Simulations
Here is Bjorn Kolbrek's simulation of the radiation impedance of a 420 Hz Le Cleac'h horn with a T = 0.707 and profile optimized for an Altec 288. The green curve on the right is the power factor, or the power admitted into the horn. The red curve on the left is the reflection coefficient, the power reflected back towards the diaphragm. I see this as similar to a VSWR curve for an antenna, showing the transmitted power versus power reflected back to the transmitter.
Now, it is a simulation, and I treat all simulations with a degree of skepticism. But still, it is interesting to look at, and other simulations by Bjorn Kolbrek showed strong reflections when the throat flare of the horn did not match the internal flare of the 288 driver (which is 8 degrees). Very slight mismatches between the compression driver and the throat flare of the horn create surprisingly large ripples in the 500 Hz to 3 kHz response region.
Here is Bjorn Kolbrek's simulation of the radiation impedance of a 420 Hz Le Cleac'h horn with a T = 0.707 and profile optimized for an Altec 288. The green curve on the right is the power factor, or the power admitted into the horn. The red curve on the left is the reflection coefficient, the power reflected back towards the diaphragm. I see this as similar to a VSWR curve for an antenna, showing the transmitted power versus power reflected back to the transmitter.
Now, it is a simulation, and I treat all simulations with a degree of skepticism. But still, it is interesting to look at, and other simulations by Bjorn Kolbrek showed strong reflections when the throat flare of the horn did not match the internal flare of the 288 driver (which is 8 degrees). Very slight mismatches between the compression driver and the throat flare of the horn create surprisingly large ripples in the 500 Hz to 3 kHz response region.
Attachments
Comment
The reasons for choosing a compression driver & horn bandwidth of 700~850 Hz to 5~7 kHz should be evident. Performance is very high within the bandwidth, but declines rapidly outside of it, which is normal for horns. What little I do know about horns is they do not respond to attempts to "stretch" their bandwidth.
Fortunately, neither the midbass driver nor the ribbon should be greatly stressed by this choice of crossover points. Both drivers should have acceptable performance an octave beyond the nominal crossover points, and the mid-high performance of the midbass driver, along with overall sonics, will be a determining factor in its selection.
The reasons for choosing a compression driver & horn bandwidth of 700~850 Hz to 5~7 kHz should be evident. Performance is very high within the bandwidth, but declines rapidly outside of it, which is normal for horns. What little I do know about horns is they do not respond to attempts to "stretch" their bandwidth.
Fortunately, neither the midbass driver nor the ribbon should be greatly stressed by this choice of crossover points. Both drivers should have acceptable performance an octave beyond the nominal crossover points, and the mid-high performance of the midbass driver, along with overall sonics, will be a determining factor in its selection.
Re: Good Articles
Lynn, having also worked with large PA horns, my findings are slightly different from yours.
I agree that there might be some brute force sensation coming along with that beasts but I trace down the underlying aspects differently.
- the effortless presentation I assume to be a side effect of low thermal transients = BIG voice coils.
- the sensation of "bigness" I assume to be the same like from line arrays = the room angle covered by the source.
- going up "endlessly" in SPL is just that it usually is like that (at listening room distances) due to high efficiency and usually extend excursion designs.
- some of the "lifelike clarity" I assume to be due to low Doppler IM with large radiating areas
Put one or two 15" besides your Ariel and cross at around 300-500Hz – I'll bet, you'll experience the sensation you are heading for (within the limits of the Ariel drivers).
Greetings
Michael
Lynn Olson said:
Lynn, having also worked with large PA horns, my findings are slightly different from yours.
I agree that there might be some brute force sensation coming along with that beasts but I trace down the underlying aspects differently.
- the effortless presentation I assume to be a side effect of low thermal transients = BIG voice coils.
- the sensation of "bigness" I assume to be the same like from line arrays = the room angle covered by the source.
- going up "endlessly" in SPL is just that it usually is like that (at listening room distances) due to high efficiency and usually extend excursion designs.
- some of the "lifelike clarity" I assume to be due to low Doppler IM with large radiating areas
Put one or two 15" besides your Ariel and cross at around 300-500Hz – I'll bet, you'll experience the sensation you are heading for (within the limits of the Ariel drivers).
Greetings
Michael
I have been around long enough to have read articles in the old
Wireless world etc about horn design. A lot of what was written then is much derided now but a golden rule I understood from what I read was that no horn should handle more than 3 to 31/2 octaves, probably less at low frequencies. From what has been shown lately theu seem to have at least got that part right.
jamikl
Wireless world etc about horn design. A lot of what was written then is much derided now but a golden rule I understood from what I read was that no horn should handle more than 3 to 31/2 octaves, probably less at low frequencies. From what has been shown lately theu seem to have at least got that part right.
jamikl
Hello,
We have to take care about such "universally admitted sentence" which comes from a time that gave us a lot of empirical formulas... and a lot of bad horns.
Could someone explains to us why a horn could not be used to reproduce an interval of frequency larger than 3 octaves...
Best regards from Paris, France
Jean-Michel Le Cléac'h
We have to take care about such "universally admitted sentence" which comes from a time that gave us a lot of empirical formulas... and a lot of bad horns.
Could someone explains to us why a horn could not be used to reproduce an interval of frequency larger than 3 octaves...
Best regards from Paris, France
Jean-Michel Le Cléac'h
jamikl said:
Jmmlc said:
Not to be a jerk but that's a silly question coming from you. You should know that all horn/drivers are different as well as what is expected of them. For example I could not live with this horn Lynn is using beyond 2K (a little over an octave) looking at the graph of it's narrowing directivity- it would need a treble unit because I expect better dispersion - (should I now call his horn a bad horn? No, it seem to be what he wants)Also look at the response of the Altec 288 beyond 6K - the treble is missing. In any horn it would only be good out to around 6K
Another example would be a bass horn where the driver won't do 3 or more octaves. There are unlimited examples of horns that work well (as designed) over less then 3 octaves
"Here's the measured directivity of a Le Cleac'h 320 Hz horn (top graph), courtesy of the author. The directivity pattern looks very much like a 1.4" to 2" direct-radiator dome-midrange."
Hello Lynn
With twice the DI number for the horn. The included angle 6 db point is twice the horns for a 2 in driver. At 10K you are looking at 80 degrees for the driver and less than 40 degrees for the horn.
Rob
Hello Lynn
With twice the DI number for the horn. The included angle 6 db point is twice the horns for a 2 in driver. At 10K you are looking at 80 degrees for the driver and less than 40 degrees for the horn.
Rob
Hello,
There is no silly questions only silly ******s. ;-)
From an historical point of view, to my knowledge, the increasing directivity of horns were never used as an argument to the "3 octave question"...
Most often the upper limit of the power response is one of the arguments.
If we use the empirical (again) formula for the lower frequency limit Fl and for the upper frequency limit Fh (see JBL technical paper http://www.petoindominique.fr/pdf/JBL_max-ouput-midrange&low-frequency.pdf )
Fl = Fs . Qts / 2 and Fh = 2 . Fs / Qts
Generally compression drivers are used above the resonance frequency of the (loaded) loudspeaker Fs with horns having quite the same cut off frequency: Fc ~= Fs . (well if you use an Altec 802 it can be different due to high Fs...). This means that for a horn loaded compression driver we have to consider as the bandwith the frequency interval between Fs and Fs/(2 . Qts).
A rapid calulation will indicate that Fh = 8. Fs (3 octave bandwith) for Qts = 0.25. I let you know if this is relevant in your case...
This is for the power response. Power response is very important when you have to design folded bass horns by example (see Marshall leach and others).
But is it important for straight horns? Your own taste can lead you to reply YES if you prefer the horn or the waveguide to have a response similar in shape to the power response. My own answer will be NO , I don't want my driver + horn equalized and my preference for music recorded in phase stereophony lead me to listen alone at the sweet spot on my main system so I don't care so much about the off axis response (if the reverberated field is OK to those ears).
To my knowledge there is no compression driver for which the upper limit of the power response as mesured on planar wave tube is over 4kHz or such. Does this means that the compression driver cannot reproduce 18kHz? No, see on axis response for an unaqualized TAD TD2001 driver on a Le Cléac'h horn as a good example.
Have a good week-end.
Best regards from Paris, France
Jean-Michel Le Cléac'h
There is no silly questions only silly ******s. ;-)
From an historical point of view, to my knowledge, the increasing directivity of horns were never used as an argument to the "3 octave question"...
Most often the upper limit of the power response is one of the arguments.
If we use the empirical (again) formula for the lower frequency limit Fl and for the upper frequency limit Fh (see JBL technical paper http://www.petoindominique.fr/pdf/JBL_max-ouput-midrange&low-frequency.pdf )
Fl = Fs . Qts / 2 and Fh = 2 . Fs / Qts
Generally compression drivers are used above the resonance frequency of the (loaded) loudspeaker Fs with horns having quite the same cut off frequency: Fc ~= Fs . (well if you use an Altec 802 it can be different due to high Fs...). This means that for a horn loaded compression driver we have to consider as the bandwith the frequency interval between Fs and Fs/(2 . Qts).
A rapid calulation will indicate that Fh = 8. Fs (3 octave bandwith) for Qts = 0.25. I let you know if this is relevant in your case...
This is for the power response. Power response is very important when you have to design folded bass horns by example (see Marshall leach and others).
But is it important for straight horns? Your own taste can lead you to reply YES if you prefer the horn or the waveguide to have a response similar in shape to the power response. My own answer will be NO , I don't want my driver + horn equalized and my preference for music recorded in phase stereophony lead me to listen alone at the sweet spot on my main system so I don't care so much about the off axis response (if the reverberated field is OK to those ears).
To my knowledge there is no compression driver for which the upper limit of the power response as mesured on planar wave tube is over 4kHz or such. Does this means that the compression driver cannot reproduce 18kHz? No, see on axis response for an unaqualized TAD TD2001 driver on a Le Cléac'h horn as a good example.
Have a good week-end.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Magnetar said:
Re: Re: Good Articles
Not just the big voicecoils, I think much of what makes prostyle and/or high efficiency and/or fullrange drivers so involving at low levels is probably due to the high Qms. Less memory effect in the materials, and the driver controlled more by the motor.
mige0 said:
Lynn, having also worked with large PA horns, my findings are slightly different from yours.
I agree that there might be some brute force sensation coming along with that beasts but I trace down the underlying aspects differently.
- the effortless presentation I assume to be a side effect of low thermal transients = BIG voice coils.
- the sensation of "bigness" I assume to be the same like from line arrays = the room angle covered by the source.
- going up "endlessly" in SPL is just that it usually is like that (at listening room distances) due to high efficiency and usually extend excursion designs.
- some of the "lifelike clarity" I assume to be due to low Doppler IM with large radiating areas
Put one or two 15" besides your Ariel and cross at around 300-500Hz � I'll bet, you'll experience the sensation you are heading for (within the limits of the Ariel drivers).
Greetings
Michael
Not just the big voicecoils, I think much of what makes prostyle and/or high efficiency and/or fullrange drivers so involving at low levels is probably due to the high Qms. Less memory effect in the materials, and the driver controlled more by the motor.
Lynn, et al,
Your design is quite intriguing, and I continue to follow this thread daily (at least!).
As it has developed I have enjoyed learning about the various drivers under consideration and it has lead to lots of fun research/thinking of my own.
The re-mention of the 6ND410 a few posts ago leads me to some thoughts that I hope aren't perceived as too far off-topic.
As you outlined, the appeal of a large MR driver is compelling (i.e. my unused JBL 2123's or the 10" 18sound MR units that have been discussed). My question is, in general, as part of overall system design, whether you (or others) believe in theoretically ideal XO points (or conversely, XO frequencies to be avoided). As an example, Colloms, in his book (and others) recommend as theoretically ideal the avoidance of a crossover in the ~300-to at least 3k (and preferably higher) range. Your horn choice obviously crosses a bit higher at the low end but allows the next break-point to the RAAL to be nice and high (and your previously-demonstrated skill in crossover design will certainly be of great help in your efforts).
Modelling of the 6ND410 suggests that in a baffle large enough to maintain 2 pi radiation to its lower crossover point (at least 300Hz and perhaps even 250Hz) that it could reach 120db (as a single unit) well within 2mm of excursion. While realizing that radiation pattern isn't your top design criterion, this unit could (by my preliminary/theoretical/non-measured calculations) match the XT1086's horizontal 80 degree radiation pattern at 3- 3.1 kHz and maintain high overall system sensitivity. It could also be crossed over to the larger (not double-high) RAAL unit at this level (while obviously not matching radiation pattern or as high overall system sensitivity).
I thought a bit about 2 vertically-aligned 6ND410's to match a bit better the horizontal dispersion of the large RAAL ribbon, but suspect that the driver-to-driver spacing of the 6ND410's would introduce combing if the higher crossover frequency was used, again forcing the XO frequency lower.
Any thoughts?
Sincerely,
Bill
Your design is quite intriguing, and I continue to follow this thread daily (at least!).
As it has developed I have enjoyed learning about the various drivers under consideration and it has lead to lots of fun research/thinking of my own.
The re-mention of the 6ND410 a few posts ago leads me to some thoughts that I hope aren't perceived as too far off-topic.
As you outlined, the appeal of a large MR driver is compelling (i.e. my unused JBL 2123's or the 10" 18sound MR units that have been discussed). My question is, in general, as part of overall system design, whether you (or others) believe in theoretically ideal XO points (or conversely, XO frequencies to be avoided). As an example, Colloms, in his book (and others) recommend as theoretically ideal the avoidance of a crossover in the ~300-to at least 3k (and preferably higher) range. Your horn choice obviously crosses a bit higher at the low end but allows the next break-point to the RAAL to be nice and high (and your previously-demonstrated skill in crossover design will certainly be of great help in your efforts).
Modelling of the 6ND410 suggests that in a baffle large enough to maintain 2 pi radiation to its lower crossover point (at least 300Hz and perhaps even 250Hz) that it could reach 120db (as a single unit) well within 2mm of excursion. While realizing that radiation pattern isn't your top design criterion, this unit could (by my preliminary/theoretical/non-measured calculations) match the XT1086's horizontal 80 degree radiation pattern at 3- 3.1 kHz and maintain high overall system sensitivity. It could also be crossed over to the larger (not double-high) RAAL unit at this level (while obviously not matching radiation pattern or as high overall system sensitivity).
I thought a bit about 2 vertically-aligned 6ND410's to match a bit better the horizontal dispersion of the large RAAL ribbon, but suspect that the driver-to-driver spacing of the 6ND410's would introduce combing if the higher crossover frequency was used, again forcing the XO frequency lower.
Any thoughts?
Sincerely,
Bill
Charles Hansen said:
My daughter nabbed my camera so I 'll have to post a pic when I get it back - I'll start a thread then
It's an open baffle with no wings and a little eq on the bottom - nice and compact (for me at least)
It's triamped with the KILLER Beyma 21L50 21" woofer up to 150 cycles then for mid is a modified Electro Voice SRO12 with the 'coffee can' alnico magnet. For 1500 up it is the alnico Emilar EA175 1" compression driver with aluminum diaphragm/mylar surround loaded in a really trick horn I found- the Renkus Heinz Complex Conic CCH12009 round horn. No tweeter needed.
My goal was high efficiency open baffle 3-way with bass into the 30's and mid crossover below 200 cycles - tone was also at the top of the list - It does it well..Complex Conic Horn