Need some advise from the experimented back loaded horn designers out there 
I have been playing with a design of a back loaded horn using a Fostex FE207E driver. From all the reading I have done in the past 2 months, I have not been able to decide which contour is best and what are the best sizes for the throat and compression chamber.
I have been playing in Autocad with the hyperbolic contour and formulas I have found on Dominique Petoin's site in France. Now that I am comfortable with the tools, I would like to tackle the throat size issue.
With my current design, it will be easy to play with the compression chamber size, but not the throat size, so it is essential that I find the best estimate for the throat area.
I have found multiple ways to calculate throat area:
0.5 x the piston area (smaller then Sd)
0.5 x Sd
0.7 x Sd
between 0.5 and 0.7 Sd
1 x Sd
This is all very confusing
Does anybody have practical experience with the throat sizing and the Fostex FE207E or FE206E ? Any practical info about throat shape and location ?
I am also trying to find a good source of information about the shape of the sound waves generated by a loudspeaker. Were there ever any tests performed under water to simulate the wave propagation in front behind the driver ?
If it is of some importance ... the design is similar to the Little Big Horn from Carfrae ( http://www.carfrae.com )
I have been playing with a design of a back loaded horn using a Fostex FE207E driver. From all the reading I have done in the past 2 months, I have not been able to decide which contour is best and what are the best sizes for the throat and compression chamber.
I have been playing in Autocad with the hyperbolic contour and formulas I have found on Dominique Petoin's site in France. Now that I am comfortable with the tools, I would like to tackle the throat size issue.
With my current design, it will be easy to play with the compression chamber size, but not the throat size, so it is essential that I find the best estimate for the throat area.
I have found multiple ways to calculate throat area:
0.5 x the piston area (smaller then Sd)
0.5 x Sd
0.7 x Sd
between 0.5 and 0.7 Sd
1 x Sd
This is all very confusing
Does anybody have practical experience with the throat sizing and the Fostex FE207E or FE206E ? Any practical info about throat shape and location ?
I am also trying to find a good source of information about the shape of the sound waves generated by a loudspeaker. Were there ever any tests performed under water to simulate the wave propagation in front behind the driver ?
If it is of some importance ... the design is similar to the Little Big Horn from Carfrae ( http://www.carfrae.com )
Hi,
I thought piston area=Sd.
the throat to Sd ratio deteremines a number of characteristics of the horn.
two that spring to mind are efficiency and distortion.
I think these two go in opposite directions, so that determines the first compromise.
A high ratio i.e. throat<<Sd can only be driven by some drivers.
Can someone tell us whether those Fostek models come into that range?
0.5 (Sd/St=2) should work for most decent drivers. You should not need to go as far as 0.7 (Sd/St=1.4) and certainly not 1.
I thought piston area=Sd.
the throat to Sd ratio deteremines a number of characteristics of the horn.
two that spring to mind are efficiency and distortion.
I think these two go in opposite directions, so that determines the first compromise.
A high ratio i.e. throat<<Sd can only be driven by some drivers.
Can someone tell us whether those Fostek models come into that range?
0.5 (Sd/St=2) should work for most decent drivers. You should not need to go as far as 0.7 (Sd/St=1.4) and certainly not 1.
Ah, throat area. Interesting subject. I'm not surprised you're confused though.
OK, getting back to the basic physics, believe it or not, & contrary to what most people expect, driver Sd is actually pretty much immaterial to the throat area per se. So don't get too hung up on it -Sd is more a convenient measure than anything else. Have a gander at these:
Single Driver site on horn throats & general design:
http://melhuish.org/audio/horndesign.html
Leach's horn design paper: http://users.ece.gatech.edu/~mleach/papers/HornPaper/HornPaper.pdf
Not many mentions of Sd...
Martin King's site is also required reading, but IIRC, he does not prescribe a method for selecting the ~ideal throat CSA. http://www.quarter-wave.com/Horns/Horn_Theory.html
You might find this interesting too: Olson's original patent for adding a filter chamber to a horn (note -the room is the compression chamber for a BLH). Still makes interesting reading: http://www.google.com/patents?vid=USPAT2224919&id=-cBwAAAAEBAJ&pg=PP1&dq=harry+f.+olson#PPA3,M1
Best
Scott
OK, getting back to the basic physics, believe it or not, & contrary to what most people expect, driver Sd is actually pretty much immaterial to the throat area per se. So don't get too hung up on it -Sd is more a convenient measure than anything else. Have a gander at these:
Single Driver site on horn throats & general design:
http://melhuish.org/audio/horndesign.html
Leach's horn design paper: http://users.ece.gatech.edu/~mleach/papers/HornPaper/HornPaper.pdf
Not many mentions of Sd...
Martin King's site is also required reading, but IIRC, he does not prescribe a method for selecting the ~ideal throat CSA. http://www.quarter-wave.com/Horns/Horn_Theory.html
You might find this interesting too: Olson's original patent for adding a filter chamber to a horn (note -the room is the compression chamber for a BLH). Still makes interesting reading: http://www.google.com/patents?vid=USPAT2224919&id=-cBwAAAAEBAJ&pg=PP1&dq=harry+f.+olson#PPA3,M1
Best
Scott
Using the "Single Driver" website recipe ...
Fostex FE207E Fostex FE206E
Fs= 39 Hz 39 Hz
Zn= 8 Ohms 8 Ohms
Sd= 0.0261 m2 0.0261 m2
Qts= 0.26 0.18
Vas= 56.25 L or 0.05626 m3 54.53 L or 0.05453 m3
Mms= 15.02 g 15.35 g
Xmax= 1.5 mm 1.5 mm
Re= 6.73 Ohms 6.69 Ohms
Qms= 3.86 3.73
Qes= 0.28 0.18
BL= 9.41 Tesla/m 11.82 Tesla/m
Cms= 0.00094 mm/N 0.00091 mm/N
At=(2pi*Fs*Qts*Vas)/c
At is the area of the throat (m2)
FS is the nominal air resonance of the driver
QTS is the total Q of the driver (note, some formulas use QES, the electrical Q here - Bruce Edgar for one)
VAS is the equivalent volume suspension (m3)
c is the speed of sound (344 m/s)
Fostex FE207E At= (2pi*39*0.26*0.05626)/344 = 0.01041981778047 m2 = 104.198 cm2 = 16.15 sq in
Fostex FE206E At= (2pi*39*0.18*0.05453)/344 = 0.006991897470698 m2 = 69.919 cm2 = 10.84 sq in
Using the piston area for both drivers:
0.5 x Sd = 103.5 cm2 = 16.04 sq in
0.7 x Sd = 144.27 cm2 = 22.36 sq in
Very similar result with 1/2 the piston area on the FE207E but does not work with the FE206E.
You are probably right saying that Sd is not needed in the calculation, but there is an indirect relationship. It is somewhere in the Q and Vas of the driver.
Let me play with the Leach formulas, and I will post the results.
Any thoughts on the throat shape and position ?
Fostex FE207E Fostex FE206E
Fs= 39 Hz 39 Hz
Zn= 8 Ohms 8 Ohms
Sd= 0.0261 m2 0.0261 m2
Qts= 0.26 0.18
Vas= 56.25 L or 0.05626 m3 54.53 L or 0.05453 m3
Mms= 15.02 g 15.35 g
Xmax= 1.5 mm 1.5 mm
Re= 6.73 Ohms 6.69 Ohms
Qms= 3.86 3.73
Qes= 0.28 0.18
BL= 9.41 Tesla/m 11.82 Tesla/m
Cms= 0.00094 mm/N 0.00091 mm/N
At=(2pi*Fs*Qts*Vas)/c
At is the area of the throat (m2)
FS is the nominal air resonance of the driver
QTS is the total Q of the driver (note, some formulas use QES, the electrical Q here - Bruce Edgar for one)
VAS is the equivalent volume suspension (m3)
c is the speed of sound (344 m/s)
Fostex FE207E At= (2pi*39*0.26*0.05626)/344 = 0.01041981778047 m2 = 104.198 cm2 = 16.15 sq in
Fostex FE206E At= (2pi*39*0.18*0.05453)/344 = 0.006991897470698 m2 = 69.919 cm2 = 10.84 sq in
Using the piston area for both drivers:
0.5 x Sd = 103.5 cm2 = 16.04 sq in
0.7 x Sd = 144.27 cm2 = 22.36 sq in
Very similar result with 1/2 the piston area on the FE207E but does not work with the FE206E.
You are probably right saying that Sd is not needed in the calculation, but there is an indirect relationship. It is somewhere in the Q and Vas of the driver.
Let me play with the Leach formulas, and I will post the results.
Any thoughts on the throat shape and position ?
Position of the throat in the back chamber will noticably affect the response. That's where Martin's MathCad worksheets come in, so you can model your intended enclosure. Martin charges a nominal $25 fee for the full set, which is an outright bargin given what they let you do, and the expensive mistakes they can save you from making.
For example, here's a horn (it's not an optimised box!) with the throat positioned at the opposite end of the back chamber to the driver:
For example, here's a horn (it's not an optimised box!) with the throat positioned at the opposite end of the back chamber to the driver:
Attachments
And here's the same horn, back-chamber & driver, but with the throat moved to the same end of the back-chamber as the driver. Makes quite a difference doesn't it?
As I say, this is where the software helps you select an optimal location for the throat, which will in turn affect how you fold the thing, because AFAIK, there are no formulas etc written down for establishing the best place for it -you either have to build a couple of prototypes, and measure them, or model it. The latter usually works out quicker, and a heck of a lot cheaper.
As I say, this is where the software helps you select an optimal location for the throat, which will in turn affect how you fold the thing, because AFAIK, there are no formulas etc written down for establishing the best place for it -you either have to build a couple of prototypes, and measure them, or model it. The latter usually works out quicker, and a heck of a lot cheaper.
Attachments
I always have difficulty with the interpretation of models.
In this case, I see an added dip in the 350HZ zone when the throat is at the back of the compression chamber.
In a real model, we would have to take into consideration the turbulance the the compression chamber, the fluidity of the access to the throat of the horn, the damping material and position, etc ...
Where can I get more nfo on Martin's worksheets ?
In this case, I see an added dip in the 350HZ zone when the throat is at the back of the compression chamber.
In a real model, we would have to take into consideration the turbulance the the compression chamber, the fluidity of the access to the throat of the horn, the damping material and position, etc ...
Where can I get more nfo on Martin's worksheets ?
Oh yes. All of which are accounted for in the above, and can be adjusted in MathCad. Nope -I'm not joking.
It all comes back to QW theory & standing waves generated by an enclosure. Martin's work & software is derived from this, which he's advanced our knowledge of massively in recent years. See his site at www.quarter-wave.com I cannot recommend it highly enough. Required reading, and purchase IMO. Every speaker I have designed has been modelled & optimised in his worksheets. Their staggering accuracy is proven. They can't design a cabinet for you, but they're probably the most powerful tool available to the DIY speaker designer.
It all comes back to QW theory & standing waves generated by an enclosure. Martin's work & software is derived from this, which he's advanced our knowledge of massively in recent years. See his site at www.quarter-wave.com I cannot recommend it highly enough. Required reading, and purchase IMO. Every speaker I have designed has been modelled & optimised in his worksheets. Their staggering accuracy is proven. They can't design a cabinet for you, but they're probably the most powerful tool available to the DIY speaker designer.
now take the two recommended throat areas and compare them to Sd.lowtherdream said:
the 207 comes out @ ~=2
the 206 @ ~=3.2
Both these compression ratios are just about in the range that most/many drivers can live with.
Note that the 206 with the much stronger damping (stronger magnet?) can accept the higher compression ratio.
I suspect compression ratio may have been invented to avoid the maths, that you went through, to get to 16.2 & 10.8 sq in.
The smaller throat of the 206 may give a slightly higher efficiency at the expense of slightly higher distortion.
That's where prototyping comes in where you can compare different horns trying to find the ONE that appeals to the designer the most.
I am just trying to find the best estimated throat size.
Prototyping horns takes time and money that's why I will take all the info I can get.
I understand there is no magic recipe, but I am now tempted to build a sample compression chamber and play with the chamber volume, throat size and position, and damping materials.
What is the best DIY way to measure what is coming out of the throat area of that jig ?
Prototyping horns takes time and money that's why I will take all the info I can get.
I understand there is no magic recipe, but I am now tempted to build a sample compression chamber and play with the chamber volume, throat size and position, and damping materials.
What is the best DIY way to measure what is coming out of the throat area of that jig ?
lowtherdream,
I am not sure that your experiment will yield many useful results. The front loaded horn's throat size, length, and mouth size are determined by the type of horn flare and the lower cut-off frequency desired. Then the compression chamber size is determined by the throat size and the desired higher frequency cut-off. The assumption is that the acoustic impedance at the throat is purely resistive at the higher cut-off frequency. You need an entire system to see how it all works together, a single piece will not function all alone.
For a back loaded horn things get a little more complicated. Most designs are essentially TL's at lower frequencies and transition to horn like behavior as frequency rises. But the relationship between the throat area and the compression chamber volume remains the same if the design is done correctly. A high frequency roll-off is desired to eliminate output from the horn mouth producing a smooth midrange and high frequency response from the driver/horn system. The mouth still needs to be big enough to produce essentially a resistive acoustic impedance at the higher cut-off frequency.
Lowther Owner,
MJK,
thanks for the advice (or warning). If I have acceptable results with this first enclosure, maybe I will give myself the right to buy Lowther drivers. If I get poor results, I will be able to blame the Fostex and not Lowther
I agree all the components of the enclosure are critical to the final experience, but if I can understand better how the chamber and throat work, I'll be starting with a solid base. The basic behavior will always be true, only the volume of the chamber will have to be modified for a specific contour. Maybe the high frequency cutoff will also have to be adjusted with damping material. Plus it is a cheap way to experiment with the FE207E or FE206E and learn to know them.
Martin's work tells us that the throat should be at the front of the chamber, nearest to the driver. Yet, many design back loaded folded horns with the throat area at the rear of the chamber.
I'm also assuming the shape of the chamber will have little effect on the results. Is there an advantage to a cylindrical compression chamber ?
thanks for the advice (or warning). If I have acceptable results with this first enclosure, maybe I will give myself the right to buy Lowther drivers. If I get poor results, I will be able to blame the Fostex and not Lowther
I agree all the components of the enclosure are critical to the final experience, but if I can understand better how the chamber and throat work, I'll be starting with a solid base. The basic behavior will always be true, only the volume of the chamber will have to be modified for a specific contour. Maybe the high frequency cutoff will also have to be adjusted with damping material. Plus it is a cheap way to experiment with the FE207E or FE206E and learn to know them.
Martin's work tells us that the throat should be at the front of the chamber, nearest to the driver. Yet, many design back loaded folded horns with the throat area at the rear of the chamber.
I'm also assuming the shape of the chamber will have little effect on the results. Is there an advantage to a cylindrical compression chamber ?
Wrong. Martin's work does not show that that throat should be at the front of the chamber.
The plots I showed above indicate that a specific driver, mounted into a specific chamber (not just its volume, but its shape as well), with a specific amount of damping in it, feeding a specific horn (and none of these were optimised I hasten to add) will give a different response depending on where the throat is positioned. That's not universally applicable -all I was doing is illustrating that the response will invariably change somewhat depending on where you place the throat in the chamber. I wasn't advocating always using a given location because in some designs it will be optimal, in others less so.
OK, I think you might need to step back a bit here & have a re-think. No offense intended -just trying to help out so you don't waste your time & hard-earned. You don't design a horn, then figure out the throat size. Remember, the throat size is a factor in determining the eventual length of the horn. Everything is integral.
I'd suggest you start by reading Martin's horn design papers, http://www.quarter-wave.com/Horns/Horn_Theory.html
and then the horn design pages on the Single Driver Site http://melhuish.org/audio/horn.html for additional information, such as using different flare profiles / contours, and how to define the optimum throat size based on the driver parameters, rather than using a generic one (we've already discussed that part above). Don't try building a cabinet though until you've got a good grasp of the theory / physics / formulas. It'll save you time & money in the long run & you'll get a better speaker out of it.
Basic method I use. You need to define, in this order:
1) Mouth area for the radiation space it's going to be firing into
2) Throat area
3) Flare constant (I like hypex / hyperbolic horns myself)
4) Length
5) Filter chamber volume.
You can then adjust the details from there using software like Martin's MathCad worksheets.
The plots I showed above indicate that a specific driver, mounted into a specific chamber (not just its volume, but its shape as well), with a specific amount of damping in it, feeding a specific horn (and none of these were optimised I hasten to add) will give a different response depending on where the throat is positioned. That's not universally applicable -all I was doing is illustrating that the response will invariably change somewhat depending on where you place the throat in the chamber. I wasn't advocating always using a given location because in some designs it will be optimal, in others less so.
OK, I think you might need to step back a bit here & have a re-think. No offense intended -just trying to help out so you don't waste your time & hard-earned. You don't design a horn, then figure out the throat size. Remember, the throat size is a factor in determining the eventual length of the horn. Everything is integral.
I'd suggest you start by reading Martin's horn design papers, http://www.quarter-wave.com/Horns/Horn_Theory.html
and then the horn design pages on the Single Driver Site http://melhuish.org/audio/horn.html for additional information, such as using different flare profiles / contours, and how to define the optimum throat size based on the driver parameters, rather than using a generic one (we've already discussed that part above). Don't try building a cabinet though until you've got a good grasp of the theory / physics / formulas. It'll save you time & money in the long run & you'll get a better speaker out of it.
Basic method I use. You need to define, in this order:
1) Mouth area for the radiation space it's going to be firing into
2) Throat area
3) Flare constant (I like hypex / hyperbolic horns myself)
4) Length
5) Filter chamber volume.
You can then adjust the details from there using software like Martin's MathCad worksheets.
Poor or good results will probably be independent of the drivers. Swapping a Fostex driver to a Lowther driver in a poorly performing design will probably only make the results even worse. The problem will be achieving adequate bass output from the enclosure and a smooth system response.
Unless you are really lucky, the first design will not work as well as following designs. The learning curve is very steep. The question becomes where do you want to do your early learning, in the wood shop or on the computer? How much time and money do you want to invest?
Personally, I measure and prototype on the computer and only build something if I feel it has a high probability of good performance. I can examine many different design options in a matter of hours, the only investment is my time. I am better at the computer then I am at building so this method works for me. Other people come at the problem differently.
Scottmoose said:
Ooh......... been there, done that and only a SLM, kit built 'o' scope, sine-square wave signal generator and log graph paper to measure/plot the results.
I'm here to tell you it was no 'walk in the park', consuming just about every free minute I could squeeze out of a work week for the better part of a year since I was too math challenged to learn very much of it reading textbooks, patents, JAES papers and many of the Audio magazine tech articles. Just as well as it turned out since some of the more respected works were somewhat flawed/mis-leading.The upside is that the longer I've been on the various audio forums the more I've come to rather arrogantly believe I understand the 'craft' of horn design much better than folks who try to learn via simming programs, so for any folks who have the time, perseverance, I highly recommend the prototyping process, especially with today's relatively cheap power woodworking tools and very comprehensive measurement systems available to the DIYer.
No disagreement from me guys. I'm currently restoring a 1940s galvanometer that I got my evil hands on last year. Mostly in perfect condition -it was buried under a pile of junk in an NHS store for years. Just needs cleaning up. Should be an interesting addition to the test bench in the garage.
Some people like to learn by reading ... some like to learn by experimenting with the concepts. Others will find comfort in validating the theory with practical experiments.
Depending on your background, abilities and education level, you will choose what fits your profile. It's all good
I personally like to both read, and experiment with concepts. Of course, money and time are part of that equation
After reading (not necessarily understanding) many books, papers, forums, etc ... I am still not sure if creating a "rear loaded horn" is art, science, magic, or a closely guarded secret. My trips to the audio show proved that not many have the right recipe.
One thing is for sure ... I did not chose the easy way. The rear loaded horn is a very complex project and I am crazy enough not to build a model, proven to give acceptable results.
Listening to what a lot of you recommend, I will take the time to better understand the theory and spend time simulating with programs and spreadsheets. But that will not stop me from trying a few experiments; it's in my nature
Depending on your background, abilities and education level, you will choose what fits your profile. It's all good
I personally like to both read, and experiment with concepts. Of course, money and time are part of that equation
After reading (not necessarily understanding) many books, papers, forums, etc ... I am still not sure if creating a "rear loaded horn" is art, science, magic, or a closely guarded secret. My trips to the audio show proved that not many have the right recipe.
One thing is for sure ... I did not chose the easy way. The rear loaded horn is a very complex project and I am crazy enough not to build a model, proven to give acceptable results.
Listening to what a lot of you recommend, I will take the time to better understand the theory and spend time simulating with programs and spreadsheets. But that will not stop me from trying a few experiments; it's in my nature
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.