustic impedance peaks are effectively zero.

Listen. It's All Good.
LTD02 is offline  
post #22 of 485 Old 08-12-2013, 04:20 PM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 935
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 16 Post(s)
Liked: 37
Quote:
Originally Posted by LTD02 View Post

The higher the 'acoustical' impedance, the 'lower' the efficiency of the system.

It's more complicated than that....but typically, we have a hard time getting the acoustic impedance high enough to begin with, so what you said is generally backwards. In other words, a higher acoustical impedance usually results in higher efficiency.

Maximum power transfer happens when the output impedance matches the load impedance and our efficiency is 50%. In this case, we have the electro-mechanical output impedance of the driver driving the acoustical impedance of the horn:
http://en.wikipedia.org/wiki/Impedance_matching

Technically, true efficiency can be higher than 50%, but in doing so, you actually decrease the maximum acoustic power transfer of the system:
http://en.wikipedia.org/wiki/Maximum_power_transfer_theorem
In other words, increasing efficiency above 50% only benefits you if you're not power limited on the driver....it's only at the lowest frequencies that we aren't power limited (we're excursion limited), but I don't know of many low frequency designs that even offer >50% efficiency anyway. It's an interesting thought to think about though.

All that to say, the ideal target is usually 50% efficiency.


Richard Small (from Thiele/Small parameters) wrote a good article on direct radiators a long time ago:
http://documents.jordan-usa.com/Famous-Articles/Small-Direct-Radiator-Loudspeaker-System-Analysis.pdf
In there you can see the electro-mechanical representation of the driver output impedance....the horn is basically increasing the radiation resistance as we can see in this article:
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf
The idea of acoustic impedance is important, because it's the ratio between sound pressure and volume velocity (driver excursion at a given frequency):
http://en.wikipedia.org/wiki/Acoustic_impedance


In a loudspeaker (especially at low frequencies), we have a lot of control over the excursion of the driver - which means we have a lot of control over the volume velocity. However, our ears hear sound pressure, so the ability of the system to create pressure is generally limited by the radiation impedance of a finite-excursion system. But as we raise the radiation impedance, we reduce our "control" over the volume velocity - which is why the concept of impedance matching is so important. There will be a point where increasing radiation impedance actually reduces the sound pressure because the motor is no longer strong enough to create the same excursion (if the air creates the same force as a person sitting on the driver, the driver ain't gonna move...). The only thing we can do at that point is throw more power into the driver (according to the maximum power transfer theorem).


With all of that in mind, the 'ideal bass horn' will present just enough radiation impedance to where it takes the maximum power rating of the driver to achieve its maximum excursion....and then the horn makes sure that volume velocity is maintained throughout the length of the horn. The latter part is important because this ensures the volume velocity at the throat is the same as at the mouth - which means you have a much larger surface area providing the same motion. Basically, you're making the diaphragm of the driver larger without increasing the moving mass nor creating new cone stiffness issues...etc.

When the mouth is undersized, you move into the realm of a resonant system - and now you're including phase angles to the impedance matching. This is important because the phase changes very quickly with frequency - so now you're limiting the bandwidth over which the horn is increasing radiation impedance....basically, every octave above the Fc of the horn is seeing an increase in radiation impedance, and then between those octaves you're seeing little to no increase - which is why you see so many spikes in the acoustic impedance of these horns. And if these spikes are too large, then you fall on the other end of the acoustic impedance to where you don't have enough motor to hit the maximum excursion of the driver.....and the only way to "reactance annul" the driver is to create a counter-resonance using the rear volume behind the driver so that you're not relying on the motor to fight the wavefront.



If you're following all of that, then you will see that a highly reactive horn starts to have a limited benefit when compared against a direct radiator....and you will also see why certain drivers lend themselves to horn-loading better than other drivers (they have more motor available to overcome the increased radiation impedance of the horn). Going reactive is fine, but I personally always start with the amount of impedance I'm looking for - and then limit the design of the horn to provide that. You'll have a hard time maximizing output by just looking at the frequency response and tweaking input parameters until it looks flat...
MemX likes this.

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #23 of 485 Old 08-12-2013, 04:42 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
"It's more complicated than that....but typically, we have a hard time getting the acoustic impedance high enough to begin with, so what you said is generally backwards. In other words, a higher acoustical impedance usually results in higher efficiency."

then something is wrong with hornresp because the higher efficiency always corresponds with a lower acoustic impedance.

here is the horn efficiency. each dip in gets less and less efficient the lower in frequency that one goes and the acoustic impedance gets higher and higher.




Listen. It's All Good.
LTD02 is offline  
post #24 of 485 Old 08-12-2013, 04:48 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
here is the efficiency and acoustic impedance for a near perfect midbass horn using the 2226 driver and operating at about 105db from 80 to 250hz.

acoustic impedance is down near zero and efficiency is about as high as one can get.

[remember to apply the multiplier of .04]




Listen. It's All Good.
LTD02 is offline  
post #25 of 485 Old 08-12-2013, 05:58 PM - Thread Starter
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 15,977
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 136 Post(s)
Liked: 856
Apparently, Hornresp is using the definition of acoustic impedance where:

Z = p/vS

sound pressure p divided by the particle velocity v and the surface area S.

http://en.wikipedia.org/wiki/Acoustic_impedance

It can be seen in the first plot, cone motion minima occur at 20hz, 40hz, etc...

As a result, particle velocity minima also occur at 20hz, 40hz, etc...

Since surface area is unchanging, the acoustic impedance peaks also occur at 20hz, 40hz, etc...





I'm not sure what all this is telling us...

Listen. It's All Good.
LTD02 is offline  
post #26 of 485 Old 08-13-2013, 07:28 AM
Advanced Member
 
MemX's Avatar
 
Join Date: Oct 2012
Posts: 650
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 45
This thread makes my hurt... lol

Quote:
Originally Posted by MBentz View Post

With all of that in mind, the 'ideal bass horn' will present just enough radiation impedance to where it takes the maximum power rating of the driver to achieve its maximum excursion....and then the horn makes sure that volume velocity is maintained throughout the length of the horn. The latter part is important because this ensures the volume velocity at the throat is the same as at the mouth - which means you have a much larger surface area providing the same motion. Basically, you're making the diaphragm of the driver larger without increasing the moving mass nor creating new cone stiffness issues...etc.

When the mouth is undersized, you move into the realm of a resonant system - and now you're including phase angles to the impedance matching
Is this saying that a Horn is designed to allow the air/soundwaves from the front of the driver to expand out at a constant rate and velocity, the end result being that when it exits the cabinet the air/soundwaves are acting as the face of the driver, meaning that it makes the exit point a much larger driver face 'diameter' and therefore increases output over the actual driver face?

And that if you undersize the horn, you are effectively making a large ported bandpass box instead?
MemX is offline  
post #27 of 485 Old 08-13-2013, 09:30 AM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 935
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 16 Post(s)
Liked: 37
This might be easier to see by comparing some ideal horns as this is way more complicated to see with reactive horns....

Below you will find three different horns. Each one is using the B&C 12PE32 driver mounted on an infinite baffle, firing into an ideal 31Hz exponential horn attached to an infinite long tube (0xPI space). The only difference is the compression ratio: I picked 1:1, 4:1, and 8:1.


SPL - 1:1


SPL - 4:1


SPL - 8:1



Acoustic Impedance - 1:1


Acoustic Impedance - 4:1


Acoustic Impedance - 8:1



Efficiency - 1:1


Efficiency - 4:1


Efficiency - 8:1



And then we need to introduce another concept called "Efficiency Bandwidth Product"....where we need to integrate the efficiency across the bandwidth we care about. So for a very simple example, a speaker with 50% efficiency from 50Hz to 100Hz has the same EBP as a speaker with 25% efficiency from 50Hz to 150Hz.

If you look closely between examples, the HF corner is moving higher as the compression ratio increases. You will also note that the 4:1 compression ratio has the highest sensitivity in the midband with 104dB at 100Hz. Both the 1:1 and 8:1 are 103dB at 100Hz, but the 8:1 goes higher in frequency (here's an overlay to compare):

The difference is an F3 of 700Hz versus 450Hz, so the 8:1 horn is covering 60% more bandwidth....which is why we're seeing a lower peak efficiency.



Then when we compare the MaxSPL charts, we'll see that the 1:1 horn is excursion limited over a wide bandwidth. The 4:1 horn is excursion limited right at the thermal limit of the voice coil, and then the 8:1 is limited thermally over the entire passband.

MaxSPL - 1:1


MaxSPL - 4:1


MaxSPL - 8:1


This is showing what I meant earlier by increasing the impedance not helping you because you run out of motor. The ideal compression ratio for this driver is 4:1, and the EBP is maximized with a 4:1 compression as well. This is because the "electro-mechanical" output impedance of the driver is entirely reactive at its Fs with a magnitude of 0.3176. Thiele/Small already showed us that the radiation impedance of a piston is incredibly low - which is why we're dealing with such small numbers. Also note that the units of electrical ohms and acoustic impedance are different (so you can't just compare the two directly).

If I didn't want the high frequency extension, then I could redesign my driver so that the energy shifts lower in frequency - and then you'll see more output at 100Hz instead of more output at 1kHz. A ton more energy is required to move the driver back and forth at faster rates (higher frequencies), which is ultimately the reason why I think some of your plots are a bit misleading...


That said - the bottom end of the horn's response is going to be reactive - and we even see that in these plots. And we're maximizing the horn by pushing the horn Fc close to the driver Fs (the horn is reactive where the driver is reactive). If we add the rear volume around the driver, then we introduce another reactive component - and if we port the rear volume into the throat, then there is yet another reactive component we can deal with. For a bass horn, we're always looking to push the response lower, so it's really important to understand the reactive nature and use it to our advantage....if you look closely, you can start to see which kinds of flares would be ideal for bass horns in this article:
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf

Ok, I really need to get back to work...
MarkHotchkiss and MemX like this.

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #28 of 485 Old 08-13-2013, 09:35 AM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 935
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 16 Post(s)
Liked: 37
Quote:
Originally Posted by MemX View Post

This thread makes my hurt... lol
Is this saying that a Horn is designed to allow the air/soundwaves from the front of the driver to expand out at a constant rate and velocity, the end result being that when it exits the cabinet the air/soundwaves are acting as the face of the driver, meaning that it makes the exit point a much larger driver face 'diameter' and therefore increases output over the actual driver face?

And that if you undersize the horn, you are effectively making a large ported bandpass box instead?

That is definitely one way to conceptualize it.

The weird thing is that it actually performs better than direct radiators that match the area of the mouth - and this gets brought up a few times in the literature. I think the general belief is that the "acoustic size" of the mouth appears bigger than its actual physical size due to how the air interacts near the corners of the mouth. I've not seen any published math on just that concept though.

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #29 of 485 Old 08-13-2013, 11:10 AM
Advanced Member
 
MemX's Avatar
 
Join Date: Oct 2012
Posts: 650
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 45
Awesome post on Horn Theory, thank you very much smile.gif

It certainly is not a simple science!! I think I will have to stick to WinISD and sealed setups for my first DIY attempt, they seem so much easier and I seem less likely to break something or make a complete mess of things... lol biggrin.gif
MemX is offline  
post #30 of 485 Old 08-13-2013, 11:36 AM
AVS Special Member
 
beastaudio's Avatar
 
Join Date: Dec 2010
Location: Western NC
Posts: 7,014
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 255 Post(s)
Liked: 573
Quote:
Originally Posted by MemX View Post

That is quite impressive!

My only fear with ported is bottoming the drivers out, even with all the electronic gizmos that can limit power etc!

This is a front loaded horn, therefore it is essentially a sealed box with a horn loaded on the front baffle. No ports here smile.gif

(European models do not accept banana plugs.)

 

"If you done it, it ain't bragging." ~ Walt Whitman

 

AE TD12x SEOS12 Build

Beast's DIY Master Measurement Thread

DIY Emminence Coaxial Surround Thread

beastaudio is offline  
Reply DIY Speakers and Subs

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off