Hmm, do you refers to the dip around 150Hz? Where is the crossing between the mid and the woofers? Did you tried the mids and the woofers separately, to see where that dip comes from exactly. It can be a front or side wall bounce too, imo.
Oh well, you see them even in the brochures...
A lie repeated becomes truth, after a while. But that's not the only one, so let's get it on
The phenomenon is real, but the term "standing waves" is not a good description. I prefer the term Acoustical Resonance.
An acoustical resonance will occur whenever an internal cabinet dimension equals the half wavelength of a signal. For instance, if a cabinets largest internal dimension is 20 inches, the frequency with a half wavelength of 20 inches is 338 Hz. The SPL inside the cabinet can rise as much as 20 dB at 338 Hz. This is very significant because the transmission loss (sound suppression) of a 3/4" layer of MDF or plywood is about 20 dB, so at 338 Hz, the sound coming through the cabinet walls can be nearly equal to the sound coming from the driver. This is not a structural resonance problem, just ordinary sound transmission.
j.
An acoustical resonance will occur whenever an internal cabinet dimension equals the half wavelength of a signal. For instance, if a cabinets largest internal dimension is 20 inches, the frequency with a half wavelength of 20 inches is 338 Hz. The SPL inside the cabinet can rise as much as 20 dB at 338 Hz. This is very significant because the transmission loss (sound suppression) of a 3/4" layer of MDF or plywood is about 20 dB, so at 338 Hz, the sound coming through the cabinet walls can be nearly equal to the sound coming from the driver. This is not a structural resonance problem, just ordinary sound transmission.
j.
I am not sure why there is a disagreement about "standing waves" in this thread. Sound is a pressure gradient, there is just as much a pressure gradient in the box as out. Treating that appropriately is needed for good sound because it will get out of the box, through the cones, through the side walls, through the port, coupled to the floor, wherever it can. To address that I have the sloped wall on one end of my long internal dimension, every surface has 1/2"+ of cotton plus bonded dacron, the box its self is 3/4" MDF with a 1-1/2" front panel and 80mil of foil backed bitumen. Window pane bracing, I could do more, but it's adequate. We have several threads on construction methods with extensive measurements and theory that prove out.
Standing waves is the technical term and they do occur in all enclosures at frequencies small enough. Only damping material can prevent them.
Why do you think you can have standing waves in a room, and not in an enclosure, which usually has the same shape as the room they are in?
In short: Standing waves inside enclosures are real, it is the correct word for them, it can predicted at which frequencies they will occur given the dimensions of the enclosure, and there are proven ways to get rid of them.
Why do you think you can have standing waves in a room, and not in an enclosure, which usually has the same shape as the room they are in?
In short: Standing waves inside enclosures are real, it is the correct word for them, it can predicted at which frequencies they will occur given the dimensions of the enclosure, and there are proven ways to get rid of them.
so you feed the speaker with a steady signal or impulse and if it appears to be any tail when stopping, that's the standing wave ? So how can a low frequency wave form inside an enclosure with dimensions of …1/2-1 m and then superimpose to the same coming from the bounce to the wall and encountering ƒ/2 to make +6 dB ? That happens in a much bigger box, say, a room.
Sound...
and musical sound
what's a musical sound? Are ear/brain/culture etc. involved?
Ah ah ah
Sound...
and musical sound
what's a musical sound? Are ear/brain/culture etc. involved?
What gating does SHD use? You should look at impulse/ETC to see at what delay this strong reflection occurs. Then you can perhaps recognize the source - is it floor, backwall or perhaps ceiling. Then change position of mic or speaker and see how dip/peak change. REW is very easy with this kind of "analysis".
Basically very good looking response, congratulations! My thread of a rather similar 3-way with double 8" woofers sealed https://www.diyaudio.com/community/threads/avalanche-as1-modernization.321711/
Wavelengths that are double that of a given distance between two reflective surfaces will create a standing wave. Same goes for discrete fractions of that distance. So low frequencies cannot set up standing waves in small enclosures, there is a lower limit for each enclosure.
Absolutely great Speakes, love them!
Yes, standing waves in the enclosure are the worst source of cabinet sound, because the wave's pressure maximum is at the inner surfaces which heavily excites the walls to also make sound. It gets worse to handle the larger the cabinet is, because larger wavelength = lower frequencies are more difficult to absorb effectively. The cabinet wall panels also show resonances, I often try to set the internal braces that the first panel resonances are out of the woofer's transmission range, and that the resulting panel areas are unequal to distribute the overall parasitic resonance spectrum. Some rough analysis of that effects for a large speaker may look like this:
Yes, standing waves in the enclosure are the worst source of cabinet sound, because the wave's pressure maximum is at the inner surfaces which heavily excites the walls to also make sound. It gets worse to handle the larger the cabinet is, because larger wavelength = lower frequencies are more difficult to absorb effectively. The cabinet wall panels also show resonances, I often try to set the internal braces that the first panel resonances are out of the woofer's transmission range, and that the resulting panel areas are unequal to distribute the overall parasitic resonance spectrum. Some rough analysis of that effects for a large speaker may look like this:
After getting my smaller set of speakers veneered I will be taking these apart to finally finish the big boys (other thread https://www.diyaudio.com/community/...y-as-gift-bass-alignment.412891/#post-7905486). In the course of that disassembly I was looking at doing a crossover upgrade. Already pulled the trigger on getting poly caps to replace the electrolytics (not expecting a performance benefit, was just doing it for longevity and peace of mind).
Would It be worth updating the woofer to mid crossover from 2nd to 3rd order? The speakers measure and sound great currently, but wouldn't mind pulling the breakup of the woofer further down if it made sense to.
I have been playing with this and the trouble I am running into is I can't keep the impedance reasonable. It is dipping below 2 ohms rather broadly in the lower mids. Looking for any tips or tricks I can employ to make this work, if it is even worth the effort. Things seem to work reusing my 8mH and 4mH iron core inductors, but I will be looking at getting another 370uF per speaker which is $$$ in poly.
The other update I will be making is throwing on some SVS soundpath isolators with some outriggers from parts express. I got them for a subwoofer build and tried them on the mains hoping they wouldn't do anything, but alas a very clear improvement in cleaning up the tactile bass. We have extremely live wooden floors and these dig in the low 20s in room. Sensation went from a general pleasant buzz of energy through the feet and seat to clear punches. Auditory improvements I did not measure, but I'm sure the tactile sensation tricked my ear brain. Others have shown isolation does not directly improve speaker measurements, more so contributions from other objects that are excited by speakers. Test track was Limit to Your Love - James Blake at my 80-85dB avg level. May not be as import after building my sub, but improvement in tactile base was so clear I'm going to do it.
Would It be worth updating the woofer to mid crossover from 2nd to 3rd order? The speakers measure and sound great currently, but wouldn't mind pulling the breakup of the woofer further down if it made sense to.
I have been playing with this and the trouble I am running into is I can't keep the impedance reasonable. It is dipping below 2 ohms rather broadly in the lower mids. Looking for any tips or tricks I can employ to make this work, if it is even worth the effort. Things seem to work reusing my 8mH and 4mH iron core inductors, but I will be looking at getting another 370uF per speaker which is $$$ in poly.
The other update I will be making is throwing on some SVS soundpath isolators with some outriggers from parts express. I got them for a subwoofer build and tried them on the mains hoping they wouldn't do anything, but alas a very clear improvement in cleaning up the tactile bass. We have extremely live wooden floors and these dig in the low 20s in room. Sensation went from a general pleasant buzz of energy through the feet and seat to clear punches. Auditory improvements I did not measure, but I'm sure the tactile sensation tricked my ear brain. Others have shown isolation does not directly improve speaker measurements, more so contributions from other objects that are excited by speakers. Test track was Limit to Your Love - James Blake at my 80-85dB avg level. May not be as import after building my sub, but improvement in tactile base was so clear I'm going to do it.
How about 2 separate 8mH coils, one to each woofer, instead of stacking the woofers in parallel (directly) ? Adjust values as needed.
Maybe try the above idea in 1st order, then compare the result to what happens if you add capacitors.
^It has nothing to do with the woofer, its the mid. (BTW: If you make two separate lowpass for both woofers, you must double coil values and half capacitor values to achieve the same behavior as the woofer impedance per "channel" is doubled. But this overall makes no sense as you gain nothing but increasing cost and build complicity by increasing just number of parts...)
The 3rd order option looks much worse budgeted to me with the 2x 200uF in the mid highpass. There is correlation between the mid overshoot around 400Hz and deep impedance drop at that frequency, the filter overshoot/overswing sucks the current needed for that by the low impedance. Group delay gets worse bump there too (symptom of the filter overswing energy storage, and overall higher GD due to the higher filter order. This frequency range is most sensitive to GD distortion), phase responses of woofer lowpass and mid highpass match better with the 2nd order also. Stay with the second order, sizing of the components and result looks much more reasonable and nice to me. I see influence of the impedance at the mid's resonance frequency to the electrical filter response, but overall resulting acoustical highpass filter slope looks well.
I would have a look at the very slight overemphasis in power response around 3kHz, it's not much but this frequency range ist most sensitive to hearing.
There is also potential in phase matching between mid and tweeter - well its okay, but you have good sound origin time alignment given by the waveguide and very nice drivers, so fight a bit to optimize this. 3kHz bump can be tuned with the tweeter highpass Q, highpass Q will affect phase matching and this probably contradictory, so play with the component sizing, use the optimizer and try also different topologies.
Here is an example how I optimized a similar symptom with similar midwoofer and tweeter setup:
https://www.diyaudio.com/community/...b-acoustics-sb15nbac30-4.391310/#post-7302114
I wonder a bit why there is no FR overshoot in the woofer channel around 100Hz without a serial notch in parallel to the woofers, that linearizes the impedance there.
Can you show the impedance response of the woofers alone without any filter? Its a vented enclosure, or?
The 3rd order option looks much worse budgeted to me with the 2x 200uF in the mid highpass. There is correlation between the mid overshoot around 400Hz and deep impedance drop at that frequency, the filter overshoot/overswing sucks the current needed for that by the low impedance. Group delay gets worse bump there too (symptom of the filter overswing energy storage, and overall higher GD due to the higher filter order. This frequency range is most sensitive to GD distortion), phase responses of woofer lowpass and mid highpass match better with the 2nd order also. Stay with the second order, sizing of the components and result looks much more reasonable and nice to me. I see influence of the impedance at the mid's resonance frequency to the electrical filter response, but overall resulting acoustical highpass filter slope looks well.
I would have a look at the very slight overemphasis in power response around 3kHz, it's not much but this frequency range ist most sensitive to hearing.
There is also potential in phase matching between mid and tweeter - well its okay, but you have good sound origin time alignment given by the waveguide and very nice drivers, so fight a bit to optimize this. 3kHz bump can be tuned with the tweeter highpass Q, highpass Q will affect phase matching and this probably contradictory, so play with the component sizing, use the optimizer and try also different topologies.
Here is an example how I optimized a similar symptom with similar midwoofer and tweeter setup:
https://www.diyaudio.com/community/...b-acoustics-sb15nbac30-4.391310/#post-7302114
I wonder a bit why there is no FR overshoot in the woofer channel around 100Hz without a serial notch in parallel to the woofers, that linearizes the impedance there.
Can you show the impedance response of the woofers alone without any filter? Its a vented enclosure, or?
Last edited:
The DCR can also be increased, so each coil may not be that much bigger. But it's a moot point.
I'd try measuring the in-room response with just 1 woofer connected and running full-range. Then use that woofer just to fill in the shelf for baffle step correction. Then tune the other woofer to match up with the mid-range. Not to mention, I wouldn't trust connecting 2 woofers directly in parallel, for all the microphonic and distortion anomalies that might occur.
First: There is no Impedance issue as the problem is the mid pass, so nothing to solve. But, the DCR might also be simply increased with one coil with just thinner winding by making it even cheaper, so taking two coils double the size (and more than double the costs) in parallel appears not the smartest solution to me. Or take just a cheap 20W wire resistor...
The other idea to just power one woofer when both are working to the same enclosure will result, that the disconnected woofer act like a passive driver with unpredictable resonance result, that will mess your measurements completely and have not much to do with the result when both are connected. Both Woofers in parallel is the intended use, so measure in that condition.
Honestly, please stop giving advice because what you write is just really all completely wrong technically.
The other idea to just power one woofer when both are working to the same enclosure will result, that the disconnected woofer act like a passive driver with unpredictable resonance result, that will mess your measurements completely and have not much to do with the result when both are connected. Both Woofers in parallel is the intended use, so measure in that condition.
Honestly, please stop giving advice because what you write is just really all completely wrong technically.
Playing with the crossover I think the real issue is the sensitivity of the mid. With the higher order crossover I need to boost the high pass, and as you pointed out that just leads impedance issues. Not going to pad down the rest of the system just to get a higher order crossover. I wish SB Acoustics had a true midrange in the CAC series like Revel does for the Be series.
So I think the conclusion is leave well enough alone, haven't had an issue with the 2nd order cross in application.
For people trying to spot the issue (I missed it for myself earlier), look at the 2nd chart on the right in both images, see in the electrical crossover how much boost is required to make the mid meet the woofers on the 3rd order cross over as compared to the 2nd.
Also as to the excess energy in 3khz, that isn't so pronounced in the built speakers. It is just an artifact of my data for the waveguide in vituixcad not having the perfect level match to the drivers in baffle. Was able to shape the mid/tweeter crossover with just that shunt resistor on the tweeter.
So I think the conclusion is leave well enough alone, haven't had an issue with the 2nd order cross in application.
For people trying to spot the issue (I missed it for myself earlier), look at the 2nd chart on the right in both images, see in the electrical crossover how much boost is required to make the mid meet the woofers on the 3rd order cross over as compared to the 2nd.
Also as to the excess energy in 3khz, that isn't so pronounced in the built speakers. It is just an artifact of my data for the waveguide in vituixcad not having the perfect level match to the drivers in baffle. Was able to shape the mid/tweeter crossover with just that shunt resistor on the tweeter.
Yes I've seen the 3rd order overshoot but only small difference in the overall acoustical results, so It might be fine when it sounds good.
I have put a bucking magnet to a SB15NBAC30-4 with ~7% increase in BL for a expensive N42 magnet, and 5% increase for a cheap Y35 90x36x20 ferrite magnet, which both give around half dB more efficiency (0,4 dB with the ferrite and 0,6 dB with the neodymium)
I have put a bucking magnet to a SB15NBAC30-4 with ~7% increase in BL for a expensive N42 magnet, and 5% increase for a cheap Y35 90x36x20 ferrite magnet, which both give around half dB more efficiency (0,4 dB with the ferrite and 0,6 dB with the neodymium)
(To be more correct: You can do so and measure the drivers separately. But you must short circuit the unused driver to give it the electrical damping by the amp to avoid the passive driver effect. But this is only needed in special cases, e.g. for 2.5 Way to measure like I did also in the past. In this case I use the lower Woofer also as baffle step filler. But here the woofers work just below bafflestep, so nothing to fill. It might make sense to go 3.5 way here with a small or dome mid and woofer-mid transition over 600Hz, but this is not the use case. Both woofers in parallel is the right solution here.)
Uhuh.
And the impedance dropping below 2 ohms from 120 Hz is a "mid-range issue". Good luck.
They are still connected in the bass, and still in parallel, just each using a separate coil. Above a few hundred Hz, voltage and current are significantly out of phase, so there is essentially no 'damping' or opposition to pressure on the cone.
Last edited:
Bucking magnet would be an interesting thing to try, but need more than the 0.5 dB. A true mid with less surround and shorter voice coil would give ~1-1.5 dB greater sensitivity and bring it more into possibility. I have a spare SB15CAC30-4 that was a QC miss I could try some things on after I wrap some other projects. I may try a high powered neo just to see, quality ferrites are hard from me to find locally