• 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!

Open baffle speaker pitfalls

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,641
Likes
6,311
Location
.de, DE, DEU
Open baffle (OB) speakers seem to be becoming more and more popular at the moment. The possible advantages of an OB speaker do not need to be listed here, they can be found on every manufacturer's website or audio forums.

When using or purchasing U-frame or V-frame (and to some degree H-frame) dipole designs, there is a risk that design errors will be made by DIY beginners if the necessary basic knowledge is lacking or speakers are purchased which are faulty designs.
Therefore, simple examples will be used to point this out and help you to avoid these pitfalls.

1668861722363.png


The examples are based on commercially available speakers. Unfortunate some manufacturers sell faulty designs based on U-frame dipole speakers too.
This is very bitter for the paying customers.
BUT if such a desgin error is made, it does not mean that the entire speaker will sound bad - please keep this in mind for the further sections.

I am not an expert on dipole design of loudspeakers, detailed information on dipole speakers can be found at Siegfried Linkwitz's website.
Before discussing the pitfalls of U/V-Frame (and to some extend H-frame) design, some basic thoughts on OB speaker.

Index
1. classic OB speaker
2. OB speaker - displacement volume is everything
3. Why U- and H-Frame?
4. The U-Frame pitfall
5. H-frame the U-frame killer
6. classic flat OB, U-frame or H-frame Design - OB low bass output?
7. V-frame instead of U-frame - what's the difference?
8. H frame with angled side walls
9. U-frame resonance side effects


1. classic OB speaker

First, let's look at a classic OB design - so we have a reference to compare to. That means a large flat baffle combined with a driver that can move as much air volume as possible.

Example 0
Here the simulation of an OB baffle with height, width, depth of 100cm x 44cm x 7cm (39''x17''x2.8'') with a 10'' woofer - better would be 18'', but for our comparisons this doesn't matter.

The images are,
- Front and back view of the simulated LS with reference axis as blue line,
- SPL speaker (red) and woofer front (blue) and back (pink) sound separately in 2m distance,
- Horizontal on-axis normalized sonogram,
- Horizontal normalized frequency responses 0-90° (10° steps):
1668644870525.png 1668644889428.png 1668644962981.png1668645021618.png 1668645062200.png
The reference example baffle is not ideally flat, but has a very slight U-frame shape to stay as close as possible to real designs (that use this practice to stabilize the baffle).


2. OB speaker - displacement volume is everything

Would like to list it here for the sake of completeness.
If the wavelength is large compared to the baffle, then the frequency response drops by 6dB/oct (compared to a monopole speaker) due to mutual cancellation of front and rear sound.
The consequences have been described by S. Linkwitz:
"The comparable monopole has a flat frequency response (0 dB), whereas the dipole rolls off at 6 dB/oct. To maintain constant sound pressure level (0 dB) the monopole displaces four times (+12 dB/oct) the air volume for every halving of frequency, but the dipole has to have eight times (+18 dB/oct) the displacement. Very quickly the maximum excursion capability of a driver becomes the limiting factor for maximally achievable sound level. Multiple drivers must be used to go beyond this limit."
Source: https://www.linkwitzlab.com/models.htm#A1

The larger the baffle, the later this transition occurs.
The sound pressure curve is of course individually different and depends on the baffle and driver size and the TSP values of the driver.

As an example, a simulation of a loudspeaker, with the same baffle and driver, once as an open baffle design and once as a closed box loudspeaker (CB) design.
More details can be found here ("Analysis" of cardioid speaker radiation via lateral slots) in section four.

In the example shown, the baffle still has an influence on the on-axis frequency responses of the OB and CB versions shown.
The tendency is nevertheless clearly recognizable, the CB speaker drops to low frequencies with about 6dB/oct, but the OB speaker with 10dB/oct. At 50Hz the sound pressure difference is about 6dB in favor of the CB version.
1668645314599.png 1668645399249.png
Although the OB speaker produces 6dB less on-axis sound pressure level, the excursion at 50Hz is 25% higher than the CB version (which may be due to the lack of damping effect on the cone excursion by the air sealed in the CB cabinet).


3. Why U- and H-Frame?

For the reproduction of low frequencies with good sensitivity you need large baffles. Folding the baffle reduces the width of the baffle front and allows a tolerable sensitivity at low frequencies.
The air enclosed in the U- and H-frame designs increases the mass to be moved by the driver and thus lowers the resonant frequency.
1668645607730.png1668645624542.png

This design reduces the risk of freaking out husbands, who have to limit the width of their newly to be purchased LED TV because of the OB speakers in the living room.

When using a U-frame, the path length is no longer the same at 90°, so the dipole radiation becomes asymmetrical and has more of a hypercardioid shape.
1668645749761.png
1668645811011.png

The use of H-frame designs avoids this asymmetry.
1668648279480.png

Sources:


4. The U-Frame pitfall

The following is not a revelation, but is known to everyone who is somewhat involved with U-Frame OB speaker (as I said before OB is not my expertise, so don't expect jaw-dropping news).
It's just to show you what happens if one uses U-frame (and to some extend H-frame) designs without basic knowledge and what this means for the speaker radiation.

The enclosed air volume in U-frame and H-frame designs naturally begins to resonate, which is an undesirable effect.
As a very rough approximation, one can imagine the designs, according to Martin J. King, as lambda/4 resonators (transmission line) - some kind of "pseudo-lambda/4" resonator.
However, the effective air column does not end abruptly at the edge of the cabinet, but protrudes slightly beyond it.
1668650293335.png

This characteristic can lead to very unpleasant side effects.

Example 1

The baffle and driver is almost the same as our reference example in section one. The height is slightly lower at 85cm, but the depth is enormous at 40cm U-frame.
The commercially available loudspeaker (there are probably versions with one or two 15'' woofers instead of one 10'' woofer as in the simulation), plays together with a compression driver horn. So the crossover frequency of the woofer is probably at 800-1200Hz.

The images are,
- Front and back view of the simulated LS with reference axis as blue line,
- SPL speaker (red) and woofer front (blue) and back (pink) sound separately in 2m distance,
- Horizontal on-axis normalized sonogram,
- Horizontal normalized frequency responses 0-90° (10° steps):
1668649252418.png 1668649268753.png 1668649400828.png 1668649459525.png 1668649489079.png
The frequency response is very wavy and therefore difficult to smooth with a passive XO.
The woofer shows a cardioid dipole radiation only clearly below 100Hz. Therefore the crossover frequency should not be higher.

Using the woofer up to 800Hz is a clear mistake and shows a much worse radiation than any monopole, because it is extremely uneven.


Example 2

The overall loudspeaker is described by the manufacturer as the "pinnacle of open baffle loudspeaker designs". We will only look at the low bass part of the speaker, which consists of four 6.5'' woofers. The dimensions were roughly estimated for the simulation (minor differences have no effect on the general result).
The dimensions for the simulation with height, width, depth are 85cmx25cmx40cm (33''x10''x16'').

The images are,
- Front and back view of the simulated LS with reference axis as blue line,
- SPL speaker (red) in 2m distance, topmost woofer is number four ... down to woofer one.
- Horizontal on-axis normalized sonogram,
- Horizontal normalized frequency responses 0-90° (10° steps):
1668652040785.png 1668652056178.png1668652480550.png 1668652649979.png 1668652698724.png
Again, you get a very wavy FR. The "pseudo-lambda/4" resonance of the U-frame causes the sound radiated from the rear to become so dominant that it far exceeds the sound radiated from the front of the drivers (the U-frame "acts" like a monopole because of its dominance in SPL). This widens the radiation extremely and completely destroys all the advantages of a dipole - the reduced lateral radiation.

In this case, the crossover frequency should not exceed 100-150Hz, otherwise the sound pressure level of the lateral frequency responses in the range 180-400Hz can be above that of the on-axis FR and the speaker does not show dipole behavior.


5. H-frame the U-frame killer

When you first think about it, you ask yourself "Huh? How can twice as many "resonance chambers" improve the problem?". The basic problem of the resonance frequency range does not disappear, the resonances still have in and out oscillation (just like any BR concept for example) which will show in the decay behavior.

In section 4 it was pointed out that H-frame concepts have a symmetrical radiation behavior, this of course also applies to the resulting resonances, which thus maintain the cardioid dipole radiation.

Example 3

The simulated speaker is identical to example 2, except that the 40cm cabinet depth uses an H-frame design, rather than the U-frame.
The front and back view of the simulated LS with reference axis as blue line:
1668683105561.png 1668683139034.png
The two "resonator chambers" of the H-frame design are each only 20cm deep, which shifts the "pseudo-lambda/4" resonant frequency to higher frequencies, compared to the U-frame.

When comparing on-axis frequency response of the U-frame with the H-frame design, the difference is not dramatic. Compared to the classic OB design from example 0, both FR are wavy.
1668683332771.png

The advantages of the H-frame design becomes obvious when looking at the sound radiation from the front and rear of each driver, as well as the normalized FR.
The images are,
- SPL speaker (red) in 2m distance, topmost woofer is number four ... down to woofer one
- Horizontal on-axis normalized sonogram,
- Horizontal normalized frequency responses 0-90° (10° steps):
1668683382916.png 1668683411215.png 1668683445897.png
Up to 250Hz crossover frequency the use is possible without radiation problems. If one accepts minor radiation errors, even use up to 500Hz is possible.

To compare the dramatic differences in the radiation of U-frame and H-frame side by side, here are the 50, 100, 200 and 400Hz polar diagrams (normalized to the axial frequency response) of example 2 and 3:
1668683874262.png 1668683900463.png
With the U-frame, the radiation above 100Hz is no longer a dipole, but more or less a mess.
The H-frame design shows a relatively even cardioid dipole radiation up to 400Hz.


6. classic flat OB, U-frame or H-frame Design - OB low bass output?
Using example 2, 3 and a classic flat OB version, it will be shown how the low frequency sound pressure level is affected by the different OB design concepts.
The reason for using U/H-frame design, as mentioned above, is to increase the low bass sound pressure level compared to a classic flat OB baffle.

This requires a simulation of example 2 as a classic OB speaker with a flat baffle.
Since I exceed the maximum limit of 35 images per post, this simulation is available in a separate post that can be found here as example 4.

Now we can compare the on-axis SPL of the different concepts (classic flat OB, U-frame 40cm, H-frame 40cm with always the same front baffle):
1668763929518.png
What immediately catches the eye is the very even frequency response of the version with classic flat baffle.

With the choice of the U-frame, the manufacturer of the speaker from example 2 has already made the wrong choice in terms of the radiation of the speaker (if the crossover frequency of the 6.5'' woofer is above 100-150Hz).
Was it at least the right decision to use the U-frame in terms of low frequency sound pressure level output?

For this purpose, the on-axis frequency responses are normalized to the FR of example 2, the U-frame concept chosen by the manufacturer:
1668764682944.png
If we look at the low frequency range below 150Hz, the H-frame concept is the better choice there as well, around 100Hz we expect over 1dB and around 50Hz almost 2dB higher sensitivity. Unfortunately, the manufacturer also made the wrong choice in terms of low bass output by choosing a U-frame concept (assuming my simulations do not contain errors).

When using an H-frame concept, compared to the classic flat OB baffle in the low bass range (in this case) around 4dB higher sound pressure can be achieved.


7. V-frame instead of U-frame - what's the difference?
Since I exceed the maximum limit of 35 images per post, see post#13 for this section.

8. H frame with angled side walls
For details see post#31.

9. U-frame resonance side effects
For details see post#32.
 
Last edited:
I had luck with a phy style open baffle, 6' tall, center baffle 16", wings hinged about 24" deep.

Wife liked it, said it didn't seem to big (with the wings more than 45 degrees bent back).
Must be a trick of the eye, not seeming wide due to height.....

12" Philips full range with a mass corner (ib) of maybe 100hz.

It basically had no bass.
In theory, dip at 280, peak at 140, then bass rolling off...............
My mostly 70-80's rock, well, sounded sad without that bass.

Very clean mids, especially with a 3" thick foam box I had around the back of the driver.
 
Thank you very much ctrl!
I guess you will present reasoning for multiway construction soon
 
Added a paragraph to the opening post about H-frame concept and comparison to U-frame designs "5. H-frame the U-frame killer" - I know... somewhat sensationalist headline, point 6 then will be "Aliens control our thou... ...argh.."


I guess you will present reasoning for multiway construction soon
I didn't actually plan to do that. I just wanted to point out that OB speaker with U/H-frame can lead to severe radiation problems.
Since such faulty speaker designs are offered for sale it seemed appropriate to reference that.
 
  • Like
Reactions: 617
Would You mind linking to a publicly available product so we could compare?
 
Would You mind linking to a publicly available product so we could compare?
I'm not sure, actually would like to avoid focusing the discussion on two manufacturers.
If others think this would be helpful, here are the two examples I chose as the basis for the two U-frame examples.
1668687538286.png
1668687547867.png
 
I'm not sure, actually would like to avoid focusing the discussion on two manufacturers.
If others think this would be helpful, here are the two examples I chose as the basis for the two U-frame examples.
Actually, it would be very nice to see if the concept of the spinorama holds for designs like those. I'm thinking of the problems we as CEA 2034 disciples had with in-wall mounted speakers. As a DIYer I'm currently struggling with true shelf-mounted speakers, direct v/ indirect sound proportion, Harman tilt.

(https://pierreaubert.github.io/spinorama/, anything?)
 
Last edited:
Actually, it would be very nice to see if the concept of the spinorama holds for designs like those. I'm thinking of the problems we as CEA 2034 disciples had with in-wall mounted speakers.
Why should there be problems using CTA-2034 measurements tuning such a speaker? The same goals always apply: The curves should be smooth without jumps, PIR and SP should have a smooth downward slope, DI and ERDI should be smooth with upward slope.
How much tilt is necessary for the respective curves or whether the axis frequency response drops slightly, has no slope or even rises slightly will certainly depend on the speaker concept and is ultimately decided by listening tests.
 
Example 4 - classic flat OB version of example 2 in the opening post (section 4)

The dimensions are identical to those of example 2, only the depth of the baffle was reduced to 5cm (2'').

The images are,
- Front and back view of the simulated LS with reference axis as blue line,
- SPL speaker (red) in 2m distance, topmost woofer is number four ... down to woofer one.
- Horizontal on-axis normalized sonogram,
- Horizontal frequency responses 0-90° (10° steps)
- Polar diagram, normalized (to on-axis FR): 50, 100, 200 and 400Hz
1668765691885.png 1668765703927.png 1668765726484.png 1668765855894.png 1668765881681.png 1668771367590.png
 
Last edited:

Years ago i started my open baffle journey with this design
Easy to DIY and many posibilities to try different drivers and cross-overs with the same panels
Ended up with Philips AD12202M8 fullrange, Eminence Alfa15A woofers and Swan RT2C-A tweeters (bi-amped)
I enjoyed them very much but...
WAF was not a succes ;)
DSC00011.JPG
 
Years ago i started my open baffle journey with this design
In this concept, too, the air volume resonates behind the baffle.
There is a review of the original kit with measurements up to 60°. There you can already see well that the radiation is no longer cardioid dipole like, in the range around 200-1000Hz the radiation extremely widens.
1668856422330.png
Source: Klang und Ton 2004-3

But, as I said in the opening post, how much this affects the sound I can't guess - except that it's definitely not the sound of a cardioid dipole OB speaker ;)
 
Last edited:
7. V-frame instead of U-frame - what's the difference?

In my search for manufacturers using U-frame concepts, my old eyes deceived me. The speaker I based my example 2 on actually uses a V-frame shaped concept.
So now I am forced to discuss this concept as well - I am glad I did not list a specific H-frame concept of a manufacturer, otherwise I would have to discuss X-frame concepts as well, in case I would have overlooked something crucial there as well... ;)


Example 5 - V-frame
This example is identical to example 2 except that rear width is 38cm (15''). The front baffle is still 25cm (10'') wide - hence the name V-frame concept.
1668859770668.png 1668859787904.png

To be able to assess whether it has advantages to change a design from U-frame to V-frame (rear cabinet width is increased), here first a direct on-axis SPL comparison (example 2 vs. example 5):
1668860042755.png
The V-frame seems to be slightly superior to the U-frame concept, but on the whole all the problems of this concept type remain.

The images are,
- SPL speaker (red) in 2m distance, topmost woofer is number four ... down to woofer one.
- Horizontal on-axis normalized sonogram,
- Horizontal normalized frequency responses 0-90° (10° steps):
1668860776050.png 1668860819792.png 1668860842187.png
The minimal improvements that the V-frame concept has over U-frame (example 2) can be seen most easily by comparing the normalized horizontal frequency responses.

However, since both concepts should only be used well below the air volume resonance (otherwise you will ruin the cardioid dipole radiation), there is hardly any difference in the end.
 
Last edited:
What about the Baffle "Peaking" frequency or effect?

(That narrow band, where the front and back waves REINFORCE each other and add a 6 db boost to the sound in the 200-400 hz range or there abouts)
 
What about the Baffle "Peaking" frequency or effect?

(That narrow band, where the front and back waves REINFORCE each other and add a 6 db boost to the sound in the 200-400 hz range or there abouts)
Sorry, that was my mistake, should have explained the diagrams a little better.

This "boost" between 200-400Hz only seems to exist, since the frequency responses are normalized to the axial frequency response (so if there would be a flat axial frequency response). In detail, the matter is somewhat more complex.

If we look again at the v-frame example 5 from post#13, we get the red curve ("speaker SPL") for the axial frequency response if all sound radiating sources are added up (woofer1 front + woofer2 back + woofer2 front + ...):
1668940373478.png
The resonance around 200Hz completely ruins the axial frequency response and the radiation pattern.
The resonance provides an additional phase shift (i.e. in addition to the delay difference of backward and forward radiated sound to the listener or microphone). The sound radiated from the rear of the driver is phase inverted (180° out of phase) with the sound radiated to the front. Which enables the dipole radiation pattern.

In yellow I have drawn the slope of the frequency response at 160Hz for the bottom woofer (No. 1). In very simplified terms, we can see the slope as a measure of the change in phase (for minimum phase systems).
The rear emitted sound has a large slope at 160Hz, i.e. positive phase change. The sound radiated to the front has a small slope, i.e. a small positive phase change. Thus, front and rear sound are no longer phase inverted to each other, so the summed sound pressure increases more and more.

This allows us to divide the axis frequency response into different ranges - see the pink areas in the image above.
Now let's look at the sound pressure ratios around the speaker in these four areas. We look at the speaker from above at an angle and look at the sound pressure in a plane around the speaker. Each color change corresponds to 5dB sound pressure difference.

Range 1: Dipole radiation (50-100Hz)
1668944660257.png

Range 2 (100-200Hz): Less cancellation of summed front and rear sound on axis, increasing dominance of rear radiation. The sound radiated to the rear (top) increasingly acts like a monopole, omnidirectionally (or subcardiodally) outshining everything else.
1668945805647.png 1668948826648.png

Range 3 (200-280Hz): Falling on-axis slope of resonance, phase behavior reverses, increasing cancellation of front and rear sound on axis.
1668949068525.png

Range 4 (280-800Hz): A real mess.
1668949220369.png 1668949255368.png 1668949292438.png 1668949369928.png


Or to make it easier, here are the non-normalized FR 0-90 deg in direct comparison to the normalized ones:
1668950122304.png 1668950142919.png
I probably should have started with this ;)
 
Last edited:
This is clearly seeen in all of your previous pics. And it happens in a flat baffle too, but not as severily. Freq of nondirectional radiation depends on baffle dimensions (pathlength for wavefronts too meet).

This phenomenom is avoided in a multiway speaker by using narrowing baffle for higher ways, lowpassed ranges below uniformity, until the tweeter must face it.

 
Last edited:
I have to apologize to you guys.
Have been dealing with cardioid radiating speakers so much lately that I have been referring to dipoles as cardioids here in the thread as well.
Have corrected the errors in the posts - hope I didn't confuse you completely.
 
Just a majorly ignorant question here, but isn't the room the baffle?
 
Just a majorly ignorant question here, but isn't the room the baffle?
Whatever metaphor is used to describe the influence of the listening room, it is always better to use a loudspeaker without design flaws for listening. Therefore, the hint on what to pay attention on U, V, H OB speaker.

Such a design flaw would exist, for example, if the V-frame OB bass section of example 5 is used in a 2- or 3-way loudspeaker at a crossover frequency above 150Hz (or example 1 above 100Hz).
1669066089326.png


We can also look at the frequency responses of the example 5 simulation as a CTA-2034 presentation (Spinorama). Those who are here in the forum more often should be familiar with this data presentation - only difference is 40dB instead of 50dB scaling.
Once without XO, with slightly downward sloping PIR and with semi-flat on-axis FR:
1669064308733.png 1669064378306.png 1669064391483.png
No matter how you design the crossover, the radiated sound power and on-axis FR/listening window never match - this is shown by the DI.
What you don't want is an abruptly and severely changing directivity index DI (or early reflections directivity index ERDI), because that means a sudden mismatch of direct sound to radiated sound power - which happens in example 5 above 150Hz.

Such behavior, no matter what listening room, will always be a problem.
 
Last edited:
Back
Top Bottom