I rolled in my active XO, LR4 @ 1kHz. There's some upper mid suckout (not a lot) that I think I can improve by increasing the value of the tweeter protection cap (or probably by swapping 3rd order filter boards into the HP, probably a better solution now that I think of it).
Short version... the speakers are much happier. The lower distortion and cleaner mids are immediately apparent.
Short version... the speakers are much happier. The lower distortion and cleaner mids are immediately apparent.
Watching this with interest. I may soon be messing around with horns.
Here is a picture of my Altec/Heil DIY speakers. The large MLTL cabinet houses an Altec 416b. Altec 811 midhorn w/802g CD, Heil AMT mounted on top. Crossover points are 1100 hz, 8 Khz. The speaker is actively managed and triamped.
My alignment would probably work well with a horn laded Heil.
Here is a picture of my Altec/Heil DIY speakers. The large MLTL cabinet houses an Altec 416b. Altec 811 midhorn w/802g CD, Heil AMT mounted on top. Crossover points are 1100 hz, 8 Khz. The speaker is actively managed and triamped.
My alignment would probably work well with a horn laded Heil.
The nice part about a heil, particularly a hornloaded one, is that you can get a clean 1-20kHz out of it- very few horns can manage that, even my Oblate Spheroid Waveguides with JBL 2426h drivers cannot cleanly do much more than 12kHz- diaphragm breakup becomes an issue. Go smaller phragm, and you run into problems with bottom end. Beryllium helps the issue though- wish I could afford some of those TruExtent phragms.
Hi badman,
When I first started reading this thread I got a strong sense of déjà vu.
I bought my pair of ESS AMT Towers (Bailey style1 TL 10”with a ‘great’ Heil) back in the seventies. After a while, I became unhappy with the bass/midrange of the system and started a long convoluted process of trying to fix the problem. I won’t go into all the gory details but one of the speaker arrangements was a JBL 2231 15” pro series woofer in an approx 5 Cu Ft Theile B6 alignment with LR4 active crossover (crossed at frequencies between 600 and 1k but mostly 800) and some compensation filters (in those days, a lot of NE5534ANs). I also had tried modifying the Heils with ‘waveguides’ and baffles in an attempt to fill the hole in the response curve and fix up some other problems I thought were caused by the Heil’s shape and construction.
Even though my waveguides were somewhat different from yours (smaller for a start) I think you can appreciate the Déjà vu when I saw your speakers. However, in the end, I abandoned this approach and went off into open baffle designs. In those days open baffles had not yet come back into vogue. I actually went to open baffles initially not so much for the dipole radiation pattern as to eliminate box resonances and radiation from the box panels. Again the details are not that important but once I heard the Heils combined with OB dynamics I never turned back. <shrug>. In the end, I ended up with a tri-amped system: end fire binomial array with 15” woofers (what Celestion would some years later market in the SL6000 system as double dipoles) for the bass, an 8” poly on small simple square baffle for the mid, Heil with relatively small contoured ‘waveguide’.
I do not want to put a dampener on your enthusiasm for your system so I will say that I thought the JBL Heil combo sounded quite good and perhaps it was premature of me to abandon them ( I didn’t retry them later as both JBL 2231s came to a rather ridiculous demise). Nevertheless, I will give you my thoughts in order to add grist to the mill.
The JBL cones are quite good in their higher registers but they do not seem to match that well with the Heils. My problem with my own statement was that I was never sure how much of the problem was caused by the drivers and how much caused by the boxes. I did try the JBLs quite briefly in the OBs but they were totally unsuited to the role and too valuable to risk. As such I did not really assess them in OBs long enough to form a valid opinion. I usually do this by listened to all my best test tracks (remember, it was the 70s and 80s – Vinyl and all that stuff) with just the bass speaker connected, box problems become obvious after a while. I couldn’t do this with the JBLs due to excursion issues and the sub-sonic content on some of the vinyl.
If the Heils cross below 1KHz they just don’t sound right. My experiments with my waveguides really did not seem to fix this problem although they might have helped a little. The waveguides were really there to get rid of some of the acoustic issues already discussed including the discontinuity in the acoustic impedance of the device. Of course your waveguides are much bigger than mine so they may increase the acoustic impedance in the lower register enough to solve the problem. However, my conclusion from my experimenting was that the diaphragm just wasn’t big enough to support these lower frequencies. In any case 1KHz is a terrible place to cross.
I thought about true horn loading the Heils after I saw the old Speaker Builder article when it was originally published but I decided the the Heil’s Bil was just too low. I had seen a set of the ESS Blue Ox horns so I knew how massive the magnets would have to be for true horn performance (I later tried to track them down but failed – probably land fill or mouldering away in some garage).
In the end I decided that the only solution was to build 7ft line source Heils to allow a 200Hz crossover frequency. I even managed to design them, the weight was astronomical. In any case the real world derailed all this stuff so nothing was done.
A few final comments and suggestions.
In some of your photos the Heil is upside down and back to front, is there a reason for this?
I noticed that your Heils have the V shaped cut-outs in the top only, was this standard in some models? Mine are the old fully rectangular boxes, I filled in the space between the plastic box and the pole piece to get rid of the ridge and possible disruption of the vertical wavefront. I noticed that later Heils had the V cut-out both top and bottom which is a better solution and completely eliminates any possible problem.
Even though it was a pain to do, I was much happier when I extended the waveguide all the way down the pole piece in a more classical horn shape; I believe it made a subtle improvement to the sound. Of course, as with all such things, I could have been deceiving myself by wishful thinking.
The diaphragms are not very securely mounted in the ESSs nor are they airtight. This is not a problem in normal operation but if you try to operate them at a substantially lower frequency...
If it is not against your audio philosophy and or budget I would seriously consider digital equalisation and crossover.
If you read Oskar Heil’s patent 3,832,499 starting column 5 line 40 and ending column 6 line 60 you will see an aspect of the acoustic circuit that is rarely discussed. Despite the quaint language it is clear that Dr Heil is very enthusiastic about this part of the design. His explanation does indeed sound plausible but I have long ago learned not to over rely on my intuition. Against this explanation is the later pole piece design where ESS changed from the laminations to a cast pole piece. The shape of the new pole piece is quite different and cannot IMO be explained by the requirements of the magnetic circuit. Further, it is at odds with the explanation of the acoustic design in the patent so if the patent is correct why did ESS change the shape to a more complex one?. In any case, I do not believe that you can treat the wavefront in the ‘throat’ as planar. In post 10 of this thread mige0 stated “ There are pretty nice applets out there to visualize (more or less) how the released wavefront of the Heil might look like…”, I for one would like to know what they are. The simplest view I have of the diaphragm (in cross section orthogonal to the pleats) is a multi element (say 10or so), in phase, constant amplitude, broadside array of planar surfaces, in other words a multi-element line source, this does not take into account any influence of the pleat geometry, e.g. pleat depth. AFAIK, even Beyma has not done a full acoustic simulation of the AMT; they have barely touched the surface investigating modal behavior using a mechanical model of the pleat.
In my opinion, I do not believe that any of the modern Heils have made great advances on the old ESS models. Perhaps Beyma but even then… Sure, a lot of them are better made and use rare earth magnets but they all use essentially the same diaphragm with little to no difference compared to the ESSs (Beyma have made changes but are they that much better?). None of them seem to measure any better that the original ESS units. I can sort of remember the test results from a test done on ESSs in the 70s by “Electronics Today International” Australia and the Beymas appear no better. Mind you, I have not been able to find any modern truly reliable testing of the ESSs nor good CSD waterfall plots for any of them. Anyhow, I am not in a position to audition any of the drivers in a meaningful way or at all for that matter.
Ooops, I have rambled on for far too long, happy listening,
Bob
When I first started reading this thread I got a strong sense of déjà vu.
I bought my pair of ESS AMT Towers (Bailey style1 TL 10”with a ‘great’ Heil) back in the seventies. After a while, I became unhappy with the bass/midrange of the system and started a long convoluted process of trying to fix the problem. I won’t go into all the gory details but one of the speaker arrangements was a JBL 2231 15” pro series woofer in an approx 5 Cu Ft Theile B6 alignment with LR4 active crossover (crossed at frequencies between 600 and 1k but mostly 800) and some compensation filters (in those days, a lot of NE5534ANs). I also had tried modifying the Heils with ‘waveguides’ and baffles in an attempt to fill the hole in the response curve and fix up some other problems I thought were caused by the Heil’s shape and construction.
Even though my waveguides were somewhat different from yours (smaller for a start) I think you can appreciate the Déjà vu when I saw your speakers. However, in the end, I abandoned this approach and went off into open baffle designs. In those days open baffles had not yet come back into vogue. I actually went to open baffles initially not so much for the dipole radiation pattern as to eliminate box resonances and radiation from the box panels. Again the details are not that important but once I heard the Heils combined with OB dynamics I never turned back. <shrug>. In the end, I ended up with a tri-amped system: end fire binomial array with 15” woofers (what Celestion would some years later market in the SL6000 system as double dipoles) for the bass, an 8” poly on small simple square baffle for the mid, Heil with relatively small contoured ‘waveguide’.
I do not want to put a dampener on your enthusiasm for your system so I will say that I thought the JBL Heil combo sounded quite good and perhaps it was premature of me to abandon them ( I didn’t retry them later as both JBL 2231s came to a rather ridiculous demise). Nevertheless, I will give you my thoughts in order to add grist to the mill.
The JBL cones are quite good in their higher registers but they do not seem to match that well with the Heils. My problem with my own statement was that I was never sure how much of the problem was caused by the drivers and how much caused by the boxes. I did try the JBLs quite briefly in the OBs but they were totally unsuited to the role and too valuable to risk. As such I did not really assess them in OBs long enough to form a valid opinion. I usually do this by listened to all my best test tracks (remember, it was the 70s and 80s – Vinyl and all that stuff) with just the bass speaker connected, box problems become obvious after a while. I couldn’t do this with the JBLs due to excursion issues and the sub-sonic content on some of the vinyl.
If the Heils cross below 1KHz they just don’t sound right. My experiments with my waveguides really did not seem to fix this problem although they might have helped a little. The waveguides were really there to get rid of some of the acoustic issues already discussed including the discontinuity in the acoustic impedance of the device. Of course your waveguides are much bigger than mine so they may increase the acoustic impedance in the lower register enough to solve the problem. However, my conclusion from my experimenting was that the diaphragm just wasn’t big enough to support these lower frequencies. In any case 1KHz is a terrible place to cross.
I thought about true horn loading the Heils after I saw the old Speaker Builder article when it was originally published but I decided the the Heil’s Bil was just too low. I had seen a set of the ESS Blue Ox horns so I knew how massive the magnets would have to be for true horn performance (I later tried to track them down but failed – probably land fill or mouldering away in some garage).
In the end I decided that the only solution was to build 7ft line source Heils to allow a 200Hz crossover frequency. I even managed to design them, the weight was astronomical. In any case the real world derailed all this stuff so nothing was done.
A few final comments and suggestions.
In some of your photos the Heil is upside down and back to front, is there a reason for this?
I noticed that your Heils have the V shaped cut-outs in the top only, was this standard in some models? Mine are the old fully rectangular boxes, I filled in the space between the plastic box and the pole piece to get rid of the ridge and possible disruption of the vertical wavefront. I noticed that later Heils had the V cut-out both top and bottom which is a better solution and completely eliminates any possible problem.
Even though it was a pain to do, I was much happier when I extended the waveguide all the way down the pole piece in a more classical horn shape; I believe it made a subtle improvement to the sound. Of course, as with all such things, I could have been deceiving myself by wishful thinking.
The diaphragms are not very securely mounted in the ESSs nor are they airtight. This is not a problem in normal operation but if you try to operate them at a substantially lower frequency...
If it is not against your audio philosophy and or budget I would seriously consider digital equalisation and crossover.
If you read Oskar Heil’s patent 3,832,499 starting column 5 line 40 and ending column 6 line 60 you will see an aspect of the acoustic circuit that is rarely discussed. Despite the quaint language it is clear that Dr Heil is very enthusiastic about this part of the design. His explanation does indeed sound plausible but I have long ago learned not to over rely on my intuition. Against this explanation is the later pole piece design where ESS changed from the laminations to a cast pole piece. The shape of the new pole piece is quite different and cannot IMO be explained by the requirements of the magnetic circuit. Further, it is at odds with the explanation of the acoustic design in the patent so if the patent is correct why did ESS change the shape to a more complex one?. In any case, I do not believe that you can treat the wavefront in the ‘throat’ as planar. In post 10 of this thread mige0 stated “ There are pretty nice applets out there to visualize (more or less) how the released wavefront of the Heil might look like…”, I for one would like to know what they are. The simplest view I have of the diaphragm (in cross section orthogonal to the pleats) is a multi element (say 10or so), in phase, constant amplitude, broadside array of planar surfaces, in other words a multi-element line source, this does not take into account any influence of the pleat geometry, e.g. pleat depth. AFAIK, even Beyma has not done a full acoustic simulation of the AMT; they have barely touched the surface investigating modal behavior using a mechanical model of the pleat.
In my opinion, I do not believe that any of the modern Heils have made great advances on the old ESS models. Perhaps Beyma but even then… Sure, a lot of them are better made and use rare earth magnets but they all use essentially the same diaphragm with little to no difference compared to the ESSs (Beyma have made changes but are they that much better?). None of them seem to measure any better that the original ESS units. I can sort of remember the test results from a test done on ESSs in the 70s by “Electronics Today International” Australia and the Beymas appear no better. Mind you, I have not been able to find any modern truly reliable testing of the ESSs nor good CSD waterfall plots for any of them. Anyhow, I am not in a position to audition any of the drivers in a meaningful way or at all for that matter.
Ooops, I have rambled on for far too long, happy listening,
Bob
Wow! That's a heck of a ramble you've got there!
As mine are large enough to load fully to below 1kHz, I'm not noticing any strain in playing them that low. They are actively crossed over at 1k, with 3rd order transfer function, and an overlaid 1st order that corrects the rising response the reinforcement provides with decreasing frequency.
They are oriented the way they are to minimize diffraction artifacts from the frame, as you described, later models did away with the overhanging bits.
Extending into the throat would be a bigger change to them- mine are more of a waveguide than a horn, I wanted to specifically reinforce the lower end of response, and eliminate some of the diffraction ripple from the hard edge of the assembly by extending the angled section out and making a nice soft transition to fullspace radiation by 400hz or so.
I disagree that 1kHz is a terrible place to cross.
I've tried the heil with several 15"s, the JBL 2226h, the Eminence Magnum15HO, and the BD Designs BD15. My preference is for the magnum, based on my listening thus far. It has the best bass alignment in my cabs and seems more dynamic than the impressive 2226h. The BD15 is overdamped for my usage in the bass.
My eventual goal is to incorporate a nude dipole 6.5" midrange section from about 500-1kHz. I need to run some sims but am hoping that I can get relatively flat performance in that range from a nude (providing the best constant directivity performance) transducer. Once I've developed that midrange section I can test incorporating dipole bass vis a vis W-frames, which are already built, but will likely require more shaping of the midrange as I'll need to run the mids lower.
Thanks for your thoughts!
As mine are large enough to load fully to below 1kHz, I'm not noticing any strain in playing them that low. They are actively crossed over at 1k, with 3rd order transfer function, and an overlaid 1st order that corrects the rising response the reinforcement provides with decreasing frequency.
They are oriented the way they are to minimize diffraction artifacts from the frame, as you described, later models did away with the overhanging bits.
Extending into the throat would be a bigger change to them- mine are more of a waveguide than a horn, I wanted to specifically reinforce the lower end of response, and eliminate some of the diffraction ripple from the hard edge of the assembly by extending the angled section out and making a nice soft transition to fullspace radiation by 400hz or so.
I disagree that 1kHz is a terrible place to cross.
I've tried the heil with several 15"s, the JBL 2226h, the Eminence Magnum15HO, and the BD Designs BD15. My preference is for the magnum, based on my listening thus far. It has the best bass alignment in my cabs and seems more dynamic than the impressive 2226h. The BD15 is overdamped for my usage in the bass.
My eventual goal is to incorporate a nude dipole 6.5" midrange section from about 500-1kHz. I need to run some sims but am hoping that I can get relatively flat performance in that range from a nude (providing the best constant directivity performance) transducer. Once I've developed that midrange section I can test incorporating dipole bass vis a vis W-frames, which are already built, but will likely require more shaping of the midrange as I'll need to run the mids lower.
Thanks for your thoughts!
Hi badman,
What I should have said is that I always found the 1Khz area a difficult place to crossover.
Back then (about 1978-1980, I think) there was no simulation; there was no computers to simulate it on short of mainframes and mini-computers (Richard Small did his work on an analogue computer!). Most of the designs were done using tables, graphs, nomograms, hand held calculators and paper. Even this is not completely accurate, my designs usually started with some known results (for OBs, I used Harry Olsen’s work with some input from Gilbert Briggs’ book) and scaled accordingly.
My midrange was an 8 inch poly coned driver mounted in a 10 inch square baffle with rounded corners. The driver only had a pressed basket and the baffle and other bracing had a secondary use as an attempt to eliminate any resonance present in the basket of the driver. A 6.5 inch should work well in the range you are looking at. I did consider 6.5s but the choice was limited and I got the 8s at a very good price point -they even sounded good.
Ahhh, W baffles - a rather large version of the AMT itself. The W baffle is a sort of binomial broadside array. I did look into this style of OB at the time however I shied away as there was a fair bit of cabinet involved (panel resonance?) and I was unsure of the effect of the equivalent of pleat depth. For a 15ich driver, the ‘pleat’ depth would be about 20 inches which is the quarter wavelength of the first resonance (does it exist – can it be modelled as a sort of cylindrical horn?), so resonance at about 170hz which was inside my intended pass-band
The endfire version I used had the advantage of a much simpler construction but at the price of lower max SPL. At the time, I designed it, I was studying antennas at University and the endfire binomial array just sort of leapt off the page – it was probably the only array that was simple enough to implement with a few woofers. I did try a midrange version of my bass system but I did not think it as good as the simple single baffle.
My modification to the Heil was as I said not really a horn, as far as I could tell it did not increase the efficiency in any meaningful way. I cast the waveguide up out of plaster of Paris, cloth and two-part foam. The flair was sort of freehand and vaguely resembled a Tractrix contour in the horizontal plane before curving around the side.
Interestingly, when I again read the AMT patent I noticed that nowhere does Dr Heil mention any horn loading or waveguide effect due to the shape of the magnet structure. Given that this is a patent, I would assume that if he saw any merit in a horn loading effect he would have included it. Also, his description of the acoustics of the AMT make the diaphragm appear as a point (actually line) source located in the space between the pole pieces and forward of the slots – not at all like the acoustic radiation pattern in the throat of a normal horn.
Time to stop rambling on
Regards,
Bob
What I should have said is that I always found the 1Khz area a difficult place to crossover.
Back then (about 1978-1980, I think) there was no simulation; there was no computers to simulate it on short of mainframes and mini-computers (Richard Small did his work on an analogue computer!). Most of the designs were done using tables, graphs, nomograms, hand held calculators and paper. Even this is not completely accurate, my designs usually started with some known results (for OBs, I used Harry Olsen’s work with some input from Gilbert Briggs’ book) and scaled accordingly.
My midrange was an 8 inch poly coned driver mounted in a 10 inch square baffle with rounded corners. The driver only had a pressed basket and the baffle and other bracing had a secondary use as an attempt to eliminate any resonance present in the basket of the driver. A 6.5 inch should work well in the range you are looking at. I did consider 6.5s but the choice was limited and I got the 8s at a very good price point -they even sounded good.
Ahhh, W baffles - a rather large version of the AMT itself. The W baffle is a sort of binomial broadside array. I did look into this style of OB at the time however I shied away as there was a fair bit of cabinet involved (panel resonance?) and I was unsure of the effect of the equivalent of pleat depth. For a 15ich driver, the ‘pleat’ depth would be about 20 inches which is the quarter wavelength of the first resonance (does it exist – can it be modelled as a sort of cylindrical horn?), so resonance at about 170hz which was inside my intended pass-band
The endfire version I used had the advantage of a much simpler construction but at the price of lower max SPL. At the time, I designed it, I was studying antennas at University and the endfire binomial array just sort of leapt off the page – it was probably the only array that was simple enough to implement with a few woofers. I did try a midrange version of my bass system but I did not think it as good as the simple single baffle.
My modification to the Heil was as I said not really a horn, as far as I could tell it did not increase the efficiency in any meaningful way. I cast the waveguide up out of plaster of Paris, cloth and two-part foam. The flair was sort of freehand and vaguely resembled a Tractrix contour in the horizontal plane before curving around the side.
Interestingly, when I again read the AMT patent I noticed that nowhere does Dr Heil mention any horn loading or waveguide effect due to the shape of the magnet structure. Given that this is a patent, I would assume that if he saw any merit in a horn loading effect he would have included it. Also, his description of the acoustics of the AMT make the diaphragm appear as a point (actually line) source located in the space between the pole pieces and forward of the slots – not at all like the acoustic radiation pattern in the throat of a normal horn.
Time to stop rambling on
Regards,
Bob
Hi,
Please ignore all the descriptions of the bass dipoles, they are very confused, the terminology is wrong and it would take too much now to straighten it out. Since I can't seem to edit it (even though my account says I should), it will have to stand.
I had a quick look at Linkwitz site and the "pleat depth resonance" is there although he equalizes it out with a notch filter.
The stuff about the mid dipole and the Heil is still a roughly correct description.
My apologies to you and all who read this thread -. a very senior moment. Sometimes it pays to just stay away from the keyboard.
Bob
Please ignore all the descriptions of the bass dipoles, they are very confused, the terminology is wrong and it would take too much now to straighten it out. Since I can't seem to edit it (even though my account says I should), it will have to stand.
I had a quick look at Linkwitz site and the "pleat depth resonance" is there although he equalizes it out with a notch filter.
The stuff about the mid dipole and the Heil is still a roughly correct description.
My apologies to you and all who read this thread -. a very senior moment. Sometimes it pays to just stay away from the keyboard.
Bob
Last edited:
Annnnd...... Here we go! The heils are veneered and have their counterpart 6.5" dipole mids.
An externally hosted image should be here but it was not working when we last tested it.
Very nice looking product Badman. Congrats on this design and your implementation of it. Does this end the project or will you be taking it further?
Listening to: Orchard by Phil Sheeran
Thetubeguy1954 (Tom Scata)
Thanks! You can see the felt is cut and placed so as not to interfere with the forward lobe from the drivers, but still offer some absorbtion.
The project is not yet done, I have dipole midbasses in M-frames to implement, but have not done so as yet. The midrange is greatly improved with the addition of the mids, as one might have hoped. Looking forward to continued play with these but hate to roll the midbass cabs out of the living room
Thanks! You can see the felt is cut and placed so as not to interfere with the forward lobe from the drivers, but still offer some absorbtion.
The project is not yet done, I have dipole midbasses in M-frames to implement, but have not done so as yet. The midrange is greatly improved with the addition of the mids, as one might have hoped. Looking forward to continued play with these but hate to roll the midbass cabs out of the living room
What about where the Heil is sitting on top of the woofer's cabinet? Are you or have you tried placing some felt, flat & directly on top of the triangular wood surface that is exposed right in front of the Heil? I would think you'd probably have a lot of the sound coming from the Heil reflecting off of that exposed wood surface in front of it, no?
Listening to: Orchard by Phil Sheeran
Thetubeguy1954 (Tom Scata)
I've used felt there, but since it's a smooth surfact it's not the scattering challenge that the tops of the horn structures represent to the midrange. It's just the bottom wall of the horn. I may experiment with damping that surface as well, again, sometime.
Remember, it's not THAT you encounter a boundary that's the issue, but HOW you encounter the boundary.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.