I would believe that. Flat phase, too. Here is a measurement of a two-high AMT-1 with trough-type wings at 1m taken from another thread:
Harmonic distortion at 105 dB/1m on axis (which was the Achilles heel of the single AMT-1 without a horn or wings below 2 kHz):
Looks really good to me. I don't think I could fault that kind of performance in any way, especially when looking at the accompanying phase response plotted just above. The price is right, too. (At the time, the AMT-1 was on sale for $120.)
I wouldn't sweat the modeling problem. The real issue will be the effect of the other (lower frequency) MEH drivers on the AMT diaphragm as the SPL rises. This may present an SPL upper limit for an AMT MEH design. I set out to build exactly the same type of loudspeaker to test the stiffness and SPL-induced modulation distortion of the design. XRK971 has already mentioned that his foam-core board construction AMT-MEH was the best sounding design that he'd heard (well after the time after he built and tested/listened to that configuration). I believe it should really turn some heads at audio shows, etc. for anyone that hears them.
The horn loading of one AMT-1 actually decreases the need for using two, I believe. The upside to using two AMT-1s stacked is the near 3-dB increase in SPL at a given input power, but the downside is the vertical polar coverage shrinks even further (the only real concern I've run into with these drivers is that the vertical polar coverage is pretty narrow due to the vertical length of the AMT diaphragm).
There is another use of air-motion transformers (custom design) that is discussed here. I believe the crossover frequency with the dipole-mounted woofers in a D'Appolito configuration was 350-450 Hz. My guess is that the asking price of the "Rainier" was too high considering their driver and raw material cost which worked against successful sales of these dipoles. I bet they sounded really good, however.
Chris
Hi @drumdoctor,
regarding the measurements, please see my comment in the previous post, also please note that my Heils are the first generation, and the diaphragm apparently underwent some development; therefore, my results may be different from others/yours.
Ah, I see that Chris (@Cask05) just chimed in, please read carefully all he has written on the subject, he has quite experience; in fact his post on the Klipsch forum was my inspiration.
Also, I believe that @Robh3606 has also built a system with Heils, you might try to search at the Lansingherritage.org forum.
I tend to agree with Chris on the stacking; I did contemplate it initially, but after listening to selection of my music, I abandoned the idea, as the SPL are sufficient for me. But, please understand that I live in an extremely quiet environment and my listening is at low (average) SPL.
I also agree with his assessment of them being exceptional; I tend to listen to operas/operas' singers, and the reproduction of the female voices is - IMHO - second to none.
Kindest regards,
M
regarding the measurements, please see my comment in the previous post, also please note that my Heils are the first generation, and the diaphragm apparently underwent some development; therefore, my results may be different from others/yours.
Ah, I see that Chris (@Cask05) just chimed in, please read carefully all he has written on the subject, he has quite experience; in fact his post on the Klipsch forum was my inspiration.
Also, I believe that @Robh3606 has also built a system with Heils, you might try to search at the Lansingherritage.org forum.
I tend to agree with Chris on the stacking; I did contemplate it initially, but after listening to selection of my music, I abandoned the idea, as the SPL are sufficient for me. But, please understand that I live in an extremely quiet environment and my listening is at low (average) SPL.
I also agree with his assessment of them being exceptional; I tend to listen to operas/operas' singers, and the reproduction of the female voices is - IMHO - second to none.
Kindest regards,
M
Last edited:
Note that the real reason for the two-high winged-AMT design of that other thread that I linked to above was the need to stitch together the SPL response of the AMT(s) with that of a couple of "H" open baffle woofers that the owner needed to reuse. So depressing the lower frequency "cutoff" of the AMT-1 assembly as far as possible was a design goal for that particular application.
Another use of that design was to be able to replace the top hat (i.e., tweeter and midrange horn) of typical Klipsch Heritage loudspeakers with an AMT mid-high section, which it could also do with a La Scala, Belle, Cornwall or even Heresy woofer or bass bin bottom end.
The real acid test of the design was whether a double-high set of AMT-1s with wings could replace the top two horns on top of a Klipschorn bass bin. It turns out, after measuring the raw performance of the Klipschorn bass bin, I found that the 400 Hz crossover frequency that's been used for many decades is really pushing it. That bass bin really needs to be crossed at a lower frequency. About 220 Hz would be a much better crossover frequency to avoid the two-mouth polar lobing issue of the Khorn bass bin, and to keep the phase response of the bass bin from going off (i.e., the Khorn bass bin experiences non-flat phase performance at the top of its frequency band).
So while I would be careful about using a dual AMT-1 with wings on top of an old Khorn bass bin, using a miniDSP 2x4 HD to cross the dual AMT-1 with wings was in fact a viable "kit" for those wanting to avoid all the issues of the Heritage top hats (and there are many, not the least of which is that they really don't sound that good to my ears).
That was the proposal that I pitched to Rudy in December 2019 that resulted in the thread that I linked, above.
I still believe that an AMT-1-based top end design using a miniDSP 2x4 HD to dial everything end just blows away the old Klipsch Heritage models in terms of consistent controlled directivity polars, phase response, inherent sound quality, and dialing everything in, which was the underlying reason for that K-forum thread.
I found that the sound of horn-loaded woofers--like La Scalas and Belles, perhaps even Khorn bass bins--would sound a lot better than direct radiating woofers, like Cornwalls and Heresies. I was right (I've used dual-stacked AMT-1s now on top of Belle bass bins for my surround loudspeakers, crossed over with a miniDSP 2x4 HD with two used Crown D75-A amplifiers driving them--leaning the top AMT-1s back about 15 degrees to get better vertical coverage for surround loudspeakers. The only thing that I would recommend to others would be to remove the Belle top hats to lower the AMT-1s closer to the centerline of the bass bin mouths, thus improving the crossover polar coverage performance. Then the bass bin could sit on top of a closed-box riser to raise its AMT-1s back up to seated ear height (or slightly higher).
JMTC.
Chris
Another use of that design was to be able to replace the top hat (i.e., tweeter and midrange horn) of typical Klipsch Heritage loudspeakers with an AMT mid-high section, which it could also do with a La Scala, Belle, Cornwall or even Heresy woofer or bass bin bottom end.
The real acid test of the design was whether a double-high set of AMT-1s with wings could replace the top two horns on top of a Klipschorn bass bin. It turns out, after measuring the raw performance of the Klipschorn bass bin, I found that the 400 Hz crossover frequency that's been used for many decades is really pushing it. That bass bin really needs to be crossed at a lower frequency. About 220 Hz would be a much better crossover frequency to avoid the two-mouth polar lobing issue of the Khorn bass bin, and to keep the phase response of the bass bin from going off (i.e., the Khorn bass bin experiences non-flat phase performance at the top of its frequency band).
So while I would be careful about using a dual AMT-1 with wings on top of an old Khorn bass bin, using a miniDSP 2x4 HD to cross the dual AMT-1 with wings was in fact a viable "kit" for those wanting to avoid all the issues of the Heritage top hats (and there are many, not the least of which is that they really don't sound that good to my ears).
That was the proposal that I pitched to Rudy in December 2019 that resulted in the thread that I linked, above.
I still believe that an AMT-1-based top end design using a miniDSP 2x4 HD to dial everything end just blows away the old Klipsch Heritage models in terms of consistent controlled directivity polars, phase response, inherent sound quality, and dialing everything in, which was the underlying reason for that K-forum thread.
I found that the sound of horn-loaded woofers--like La Scalas and Belles, perhaps even Khorn bass bins--would sound a lot better than direct radiating woofers, like Cornwalls and Heresies. I was right (I've used dual-stacked AMT-1s now on top of Belle bass bins for my surround loudspeakers, crossed over with a miniDSP 2x4 HD with two used Crown D75-A amplifiers driving them--leaning the top AMT-1s back about 15 degrees to get better vertical coverage for surround loudspeakers. The only thing that I would recommend to others would be to remove the Belle top hats to lower the AMT-1s closer to the centerline of the bass bin mouths, thus improving the crossover polar coverage performance. Then the bass bin could sit on top of a closed-box riser to raise its AMT-1s back up to seated ear height (or slightly higher).
JMTC.
Chris
Last edited:
Nice work. My only comment is that the transition from the AMT tweeter throat to the horn/waveguide needs to be smooth and free of any small lip/sharp edges. I used modeling putty on my throat transition and made measurements as I applied more and more. The smoother it was in putty, the smoother the measured response was. I found that the horn did not provide any SPL gain to the AMT, only improved directionality and shaping of the response curve. The sound was magical though. One of the best tweeters I have heard.
We measured about a 160 Hz depression of the cutoff frequency with dual AMTs and wings--from 610 Hz down to ~450 Hz (i.e., comparing -3 dB points) relative to a single bare AMT-1 without wings or horn. The real issue is controlling the AMT-1 polars below ~2 kHz--without using a horn or wings. The AMT-1 begins to really lose it at 100 dB/1m on axis due to loss of "horn sidewalls" below 2 kHz.
The dual-high configuration with wings would have been within striking distance of using it on top of a Khorn bass bin...if I hadn't also discovered that the Khorn bass bin crossover point really needs to be ~220 Hz instead. YMMV.
I also found that AMTs need to be broken in for a few days to get the lower end performance that I show. In fact, AMT-1s experience the most change in their SPL response with break-in of any driver that I can remember. I assume it's the diaphragm loosening up the silicone adhesive that glues the voice coil to the pleats, or perhaps it's the base diaphragm material itself that loosens up.
Chris
Hi Chris,
Hi @xrk971,
Kindest regards,
M
I would love to lower the cross-over point, currently at about 620 Hz, to about 380 Hz, because if the data for TAD1601A are to be believed, it is in a piston region up to that frequency. Hence the idea of MEH.
Hi @xrk971,
I have identical experience. My Heil has a raised edge from both the top and the bottom of the case, and the measured response significantly changed, when I filled both with a properly shaped insert.
Kindest regards,
M
So is it fair to summarize lower limit/recommended crossover point as follows:
Single Heil without wings or waveguide = 800-2000hz (depending who you believe)
Single Heil with waveguide = about 620 hz
Double Heil with lateral wings = about 450 hz
There are some posts where people have mused about a triple stack, but it doesn’t seem anyone has actually tried it. Besides cost, perhaps vertical lobes might become a significant problem?
Single Heil without wings or waveguide = 800-2000hz (depending who you believe)
Single Heil with waveguide = about 620 hz
Double Heil with lateral wings = about 450 hz
There are some posts where people have mused about a triple stack, but it doesn’t seem anyone has actually tried it. Besides cost, perhaps vertical lobes might become a significant problem?
Actually, a single bare AMT-1 will cut off around 610 Hz by itself at nominal SPL in-room. It is however limited in SPL output to below ~95-100 dB (1m) due to disproportionately rising harmonic distortion (HD) with SPL below 2 kHz.
The issue is that the AMT-1 runs out of its own sidewall acoustic support below 2 kHz, and the diaphragm essentially becomes direct-radiating below 2 kHz, leading to the disproportionately rising harmonic distortion below 2 kHz as the diaphragm runs out of available travel, and thus linearity.
If you provide added sidewall support length (an added extension of the existing sidewalls), it's good up to at least 105 dB (1m) without disproportionate rise in harmonic distortion.
Adding an AMT-1 on top of another and driven in series will lower the cutoff frequency slightly over the single bare AMT-1, but leaves the issue of rising HD below 2 kHz, albeit at lower HD levels. Adding wings or a full four-sided horn to a double-stacked AMT-1 will lower the cutoff to about 450 Hz (as seen in the plot above).
I believe that I was the source on the triple-stack idea in order to achieve a "shaded array" of three vertical elements that could be set up to approximate a line array to increase vertical polar coverage using DSP, like configuration (f) in this diagram from Toole's book (1st Ed.), figure 18.3:
![post-139669-0-48344300-1427413609[1].jpg](https://www.diyaudio.com/community/attachments/post-139669-0-48344300-1427413609-1-jpg.1277978/ "post-139669-0-48344300-1427413609[1].jpg")
Chris
The issue is that the AMT-1 runs out of its own sidewall acoustic support below 2 kHz, and the diaphragm essentially becomes direct-radiating below 2 kHz, leading to the disproportionately rising harmonic distortion below 2 kHz as the diaphragm runs out of available travel, and thus linearity.
If you provide added sidewall support length (an added extension of the existing sidewalls), it's good up to at least 105 dB (1m) without disproportionate rise in harmonic distortion.
Adding an AMT-1 on top of another and driven in series will lower the cutoff frequency slightly over the single bare AMT-1, but leaves the issue of rising HD below 2 kHz, albeit at lower HD levels. Adding wings or a full four-sided horn to a double-stacked AMT-1 will lower the cutoff to about 450 Hz (as seen in the plot above).
I believe that I was the source on the triple-stack idea in order to achieve a "shaded array" of three vertical elements that could be set up to approximate a line array to increase vertical polar coverage using DSP, like configuration (f) in this diagram from Toole's book (1st Ed.), figure 18.3:
Chris
What do you recommend for a break-in procedure (e.g. frequency, SPL, time duration)?
Just normal music or perhaps pink noise will do, since most of the change in driver output occurs between 600-2500 Hz--the lowest end of its passband (pink noise has amplitude proportional to 1/f). I would be careful about driving the diaphragms at too high a level: something around 96-100 dB/1m is probably sufficient. The duration should probably be something like a few hours instead of merely minutes. The easiest way to know if it's broken-in is when the SPL output stops changing, of course.
You can throw an acoustically absorbing blanket over the whole driver assembly during break-in to attenuate the pink noise SPL, since they are so efficient: about 96 dB/1m @ 2.83V. The voice coil is fairly large in area--not like a circular voice coil driver--and is well ventilated. It doesn't get hot like conventional cone/dome driver voice coils do.
Chris
You can throw an acoustically absorbing blanket over the whole driver assembly during break-in to attenuate the pink noise SPL, since they are so efficient: about 96 dB/1m @ 2.83V. The voice coil is fairly large in area--not like a circular voice coil driver--and is well ventilated. It doesn't get hot like conventional cone/dome driver voice coils do.
Chris
I bought a pair and I would like to ask 1) if you used only side-wings or a complete horn/waveguide 2) has the narrow vertical dispersion been a problem in your applications?
Lack of vertical dispersion and lack of control of horizontal dispersion below 2 kHz seem to be the most common objections. I am asking myself if both aspects could be improved with a waveguide to cross them at around 800 - 1000 Hz. Perhaps a waveguide with horizontal fins?
Because the diaphragm has height (like a line array), it has directivity down to the frequency corresponding to a half wavelength equal to the diaphragm height. So the real issue is that the AMTs all tend to have a fairly narrow vertical coverage angle.
So, down to about 1 kHz for a single AMT-1, it has directivity vertically and doesn't need a constraining horn in the vertical direction.
If you stack two AMT-1s on top of each other, the frequency of loss of directivity control vertically is decreased by one octave (i.e., 500 Hz), so there is a reason to stack them if you want to avoid a full 4-sided horn. If you want to use only one AMT-1, then you're looking at adding top and bottom sides to the horn aperture. If you look at the old Hawthorn Audio (now defunct) "Rainier", you will see a single (custom) AMT with a full 4-sided horn:
![images[1].jpg](https://www.diyaudio.com/community/attachments/images-1-jpg.1365186/ "images[1].jpg")
I've found the way to mitigate the narrower vertical coverage due to stacking two or more AMT-1s is to simply lean the top AMT-1 back about 15-20 degrees to spread the vertical coverage (I used pieces of neoprene from an old mouse pad to form a wedge). This is easy to hear the difference, and you can adjust this angle to cover the vertical directivity needs for your room.
Chris
So, down to about 1 kHz for a single AMT-1, it has directivity vertically and doesn't need a constraining horn in the vertical direction.
If you stack two AMT-1s on top of each other, the frequency of loss of directivity control vertically is decreased by one octave (i.e., 500 Hz), so there is a reason to stack them if you want to avoid a full 4-sided horn. If you want to use only one AMT-1, then you're looking at adding top and bottom sides to the horn aperture. If you look at the old Hawthorn Audio (now defunct) "Rainier", you will see a single (custom) AMT with a full 4-sided horn:
I've found the way to mitigate the narrower vertical coverage due to stacking two or more AMT-1s is to simply lean the top AMT-1 back about 15-20 degrees to spread the vertical coverage (I used pieces of neoprene from an old mouse pad to form a wedge). This is easy to hear the difference, and you can adjust this angle to cover the vertical directivity needs for your room.
Chris
Last edited:
Hi @swak,
for at least the reason stated by @Cask5, I have attached a full horn; the other reason being that I have been contemplating adding mid-range drivers and since I am using the AMT to lower frequency, I need all four surfaces for the correct 1/4 distance between the mid-range drivers.
The lower vertical dispersion does not bother me (too much), as I tend to listen at one spot and not jumping up and down, walking about, etc.
Kindest regards,
M
for at least the reason stated by @Cask5, I have attached a full horn; the other reason being that I have been contemplating adding mid-range drivers and since I am using the AMT to lower frequency, I need all four surfaces for the correct 1/4 distance between the mid-range drivers.
The lower vertical dispersion does not bother me (too much), as I tend to listen at one spot and not jumping up and down, walking about, etc.
Kindest regards,
M
No problem, I only have to find it🙂 I can also send you the (chaotic) F360 project. Just send me a PM with e-mail and I will send it tomorrow when I am at the work computer.
this is what I might try. I believe I am going to cross the AMT to a 414 (12"). If I place the 12" in a box, I might need to choose a higher XO point (around 1.1k) or try a cardioid box with resistive ports at the side. If OB, then I would be able to cross lower and don't worry so much for the horizontal directivity match. BTW, do you have measurements of the horizontal dispersion of the AMT (specially interested in the 800 to 1.2 region)?
Good to hear this. As long as the vertical dispersion does not affect too much the spectral balance of the reflected signal I am fine too with sitting height.
Your horn is larger than the one I would like to try and I understand you had the idea of eventually mounting MF drivers on it. While this is interesting, it is not my plan for now.
My plan is to check the ATH thread and aim at the shortest waveguide that would give horizontal control at the xo frequency. I have not seen many amt waveguides with fins but here is one. No measurement of vertical dispersion effects, and judging from the form, this was probably not the effect that was aimed at.
I haven't even seen a naked Heil AMT so I have no idea how difficult it would be to mount a different WG or if it is even feasible. It will be some time before getting to play with mine.
Hi @swak,
yeah, I saw the wave-guide that you found, but unless I am misreading the plots, the fins do not have much of - if any - effect.
The mounting is not difficult, provided that you can print the wave-guide, and not using the crude manufacturing that I was forced to do not having a printer. Now, that I am satisfied with the result, I am thinking about designing and having printed a wave-guide that will be partially inserted into the AMT since I want to decrease the horizontal dispersion from the default 90 deg.
Please pay attention to @xrk971 post #24, I have exactly the same experience.
I do not know whether ATH can model square wave-guide throat. When I played with ATH, I was unable to figure out how to design a rectangular wave-guide, so if you could advise me after you have modeled yours, I would appreciate it.
Kindest regards,
M
yeah, I saw the wave-guide that you found, but unless I am misreading the plots, the fins do not have much of - if any - effect.
The mounting is not difficult, provided that you can print the wave-guide, and not using the crude manufacturing that I was forced to do not having a printer. Now, that I am satisfied with the result, I am thinking about designing and having printed a wave-guide that will be partially inserted into the AMT since I want to decrease the horizontal dispersion from the default 90 deg.
Please pay attention to @xrk971 post #24, I have exactly the same experience.
I do not know whether ATH can model square wave-guide throat. When I played with ATH, I was unable to figure out how to design a rectangular wave-guide, so if you could advise me after you have modeled yours, I would appreciate it.
Kindest regards,
M
I've got a couple of plots of this with fairly noisy in-room measurements--without external wings or horn to constrain the output horizontally. The first is a normalized polar sonogram, the second is the same data in a normalized polar plot:
Chris
Thank you Chris. Quite constant 70 deg. dispersion down to 2.4 k, below that almost no control.
It seem it can (link).
I have no idea how the AMT default waveguide is assembled, do you mean it is not difficult mounting it on the original waveguide or on the "naked" driver?
Here is another WG for the Mundorf. Nice horizontal dispersion down to 1k and minimal size. This is what I would like to achieve.
The designer participates in this forum (NicoB).
I am starting to think that even using a 12" mid I would need a small waveguide for it to match the directivity of the AMT if crossing low (800-1.1k). Perhaps the simplest solution would be to cross higher, given that the 12" seems to handle that well.
It seem it can (link).
I have no idea how the AMT default waveguide is assembled, do you mean it is not difficult mounting it on the original waveguide or on the "naked" driver?
Here is another WG for the Mundorf. Nice horizontal dispersion down to 1k and minimal size. This is what I would like to achieve.
The designer participates in this forum (NicoB).
I am starting to think that even using a 12" mid I would need a small waveguide for it to match the directivity of the AMT if crossing low (800-1.1k). Perhaps the simplest solution would be to cross higher, given that the 12" seems to handle that well.
Last edited:
Hi @swak,
the mounting depends on the type of AMT you have. If you look at the pictures in post #19, you can see that the AMT has flat top and bottom with two through holes. I attached rectangular plates with holes matching the holes in the AMT to the wave-guide, and used crudely made ex-centers to secure the plates to the AMT, see the attached picture.
What I plan to do is print the wave-guide and use the front surfaces with some non-hardening glue (silicon based?) to adhere them together. Search for @pelanj, who made an interface for a similar AMT.
Kindest regards,
M
the mounting depends on the type of AMT you have. If you look at the pictures in post #19, you can see that the AMT has flat top and bottom with two through holes. I attached rectangular plates with holes matching the holes in the AMT to the wave-guide, and used crudely made ex-centers to secure the plates to the AMT, see the attached picture.
What I plan to do is print the wave-guide and use the front surfaces with some non-hardening glue (silicon based?) to adhere them together. Search for @pelanj, who made an interface for a similar AMT.
Kindest regards,
M
