-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
"Analysis" of cardioid speaker radiation via lateral slots - like D&D 8c
- Thread starter ctrl
- Start date
Biblob
Addicted to Fun and Learning
- Joined
- Sep 13, 2018
- Messages
- 641
- Likes
- 614
Biblob
Addicted to Fun and Learning
- Joined
- Sep 13, 2018
- Messages
- 641
- Likes
- 614
@ctrl regarding the R2 speaker-project, do you think using a cardioid design will negatively impact it's max SPL, using only a 5" inch driver? Could it reach 105 dB at 1m at 200hz, still?
Even when crossed around 200 hz it will reach it max excursion pretty quickly, I'd think. Erin found out that these SB Acoustics drivers don't really meet spec regarding their x-max.
Even when crossed around 200 hz it will reach it max excursion pretty quickly, I'd think. Erin found out that these SB Acoustics drivers don't really meet spec regarding their x-max.
OP
- Thread Starter
- #44
This should not be a problem at all as far as drive excursion is concerned. However, the voice coil is unlikely to withstand a sound pressure level of 105dB for long.
As a rough estimate for the excursion and load of the 5'' driver at 105dB@1m free field:
1. Here is the diffraction simulation of the Directiva r2 baffle as open baffle (the excursion of the drivers in speakers with slots is about the same as OB speaker):
2. The combination of half-space frequency response of the driver on an infinite baffle and the diffraction correction to simulate free field conditions and an LR4@200Hz high pass, results for 105dB@1m free field (used some 5'' driver, do not know off the top of my head which SB Accoustic driver it is):
A high thermal load for the voice coil of the 5'' woofer and an acceptable excursion at 105dB@1m free field.
More details on how to simulate the excursion of the driver, can be found here.
Last edited:
Biblob
Addicted to Fun and Learning
- Joined
- Sep 13, 2018
- Messages
- 641
- Likes
- 614
That's not too bad actually! Especially the excursion, it isn't very high at those SPL levels. Although it will be at the limit of the driver.
The spec is 5mm if I recall correctly from the top of my head. Erin measured about half the distance from the manufacturer specs, so that would be around 3 mm.
Last edited:
If my recollection doesn't trick me, the limitation in excursion was due to the suspension, which isn't a major problem. It doesn't generate intermodulation, and HD is to a vast percentage masked by the musical content, which brings with it an overwhelming portion of harmonics (hence the name) anyway.
Do You aim to swamp an airport terminal with the speaker? I would argue that the combination of high force (acceleration) and high temperature could shut down a 'last call' announcement preliminary.
Really interesting thread ! Thanks ctrl. I have two questions:
- you mentioned that slot area is often closed to cone area, do you have some link or paper which detail slot design rules (area and position) ?
- what would look the v and h sonogram of a front slot vented box (for instance a cube with woofer at the center and four slots around the woofer on front side) ?
- you mentioned that slot area is often closed to cone area, do you have some link or paper which detail slot design rules (area and position) ?
- what would look the v and h sonogram of a front slot vented box (for instance a cube with woofer at the center and four slots around the woofer on front side) ?
OP
- Thread Starter
- #48
I think there might be something in D&D's European patent application - but I'm not sure.
But my information was only a very rough comparison to aperiodic speaker concepts with Variovent, whose area is rather small - compared to the cabinet size (roughly 80cm² at 50L volume).
During the design of the Directiva r2 it turned out that a slot area of almost that of the 5''-6'' woofer is optimal. In another project with a 10'' woofer I deliberately reduced the slot area a bit, so that it is only 60% of the woofer Sd - because in this way a better radiation is achieved.
There is no "real rule" for the relationship between driver area and slot area. Especially since the cabinet surface areas , cabinet dimensions and the position of the slots also still play a major role.
If one is "unscientific", one can summarize the few data points (the D&D 8c represents another data point) and classify very roughly in which ratio slot area to Sd should be...
Well, someone should simulate that
If the sound sources are arranged symmetrically, then hor and ver result in the same radiation - obviously.
Depending on the arrangement of the slots, the driver would be surrounded circularly or rectangularly by sound sources radiating with inverse phase. No idea how close the radiation then comes to an OB speaker without baffle (freely mounted driver).
I am intrigued by this page, describing a cardioid result from an 18” damped uframe, which is similar but different from the items so far mentioned here.
Intrigued doesn’t mean I fully understand it yet. The part about group delay is interesting.
Intrigued doesn’t mean I fully understand it yet. The part about group delay is interesting.
OP
- Thread Starter
- #50
First, this is the most unreadable website I've ever seen - who uses images to display text!
That's an interesting approach to using U-frames. The only thing that may not be quite optimal is the jump in sound pressure level of the rear radiation starting at 60Hz in the last image. This is where the U-frame resonance starts to interfere with the cardioid radiation, which would have to be looked at more closely in a simulation or in complete measurements.
The additional group delay by the damping material is a consequence of the phase shift (caused by the damping material). The damping material (which does not damp all frequencies equally, frequency dependence of the absorption coefficient) acts like a lowpass filter, i.e. it damps higher frequencies more strongly (which somewhat mitigates the extreme U-frame resonance). But a lowpass filter always causes a phase shift, which is also shown in the images. Additional phase shift also means additional group delay because group delay is the negative derivative of the phase function over the angular frequency.
OP
- Thread Starter
- #51
Does each cardioid speaker design, using slots, has increased distortion?
If you don't feel like reading the whole text (don't even think about it), the answer is 'no' (and not 42)
I was very intrigued by this question, because if true, sloted design concepts would have a real disadvantage.
In previous posts it was suggested that the main reason for increased distortion in sloted designs (if it happens) is most likely caused by increased driver excursion (and a driver that cannot cope with this excursion, even at low SPL). Because the sloted design for cardioid radiation behaves more or less like an open baffle speaker at low frequencies.
To further support this assumption, a sloted speaker design was created and the distortion behavior of the bass-midrange driver was intensively investigated.
The speaker uses a 10'' midrange driver with very stiff suspension and strong motor (and twin demodulation rings for lower distortion). The strong motor ensures that the damping of the oscillation is mainly electrical and not via the mechanical resistance Rms or by dampening effect of the enclosure - Rms = 5.5 kg/s and Rel = 42.6 kg/s. Therefore, no cabinet with additional damping resistor is needed to control the oscillation behavior of the driver cone - to list a few points why the driver was chosen. With specific OB drivers (low fs, Qts >0.6), Rel is often only half as large, but they provide higher SPL in low bass reproduction in OB speaker concepts.
For reference, here again is the harmonic distortion (HD) measurement of the D&D 8c measured by Erin.
At 96dB, the speaker reaches at 100Hz almost 10%** HD3 and well over 1%** HD5 (which is really unusual).
** unfortunately Erin's HD measurements never show the FR of the fundamental, which makes evaluation and comparability very difficult. The percentages refer to an ideal linear frequency response of the fundamental (which is never the case) - to be able to interpret the HD measurements, one needs the response of the fundamental. Since in case of the 8c the on-axis FR is very linear down to 30Hz, the statement is sufficiently accurate.
The crossover frequency to the subwoofers is 100Hz with fourth order Linkwitz-Riley filter. So at 100Hz the SPL of the 8'' bass-midrange of the D&D 8c is reduced by 6 dB.
If the loudspeaker concept of cardioid radiation via lateral slots and especially the slots themselves (as was also assumed here in the thread) would cause increased distortion, all corresponding speaker designs would have to show this.
Therefore, we now look at the harmonic distortion of the speaker with side slots shown above. I know a 8'' (D&D 8c) to 10'' (speaker above) midwoofer comparison is not fair, but as I said, the point is to determine whether the speaker concept provides for increased distortion.
Measured at 0.3m distance (no gate) with sine sweep at SPL of 100dB@1m@1kHzSine, only the 10'' midrange driver without XO:
Down to 180Hz the SPL is at 100dB or above and (T)HD is way below 1% and there is no significant rise of HD (the 8c already shows a few percent THD at 200Hz).
To be absolutely sure, we measure the driver cone and the lateral slots separately in the near field (a few centimeters away) with higher details.
a) speaker 10'' midrange driver near field - first image fundamental with HD, second is HD percentage relative to the fundamental:
b) speaker lateral slots - first image fundamental with HD, second is HD percentage relative to the fundamental:
No problems there either, the lateral slots do not cause increased distortions. As expected, the distortion increases with the excursion of the driver at low frequencies.
If the driver quality is good and care is taken to keep the excursion of the driver within its capabilities, then a cardioid speaker design with side slots will not show increased distortion.
If you don't feel like reading the whole text (don't even think about it), the answer is 'no' (and not 42
)I was very intrigued by this question, because if true, sloted design concepts would have a real disadvantage.
In previous posts it was suggested that the main reason for increased distortion in sloted designs (if it happens) is most likely caused by increased driver excursion (and a driver that cannot cope with this excursion, even at low SPL). Because the sloted design for cardioid radiation behaves more or less like an open baffle speaker at low frequencies.
To further support this assumption, a sloted speaker design was created and the distortion behavior of the bass-midrange driver was intensively investigated.
The speaker uses a 10'' midrange driver with very stiff suspension and strong motor (and twin demodulation rings for lower distortion). The strong motor ensures that the damping of the oscillation is mainly electrical and not via the mechanical resistance Rms or by dampening effect of the enclosure - Rms = 5.5 kg/s and Rel = 42.6 kg/s. Therefore, no cabinet with additional damping resistor is needed to control the oscillation behavior of the driver cone - to list a few points why the driver was chosen. With specific OB drivers (low fs, Qts >0.6), Rel is often only half as large, but they provide higher SPL in low bass reproduction in OB speaker concepts.
For reference, here again is the harmonic distortion (HD) measurement of the D&D 8c measured by Erin.
At 96dB, the speaker reaches at 100Hz almost 10%** HD3 and well over 1%** HD5 (which is really unusual).
** unfortunately Erin's HD measurements never show the FR of the fundamental, which makes evaluation and comparability very difficult. The percentages refer to an ideal linear frequency response of the fundamental (which is never the case) - to be able to interpret the HD measurements, one needs the response of the fundamental. Since in case of the 8c the on-axis FR is very linear down to 30Hz, the statement is sufficiently accurate.
The crossover frequency to the subwoofers is 100Hz with fourth order Linkwitz-Riley filter. So at 100Hz the SPL of the 8'' bass-midrange of the D&D 8c is reduced by 6 dB.
If the loudspeaker concept of cardioid radiation via lateral slots and especially the slots themselves (as was also assumed here in the thread) would cause increased distortion, all corresponding speaker designs would have to show this.
Therefore, we now look at the harmonic distortion of the speaker with side slots shown above. I know a 8'' (D&D 8c) to 10'' (speaker above) midwoofer comparison is not fair, but as I said, the point is to determine whether the speaker concept provides for increased distortion.
Measured at 0.3m distance (no gate) with sine sweep at SPL of 100dB@1m@1kHzSine, only the 10'' midrange driver without XO:
Down to 180Hz the SPL is at 100dB or above and (T)HD is way below 1% and there is no significant rise of HD (the 8c already shows a few percent THD at 200Hz).
To be absolutely sure, we measure the driver cone and the lateral slots separately in the near field (a few centimeters away) with higher details.
a) speaker 10'' midrange driver near field - first image fundamental with HD, second is HD percentage relative to the fundamental:
b) speaker lateral slots - first image fundamental with HD, second is HD percentage relative to the fundamental:
No problems there either, the lateral slots do not cause increased distortions. As expected, the distortion increases with the excursion of the driver at low frequencies.
If the driver quality is good and care is taken to keep the excursion of the driver within its capabilities, then a cardioid speaker design with side slots will not show increased distortion.
Last edited:
jaakkopetteri
Active Member
- Joined
- Apr 10, 2022
- Messages
- 254
- Likes
- 175
WRT distortion: https://www.diyaudio.com/community/threads/a-3-way-design-study.376620/page-61#post-7343115
also see post #1199
also see post #1199
@ctrl
Nice experiment! I see that I may not have mentioned it explicitly in the thread before, but during the development of the Manta we did distortion tests with a sealed cabinet first and then with the caridioid slots. We saw no significant change in distortion.
In other words, same conclusion as your experiment.
Nice experiment! I see that I may not have mentioned it explicitly in the thread before, but during the development of the Manta we did distortion tests with a sealed cabinet first and then with the caridioid slots. We saw no significant change in distortion.
In other words, same conclusion as your experiment.
OP
- Thread Starter
- #54
Thanks for the link, didn't know that thread.
It is always a good thing when simulations and conclusions are independently confirmed (or disproved) - even if it was in this case probably more by chance.
Thank you for confirming the measurements. With this, we can assume that the next version of the D&D 8c will no longer exhibit the "distortion behavior", as it is relatively easy to fix
Does each cardioid speaker design, using slots, has increased distortion?
If you don't feel like reading the whole text (don't even think about it), the answer is 'no' (and not 42
I was very intrigued by this question, because if true, sloted design concepts would have a real disadvantage.
In previous posts it was suggested that the main reason for increased distortion in sloted designs (if it happens) is most likely caused by increased driver excursion (and a driver that cannot cope with this excursion, even at low SPL). Because the sloted design for cardioid radiation behaves more or less like an open baffle speaker at low frequencies.
To further support this assumption, a sloted speaker design was created and the distortion behavior of the bass-midrange driver was intensively investigated.
The speaker uses a 10'' midrange driver with very stiff suspension and strong motor (and twin demodulation rings for lower distortion). The strong motor ensures that the damping of the oscillation is mainly electrical and not via the mechanical resistance Rms or by dampening effect of the enclosure - Rms = 5.5 kg/s and Rel = 42.6 kg/s. Therefore, no cabinet with additional damping resistor is needed to control the oscillation behavior of the driver cone - to list a few points why the driver was chosen. With specific OB drivers (low fs, Qts >0.6), Rel is often only half as large, but they provide higher SPL in low bass reproduction in OB speaker concepts.
View attachment 300821
For reference, here again is the harmonic distortion (HD) measurement of the D&D 8c measured by Erin.
At 96dB, the speaker reaches at 100Hz almost 10%** HD3 and well over 1%** HD5 (which is really unusual).
View attachment 300825
** unfortunately Erin's HD measurements never show the FR of the fundamental, which makes evaluation and comparability very difficult. The percentages refer to an ideal linear frequency response of the fundamental (which is never the case) - to be able to interpret the HD measurements, one needs the response of the fundamental. Since in case of the 8c the on-axis FR is very linear down to 30Hz, the statement is sufficiently accurate.
The crossover frequency to the subwoofers is 100Hz with fourth order Linkwitz-Riley filter. So at 100Hz the SPL of the 8'' bass-midrange of the D&D 8c is reduced by 6 dB.
If the loudspeaker concept of cardioid radiation via lateral slots and especially the slots themselves (as was also assumed here in the thread) would cause increased distortion, all corresponding speaker designs would have to show this.
Therefore, we now look at the harmonic distortion of the speaker with side slots shown above. I know a 8'' (D&D 8c) to 10'' (speaker above) midwoofer comparison is not fair, but as I said, the point is to determine whether the speaker concept provides for increased distortion.
Measured at 0.3m distance (no gate) with sine sweep at SPL of 100dB@1m@1kHzSine, only the 10'' midrange driver without XO:
View attachment 300885
Down to 180Hz the SPL is at 100dB or above and (T)HD is way below 1% and there is no significant rise of HD (the 8c already shows a few percent THD at 200Hz).
To be absolutely sure, we measure the driver cone and the lateral slots separately in the near field (a few centimeters away) with higher details.
a) speaker 10'' midrange driver near field - first image fundamental with HD, second is HD percentage relative to the fundamental:
View attachment 300887 View attachment 300888
b) speaker lateral slots - first image fundamental with HD, second is HD percentage relative to the fundamental:
View attachment 300889 View attachment 300890
No problems there either, the lateral slots do not cause increased distortions. As expected, the distortion increases with the excursion of the driver at low frequencies.
If the driver quality is good and care is taken to keep the excursion of the driver within its capabilities, then a cardioid speaker design with side slots will not show increased distortion.
I think the quality of midwoofer/hifi segment driver design has gotten good enough that these designs are now much more realistic. I wonder how well they would have worked in the pre-skaaning era - of course back then people simply used larger woofers to achieve high sound quality.
The designs we discuss on ASR tend to be difficult to commercialize, state-of-the-art type things, but I wonder if we'll start to see affordable active monitors with cardioid radiation in the near future. Maybe @KaliAudio_Official will create something. I doubt we'll be seeing an affordable speakers akin to the 8381 - arrays of drivers and huge waveguides are intrinsically expensive - but I see nothing about the directiva design or the sigberg design that couldn't be translated into a mass-market product.
I have been developing and measuring various passive cardioids for the past 2 years and this thread has been really usefull - thanks! I intent to open a separate thread to share results & insights but for now allow my to join in: Whilst both (nice!) designs lead indeed to low distortion in the lower mid-bass region - based on my findings it is for different reasons (initially I worked on same hypothesis). My conclusion/hypothesis is that provided a low distortion driver (I have used Purifi PTT6.5-X for prototypes) is selected, the main potential cause for distortion is due to noise created by the passive resistant ports notably when relative small box volume is applied as is the case for the 8C (which I have validated with prototypes with similar Vbox to Sd ratios). In this case when applying appropriate damping at the slots which creates large flow-resistance (needed to sufficiently attenuate the backwave to achieve cardioid - otherwise radiation shifts back to dipole-like) as there is limited air-cushion to provide compliance it leads to high cyclical dPs over the slots with excursion under 300 Hz indeed increasing rapidly. Both your and the Sigberg designs have directionally large midrange enclosures, which allows for mild compression which in turns dramatically reduces dp and airflow over the slots. It was quite revealing when I started to use the Purifi driver to measure nearfield distortion for a ~4 ltr enclosure at the cone at 8V input to be -50dB to well below 125Hz (consistent with Purifi’s specs and HiFi Compass) whilst ports (8x20cm with 4cm Akotherm on both sides) measured distortion well in access of 3% - and moving to indoor ‘midfield’ 315mm distance already the combined vectors showing distortion in access of 2% which stays about constant when moving to 80cms. In fact it led me to develop a simulation model (which I use in combination with VituixCad) which clearly shows the impact of enclosure size on air velocities through the ports. To be continued.Thanks for the link, didn't know that thread.
It is always a good thing when simulations and conclusions are independently confirmed (or disproved) - even if it was in this case probably more by chance.
Thank you for confirming the measurements. With this, we can assume that the next version of the D&D 8c will no longer exhibit the "distortion behavior", as it is relatively easy to fix
For your entertainment : my current project, a 3-way with a horn-loaded 1'' compression driver (BMS 4550 + AudioHorn X-Shape X25), 8'' midrange (B&C 8MBX51) and a serious 15" from SB Audience.
Problem was matching the horn's near-constant directivity to the wider and less controlled one of the midrange. Thanks to inspiring postings by @ctrl and others , I decided to learn a bit of Akabak and give resistive enclosure cardioidism a simulated try for this midrange (intended range : 300 - 1300 Hz).
Total slot area is a few cm² above Sd.
Here is the modelled box, after a few iterations and pulling back of the slots to allow for more bass extension :
Simulated polar map (click to enlarge) :
And my favorite, horizontal plane SPL distribution @300 Hz, 600 Hz and 1.2 kHz, wall impedance = 0.2 :
Same frequencies, wall impedance set to 0.7 :
Now a few questions to you, seasoned users, given the variability of (specially low end) SPL depending on software settings : in your experience, how does the wall impedance as defined in Akabak correlate with real-life foams' and wools' absorbance/resistivity ?
Does a lowish-QTs driver (0.29 here IIRC) tend to be more prone to SPL variability depending on back loading ?
Problem was matching the horn's near-constant directivity to the wider and less controlled one of the midrange. Thanks to inspiring postings by @ctrl and others , I decided to learn a bit of Akabak and give resistive enclosure cardioidism a simulated try for this midrange (intended range : 300 - 1300 Hz).
Total slot area is a few cm² above Sd.
Here is the modelled box, after a few iterations and pulling back of the slots to allow for more bass extension :
Same frequencies, wall impedance set to 0.7 :
Now a few questions to you, seasoned users, given the variability of (specially low end) SPL depending on software settings : in your experience, how does the wall impedance as defined in Akabak correlate with real-life foams' and wools' absorbance/resistivity ?
Does a lowish-QTs driver (0.29 here IIRC) tend to be more prone to SPL variability depending on back loading ?
