Hi
So I start my thread now. For years I have been playing around with dipole concepts, just for fun, always with the intention to find some solutions and to feed my curiosity. Learning by doing. Learning by trial and error. Learning by sawing and welding, measing and reading. Like a real hobbyist. And learning by sharing ... and getting answers, like this one:
Thanks for this one. I never considered ETC, and indeed, having peeped into it, it seems to be another, for me new and useful information well. And I come to this forum in a hope to share and tap new infos with and from another community than the one I met at www.aktives-hoeren.de, where I was active for some years until now.
DIY-wise, in my shack there actually is a heap of Yamaha AX amplifiers to feed my ongoing-neverending multiway dipole project, some linux-friendly M-audio Delta 1010's soundcards, some drivers, some books, some laptops and the like, but no really operative system. My ongoing project is never operative, because it is in steady transformation ...The last release of a really operative system occurred some years ago, still functional, and since then, there is production-wise nothing happening but joyful construction time again.
For the measurements I use Uli Brueggemann's fantastic Acourate Sofware, for rapid prototyping of Xo's it's Acourate Convolver counterpiece on an old W8.1 Laptop, along with outdated RME Multiface ADDA's
So? This is my actual concept project: A 5- to 6- way dipole. Mutations occur nearly every week-end.
When finished, which I hope will never be the case, there will be two units per channel: A woofer two-way unit 60-170/170-285, and a mid-treble three-way unit 285-710/710-2200/2200. Plus maybe a subwoofer elswhere in the room, where it's best suited. Specs and distances from driver to driver are chosen in order to perform the transitions with lobing <1dB (except for the highmid-tweeter xover), and for a target SPL of 110dB@f_crossover@Xmax/2. Not @ Xmax, but Xmax/2, in order to keep distortion low, even at a theoretical volume-displacement-limited SPL of 116dB for the whole system.
So, basically that's it. Little bit crazy, and maybe welcomed weird (as a forum's member wrote). Might be. But no worries. In my professional life I am a shrink. That's maybe one of the reasons for embracing crazy things.
So I start my thread now. For years I have been playing around with dipole concepts, just for fun, always with the intention to find some solutions and to feed my curiosity. Learning by doing. Learning by trial and error. Learning by sawing and welding, measing and reading. Like a real hobbyist. And learning by sharing ... and getting answers, like this one:
Thanks for this one. I never considered ETC, and indeed, having peeped into it, it seems to be another, for me new and useful information well. And I come to this forum in a hope to share and tap new infos with and from another community than the one I met at www.aktives-hoeren.de, where I was active for some years until now.
DIY-wise, in my shack there actually is a heap of Yamaha AX amplifiers to feed my ongoing-neverending multiway dipole project, some linux-friendly M-audio Delta 1010's soundcards, some drivers, some books, some laptops and the like, but no really operative system. My ongoing project is never operative, because it is in steady transformation ...The last release of a really operative system occurred some years ago, still functional, and since then, there is production-wise nothing happening but joyful construction time again.
For the measurements I use Uli Brueggemann's fantastic Acourate Sofware, for rapid prototyping of Xo's it's Acourate Convolver counterpiece on an old W8.1 Laptop, along with outdated RME Multiface ADDA's
So? This is my actual concept project: A 5- to 6- way dipole. Mutations occur nearly every week-end.
An externally hosted image should be here but it was not working when we last tested it.
When finished, which I hope will never be the case, there will be two units per channel: A woofer two-way unit 60-170/170-285, and a mid-treble three-way unit 285-710/710-2200/2200. Plus maybe a subwoofer elswhere in the room, where it's best suited. Specs and distances from driver to driver are chosen in order to perform the transitions with lobing <1dB (except for the highmid-tweeter xover), and for a target SPL of 110dB@f_crossover@Xmax/2. Not @ Xmax, but Xmax/2, in order to keep distortion low, even at a theoretical volume-displacement-limited SPL of 116dB for the whole system.
So, basically that's it. Little bit crazy, and maybe welcomed weird (as a forum's member wrote). Might be. But no worries. In my professional life I am a shrink. That's maybe one of the reasons for embracing crazy things.
Moved to multi-way forum
Right! This baffle particularity is a meaningless leftover from a former iterative step ...
An externally hosted image should be here but it was not working when we last tested it.
This version was a failure due to gross vertical lobing of the low midrange drivers at the transition to the high midrange. So I ommitted one or the lowmid drivers to minimize vertical dimensions.
In order to
1. maintain the target spl specs ...
2. improve lobing between the W-Bass and the low-mid driver ...
I added a 5th way, some kind of filler upper-bass 12" driver.
So the (old) baffle in the new interative step still has 7"-spaced slots, whereas the new stand of the 12" has it's tubes spaced by 10". There soon will be a more consequent baffle for the W-Bass in order to improve it's efficiency. This is now possible due to the lower xover between w-bass and upper-bass (170Hz)
Horbach-Keele and the ground reflection
Now for something completely different ... Having left Horbach-Keele-Designs (HK) aside some time ago because of the practical limitations inherent within the top frequency range (size of drivers, spl), I recently took another approach with Dipole-HK in order to minimize early reflections from room boundaries. Within this iteration, I try to deal with the floor (and ceiling) reflection.
HK theoretically provides constant directivity within the vertical plane. By choosing the appropriate parameters for the filters, it should be possible to design an array such as a lobing null is directly headed towards the point(s) of reflection between the center of the speaker and the hearing position, located at the floor (and/or the ceiling). With such an array, consequently no sound will be fed into this reflections at all.
Figure e) e.g. shows a configuration, where the null occurs at approx. 55°. Designing the system for a real-world situation with the central tweeter at 128cm, a listening distance of 215cm and an ear height at 128cm, the needed angle of the null would be at approx. 50°. This transform into D/Lambda=0.65 and a beamwidth (BW) of 62° for the design of the HK-parameters.
This system should be considered as a very early prototype, and so are it's filters. It is only a partial HK-System, because of the three pairs of drivers provide constant directivity only in the range of approx. 200Hz...1100Hz. Therefore, the first SPL-cancellation due to the ground reflection, which in this setup will occur at 144Hz, cannot be avoided. To circumvent this dip in a final version, the single woofer could cross over at some 150Hz ... 170Hz in order to provide some spl of -6db ... -4dB pressure at 144Hz.
Now, does this theoretical approach work? For a first measurement, accordingly to HK the full array was fed (red). For a second measurement, the upper half of the array was disconnected. So only the bottom half of the array was fed, making up for some kind of a "standard" setup with the bass drivers low above ground, up to the tweeter at ear height (green).
As mentionned before, red trace is HK, green trace is the standard-system. The marked section of the pulse response shows both ground reflections artefacts (the spike at 0.0682 is an unavoidable room artefact which has nothing to do with the ground/ceiling).
HK seems to work as expected ... That's good news for the search of minimizing early room artefacts (and so are dipoles, anyway).
Now for something completely different ... Having left Horbach-Keele-Designs (HK) aside some time ago because of the practical limitations inherent within the top frequency range (size of drivers, spl), I recently took another approach with Dipole-HK in order to minimize early reflections from room boundaries. Within this iteration, I try to deal with the floor (and ceiling) reflection.
HK theoretically provides constant directivity within the vertical plane. By choosing the appropriate parameters for the filters, it should be possible to design an array such as a lobing null is directly headed towards the point(s) of reflection between the center of the speaker and the hearing position, located at the floor (and/or the ceiling). With such an array, consequently no sound will be fed into this reflections at all.
An externally hosted image should be here but it was not working when we last tested it.
Figure e) e.g. shows a configuration, where the null occurs at approx. 55°. Designing the system for a real-world situation with the central tweeter at 128cm, a listening distance of 215cm and an ear height at 128cm, the needed angle of the null would be at approx. 50°. This transform into D/Lambda=0.65 and a beamwidth (BW) of 62° for the design of the HK-parameters.
An externally hosted image should be here but it was not working when we last tested it.
This system should be considered as a very early prototype, and so are it's filters. It is only a partial HK-System, because of the three pairs of drivers provide constant directivity only in the range of approx. 200Hz...1100Hz. Therefore, the first SPL-cancellation due to the ground reflection, which in this setup will occur at 144Hz, cannot be avoided. To circumvent this dip in a final version, the single woofer could cross over at some 150Hz ... 170Hz in order to provide some spl of -6db ... -4dB pressure at 144Hz.
Now, does this theoretical approach work? For a first measurement, accordingly to HK the full array was fed (red). For a second measurement, the upper half of the array was disconnected. So only the bottom half of the array was fed, making up for some kind of a "standard" setup with the bass drivers low above ground, up to the tweeter at ear height (green).
An externally hosted image should be here but it was not working when we last tested it.
As mentionned before, red trace is HK, green trace is the standard-system. The marked section of the pulse response shows both ground reflections artefacts (the spike at 0.0682 is an unavoidable room artefact which has nothing to do with the ground/ceiling).
HK seems to work as expected ... That's good news for the search of minimizing early room artefacts (and so are dipoles, anyway).
Stumbled on this very interesting thread searching for diy horbach keele implementations, sadly all the pictures in this thread seem to be dead (could you perhaps upload them again so the reader can better understand the system and the story?).
In your last post you mention leaving HK because of practical limitations with the top frequency range, as someone concerned about the same problem, what was the best/closest solution you found for this problem and what were issues? Also very interested in seeing/hearing about what the final system looked like and if you're still pursuing HK Dipoles or found something better!
In your last post you mention leaving HK because of practical limitations with the top frequency range, as someone concerned about the same problem, what was the best/closest solution you found for this problem and what were issues? Also very interested in seeing/hearing about what the final system looked like and if you're still pursuing HK Dipoles or found something better!