Please tell me more about, "correct the MIB measurement"
It is my impression that the the high pass filter nature of the sealed enclosure is caused by the increasing pressure inside the box pushing back on the driver cone and limiting the cone excursion. High pressure inside reduces dBSPL outside.
Yes you can see the ripples in both the time and frequency domains.
The experiments are for fun. It is like being back in the physics lab.
It is not just enclosure wall resonance. It is also equal and opposite action and reaction. Even if the the enclosure is perfectly ridged, as in no flexing vibrations; when the speaker cone is pushed foreword the speaker frame is pushed backward. The speaker frame is attached to the enclosure and it moves along with the enclosure in the opposite direction of the speaker cone.
the GRAS microphone is a pressure sensor. The microphone is calibrated in Pascal/Volt. The Y Scale of the plot can be be changed from reading dBSPL to Pa/V, both are pressure measurements.
My purpose for using the MIB tool is for fun and to see what is going on pressure wise inside the various types of enclosures. Sealed is first.
Thanks DT
Final assembly is complete.
I am fairly pleased with how they look. The cherry has a lot of grain character, waviness, some curl, some spalting. The all-wood exterior gives a warm, soft appeal to a shape which would otherwise be a harsh industrial look. The basic shape of the speaker reminds me of an air-defense radar installation, or a microwave communication equipment. The polished wood makes it livable.
The bracket which supports the waveguide was a hard decision. I experimented with a couple of metal brackets, and I tried several shapes of wood bracket. I needed something that securely attached the waveguide to the box, concealed the wiring, was "buildable", and yet looked good. What I made is my best compromise between aesthetics, stiffness, and my ability to build it with my tools and skills.
Next step is to make polar measurements of each driver in the completed cabinet.
- - -
j.
I am fairly pleased with how they look. The cherry has a lot of grain character, waviness, some curl, some spalting. The all-wood exterior gives a warm, soft appeal to a shape which would otherwise be a harsh industrial look. The basic shape of the speaker reminds me of an air-defense radar installation, or a microwave communication equipment. The polished wood makes it livable.
The bracket which supports the waveguide was a hard decision. I experimented with a couple of metal brackets, and I tried several shapes of wood bracket. I needed something that securely attached the waveguide to the box, concealed the wiring, was "buildable", and yet looked good. What I made is my best compromise between aesthetics, stiffness, and my ability to build it with my tools and skills.
Next step is to make polar measurements of each driver in the completed cabinet.
- - -
j.
The virtually nonexistent expanding and contracting of the enclosure due to inside sound pressure variations.
That puts it into perspective. The balloon effect is the movement of cabinet walls at low frequencies. It is not a structural resonance movement, it is simply that the pressure inside the box has overcome the stiffness of the box, to the point where some detectable motion is created. As @markbakk says, it is very rare. An enclosure has to be very flimsy and poorly built for this to be a problem.
Here is the Joe D'Appolito article.
I don't want to take this out of context but clearly, as a mode, it is an effective one.
It will be a while before I can listen to them... first I have to complete my measurements, then design the DSP filtering. I intend to share all the relevant details, but I do not have a set of plans (engineering drawings). These speakers, like all of my speaker designs, are one of a kind custom fabrications. I start with simulations, I make prototypes, I make acoustical measurements of the prototypes, then I make more simulations. Based on that, I make some rough drawings, but the actual construction process is cut-to-fit, trim-to-fit.
j.
j.
the mib to far field correction is to tilt up the measurement by 12dB/Octave. According to recent talks with Andrew Jones, this technique was proposed by Richard Small in a 1972 paper.
Another neat trick for identifying resonance modes is to use the driver as sensor and tap the box. It’s mind blowing what this method reveals in a few seconds
Another neat trick for identifying resonance modes is to use the driver as sensor and tap the box. It’s mind blowing what this method reveals in a few seconds
I downloaded the 1972 Small AES paper. Small gives credit for the MIB Tool to Thiele.
Gave some thought to correcting the MIB plot to better estimate woofer output.
I am mostly sure that there is not a one size MIB correction that fits all. There are too many variables; driver parameters, Box size, type of enclosure; sealed, BR, PR and resistive port to name a few.
I am going to look at a few more MIB measurements looking for patterns; In box speaker parameters, impedance curves, voice coil excursion, FR and distortion will change.
The intension is to observe the pressure inside the box and distortion outside the box for different types of enclosures; sealed, BR, PR and resistive port.
Correction curves will be later.
Thanks DT
What's the relevance of MIB spl and it's curve? What do you have as reference? Same applies to using knocking and driver signal as mic.
Resonances and modes will show as peaks/dips easily, but what should the curve look like? Testing the effect of brace and damping materials would be interesting!
Reflex and TL tuning should be esasy to see... I have never tried these myself.
Resonances and modes will show as peaks/dips easily, but what should the curve look like? Testing the effect of brace and damping materials would be interesting!
Reflex and TL tuning should be esasy to see... I have never tried these myself.
Or just measure very nearfield and look at the Burst Decay graph (waterfall shown in periods).
Although there isn't much fancy to see, because those inbox standing waves line up extremely well with one can expect from theory. Sometimes they can show up in the impedance plot.
As for wall resonances, the only proper ways of doing this, is either with an accelerometer, or with a velocity sensor.
Yet still it's hard to compare apples with apples.
Like said before, just some decent (but simple) bracing will solve this problem entirely.
Bending or denting a material will have a similar effect, but not an interesting solution at all in speaker building.
It's a bit like splitting hairs at this point imo.
Btw, another great compromise for a full-range 2-way system.
Because the lower the frequency gets plus the more energy (cone excursion) is putting into, the more issues one can expect.
Last edited:
I think of it this way; the driver cone is a piston compressing the air in the sealed box.
Holding as many variables constant as possible including amplifier power input, sweep the speaker in a sealed box across the test bandwidth. The cone excursion at 100Hz will be 10 times greater than at 1000HZ yet the power input is the same. Why? Per cycle the cone does 10 times the work at 100Hz than it does per cycle at 1000Hz. The cone at 1000Hz cycles 10 times more often than at 100Hz. the amount of work done per unit of time at 100Hz is equal to the work done at 1000Hz. The measured frequency response outside the box is mostly equal.
The MIB is a pressure sensor, with equal power input the Pascals measured in the box decrease with increasing frequency.
Thanks DT
My mistake.
Not the first time I looked at Log graph paper and thought linear relationship.
Thanks DT
As this thread approaches 300 posts, I decided to add an index to the top (post #1). As I did that, I realized I never discussed my list of design constraints for this project.
I find it difficult to enumerate performance targets or design goals for a speaker project. In general, we are looking to maximize performance, and the balance among the different performance attributes is the "art" part of the art and science of speaker design. Making tradeoffs is necessarily a judgment call, so I find it very hard to define the performance goals in advance.
But design constraints are easy to define, and are actually more important. So here were the constraints I started with in this project.
Design Constraints
The new Mid-Tweeter system shall use a single PTT6.5M04-NFA-01 as the mid driver and a single Satori TW29BNWG-4 as the tweeter.
The new Mid-Tweeter system shall utilize the existing SB34NRX75-6 woofer bins, in their current location.
The new system shall cross from woofer to mid no higher than 200 Hz.
The new system shall use no passive filter components. All filtering shall be IIR DSP filtering as supported by Hypex Fusion HFD software.
The new system must utilize the existing 30” high speaker stands with the built-in Hypex amps.
The new system shall position the tweeter at a height of 38” minimum to 48” maximum.
The maximum width of the mid-tweeter system shall be 13”, the maximum depth shall be 13”, and the maximum weight shall be 35 lb.
The new system must accept a 5/16” bolt through the bottom to secure the system to the stand. The bolt is located at the center of the stand top plate. The speaker shall contain a threaded insert or T-nut for the bolt.
The wiring for the new system shall pass through the bottom of the speaker and through an access hole in the top plate. The wires shall be 12” to 16” in length, marked for polarity, and be terminated with #8 ring terminals.
The system exterior shall be designed to minimize the use of veneer and maximize the use of solid wood cladding or panels. The wood shall be cherry. No painted surface will be visible
The physical size, shape, layout, and configuration of the system shall be such that I can construct it in my shop, using my available tools. No new tool or equipment shall be required, except those costing less than $500.
The maximum machine-cut radius of any external roundover shall be 37 mm (1.5”). Machine cut profiles will be either circular profile, simple bevels, or compound bevels. Elliptical, parabolic, hyperbolic, or exponential curve profiles shall not be utilized.
I find it difficult to enumerate performance targets or design goals for a speaker project. In general, we are looking to maximize performance, and the balance among the different performance attributes is the "art" part of the art and science of speaker design. Making tradeoffs is necessarily a judgment call, so I find it very hard to define the performance goals in advance.
But design constraints are easy to define, and are actually more important. So here were the constraints I started with in this project.
Design Constraints
The new Mid-Tweeter system shall use a single PTT6.5M04-NFA-01 as the mid driver and a single Satori TW29BNWG-4 as the tweeter.
The new Mid-Tweeter system shall utilize the existing SB34NRX75-6 woofer bins, in their current location.
The new system shall cross from woofer to mid no higher than 200 Hz.
The new system shall use no passive filter components. All filtering shall be IIR DSP filtering as supported by Hypex Fusion HFD software.
The new system must utilize the existing 30” high speaker stands with the built-in Hypex amps.
The new system shall position the tweeter at a height of 38” minimum to 48” maximum.
The maximum width of the mid-tweeter system shall be 13”, the maximum depth shall be 13”, and the maximum weight shall be 35 lb.
The new system must accept a 5/16” bolt through the bottom to secure the system to the stand. The bolt is located at the center of the stand top plate. The speaker shall contain a threaded insert or T-nut for the bolt.
The wiring for the new system shall pass through the bottom of the speaker and through an access hole in the top plate. The wires shall be 12” to 16” in length, marked for polarity, and be terminated with #8 ring terminals.
The system exterior shall be designed to minimize the use of veneer and maximize the use of solid wood cladding or panels. The wood shall be cherry. No painted surface will be visible
The physical size, shape, layout, and configuration of the system shall be such that I can construct it in my shop, using my available tools. No new tool or equipment shall be required, except those costing less than $500.
The maximum machine-cut radius of any external roundover shall be 37 mm (1.5”). Machine cut profiles will be either circular profile, simple bevels, or compound bevels. Elliptical, parabolic, hyperbolic, or exponential curve profiles shall not be utilized.