Open baffle without limits

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The holy grail of open baffle threads on this site

This thread :
My open baffle dipole with Beyma TPL-150

has all the tech info plus multiple examples of extreme open baffle designs... Stig Erik is a Beyma dealer so has access to great drivers at low cost.
He is also computer / DSP centric so has all the skills to maximise the drivers potential.
I think he ended up building a dedicated listening room with one single chair as the listening position and the speakers about 2 meters from his head....
The open baffle speakers were incredible....

Multiple Hypex Ncore driving approx 1.5 SQUARE METERS of 18 inch or 21 inch drivers... !!!

The thread attracted lots of clever guys and the comments and insights are top drawer.
 
ATC's midrange is a dome with closed backside. It doesn't work as dipole at all, just like most tweeters don't.

Dipole speakers are more like luthier's artesanery rather than blacksmith's hammering! Delicate and almost a mission impossible! They don't sound special in a well damped room, they need those front and sidewall reflections to sound right - "dipolic" and to "enlighten the room with music"

Yes, I'm well aware of how the ATC style mid works. I have some other ideas that are a little non traditional in that regard but more on that later.

I guess what I"m really interested in is hearing how the low end sounds without a box, as it's something I feel I can hear even in very finely engineered traditional speakers. With that in mind it's more about the low end being dipole. I may also experiment with open back mid range drivers too but my initial experiments are going to be based around absorbing a lot of what comes from the rear, not using it in the Linkwitz sense.
 
This thread :
My open baffle dipole with Beyma TPL-150

has all the tech info plus multiple examples of extreme open baffle designs... Stig Erik is a Beyma dealer so has access to great drivers at low cost.
He is also computer / DSP centric so has all the skills to maximise the drivers potential.
I think he ended up building a dedicated listening room with one single chair as the listening position and the speakers about 2 meters from his head....
The open baffle speakers were incredible....

Multiple Hypex Ncore driving approx 1.5 SQUARE METERS of 18 inch or 21 inch drivers... !!!

The thread attracted lots of clever guys and the comments and insights are top drawer.

Fantastic! I'll dig into that. Cheers.
 
I guess what I"m really interested in is hearing how the low end sounds without a box, as it's something I feel I can hear even in very finely engineered traditional speakers. With that in mind it's more about the low end being dipole. I may also experiment with open back mid range drivers too but my initial experiments are going to be based around absorbing a lot of what comes from the rear, not using it in the Linkwitz sense.
The made advantage of dipole bass is less interaction with the room due to it's directivity, so long as the rear radiation is not absorbed it will work in this way. Linkwitz advocates using monopoles below 40Hz.
 
I recently posted my design approach to "open baffle" type loudspeakers here:
Multi-way OB

This was written in response to someone wanting to use a fullranger in a large baffle. I advise against that, and the approaches of "FAST" (full range and subwoofer technology) and "WAW" (woofer assisted wideband).

My approach typically gives rise to a 3- or 4-way loudspeaker. This more or less means active crossovers, multiple channels of amplification, etc. It's not a great match to a passive crossover.
 
I guess what I"m really interested in is hearing how the low end sounds without a box, as it's something I feel I can hear even in very finely engineered traditional speakers. With that in mind it's more about the low end being dipole. I may also experiment with open back mid range drivers too but my initial experiments are going to be based around absorbing a lot of what comes from the rear, not using it in the Linkwitz sense.

If you want dipole bass, the you need something rather large. There is just no way to get around the physics of it, and the situation is much worse than for a boxes loudspeaker due to the front-back cancellation. If you can keep the pathlength long you can get dipole bass all the way down to 20Hz, but at the same time that driver can't be used much above 150Hz.

As someone notes, Linkwitz said that sealed subs could be used at the lowest frequencies. This is partly because he uses a very short woofer M-frame (or which ever frame it is). Thus at low frequencies there is lots of cancellation and his twin 10" woofers cannot keep up - they are excursion limited. A larger driver (or drivers) in longer/larger frames of some type (I suggest an H-frame) can be very impressive when the right driver is used. This is not your average home or HT subwoofer driver, however.
 
Use 10 x 15" GPA for each side in SLOB (slot loaded open baffle) and 2 or 3 12" Acoustic Elegance or GPA 414 for mids.
Horns from mids and up. JMLC400 with a good CD and TAD ET-703 or Fostex T500A mkll on top.

Here is my compact system, sorry for bad looks. Not finish yet.

488428d1522602156-stemningsrapport-med-bilde-20180401_185230.jpg



I use 6x12" GPA414 for bass (SLOB), 2x12" GPA414 for midbass, TAD 2001 in waveguide and T500A mkll on top.
DSP: Audiolense XO, RME pro USB soundcard.
XO: 200, 1400 and 10k.
Sounds ok, but later I will rebuild the satellites and get a different color.

OB works best from midbass and down.
 
Last edited:
If you want dipole bass, the you need something rather large. There is just no way to get around the physics of it, and the situation is much worse than for a boxes loudspeaker due to the front-back cancellation. If you can keep the pathlength long you can get dipole bass all the way down to 20Hz, but at the same time that driver can't be used much above 150Hz.

As someone notes, Linkwitz said that sealed subs could be used at the lowest frequencies. This is partly because he uses a very short woofer M-frame (or which ever frame it is). Thus at low frequencies there is lots of cancellation and his twin 10" woofers cannot keep up - they are excursion limited. A larger driver (or drivers) in longer/larger frames of some type (I suggest an H-frame) can be very impressive when the right driver is used. This is not your average home or HT subwoofer driver, however.

Excellent points all, thanks Charlie.

I'm not really interested in sealed subs for now, or practical solutions! I have access to class leading sealed, ported and transmission line speakers through my work. For now, I'm curious to see what open baffle can do, and I have no limits on size. I don't actually need huge SPL either but I do need the LF extension all the way down.

Why is that you say you cannot have the big baffle, dipole bass down to 20Hz but not have that same driver go higher than 150 Hz (ish). Is it limited by the size of the driver required to move that air?

Thanks all, really enjoying this exploration and I realize that I am far from a typical DIYer.

Cheers,
Ruairi
 
Why is that you say you cannot have the big baffle, dipole bass down to 20Hz but not have that same driver go higher than 150 Hz (ish). Is it limited by the size of the driver required to move that air?

This is more or less explained in the post I linked to and was under the assumption that to achieve the long pathlengths you will use a folded baffle. These are referred to as U-frame, H-frame, M-frame, Z-frame - some letter and then "frame". I just refer to them generically as 'frames (the apostrophe represents a missing letter).

In these systems, the response can be thought of as a combination (sum of SPLs) from the driver and the baffle or 'frame. For planar open baffle, and all the 'frames, the response looks the same at low frequencies and is determined by the front-to-back pathlength. What these look like, in general, is this:
2pt-src2.gif

Credit: Linkwitzlab, Phoenix, dipole woofer, electro-acoustic models
Caption: the above is the response of a dipole formed by two point sources

The figure above shows a dipole response. It's a good illustration of the basic behavior of a dipole system, no matter how you create it. This is like the "response of the 'frame" that I mentioned above because it does not include the driver's response. There are a couple main characteristics:
DIPOLE PEAK: at 680Hz in the figure, this is determined by the length of the path between the front and back of the radiator (the driver). This is at +6dB WRT the driver's reference level because the front and back waves are adding inphase at this frequency.
DIPOLE NULL: at twice the frequency of the peak there is an on-axis null. Note that on other axis the null occurs at different (higher) frequencies and the responses are crossing.
LOW FREQUENCY LOSSES: below the dipole peak the response falls off at 6dB per octave. Note that the off axis responses all perfectly follow the on axis (upper most) response. That's constant directivity.

Operating below the dipole peak allows very predictable and constant directivity response from the system.

There are other response effects not shown in this plot. The 'frame constructions have internal (transmission line) resonances that create additional nulls in the response at 1/4 wavelength of the depth of the 'frame. Even planar baffles, when much larger than the driver diameter, will have additional nulls forming in the on-axis response (and off axis too). The larger the baffle is compared to the driver, the worse the effect. This was presented by Leo L. Beranek & Tim J. Mellow in their 2012 book on acoustics and is not for a point source but a 6.5" piston source. It can be a dip of 12dB, which is very significant. The moral is: to reduce this effect don't put a small driver in a large baffle, instead use a large driver in a minimal baffle or without baffle (just hanging there by itself).

But back to your question about why a system would be limited to 150Hz. To make a system operate with sufficient SPL at 20-40Hz you need a driver with a free-air response that is not too rolled off there. But also you need a long front-to-back pathlength so that the losses from the 6dB downward trend at low frequency in the 'frame (dipole) response is not too large. Instead of using a huge planar baffle, you can fold back the baffle into a 'frame. A long front-to-back pathlength means the frame will be deep enough to form a short transmission line, and this causes the null from the 1/4 wave resonance of the frame to be low in frequency. If the frame depth is 10 inches, the null will be located at 345Hz but the response is already falling off into the null by about 150Hz, so this is about the upper end of the useful response. You can move the null higher in frequency by making the frame less deep, but the your low frequency losses will be greater.

Keep in mind in addition to the response of the frame, the DRIVER has its own response. It's similar to the free-air response that you can guesstimate from the MFG datasheet TS parameters. Most woofer/subwofoers have a Qts less that about 0.5 so that they can be used in sealed and vented boxes. The lower this number the more the free air response rolls off around resonance. Like this:
An externally hosted image should be here but it was not working when we last tested it.

Credit: mh-audio.nl, Calculate Baffle Diffraction Loss

The above shows the response of a driver (woofer) in a planar open baffle (so no line resonance like the frames have), with various values for the driver Qts and a driver Fs of around 60Hz (an arbitrary choice for the purposes of illustration I guess). Notice how rolled off the system response is when Qts=0.5? Remember, Qts=0.5 is already on the high end for most hi-fi drivers. By contrast, the Qts=2.0 response is more or less flat down until the 60Hz resonance frequency is reached. This is why using a high Qts driver can be beneficial in an open baffle or diple system. It helps to make up for the losses of the dipole system itself, and remember it is the combination of the two that is determining the overall response, including transient response. Let me repeat that for emphasis - the transient response of the system is its overall response. You would never want a closed box having a Q of 2.0! But this case is different.

One additional point about driver selection - efficiency helps offset losses. Your typical home subwoofer driver might have Qts=0.35, Fs=25Hz, and SPL@1m,1W=85dB (or less!). This is not very well suited to a diple system, especially a planar one (meaning NOT one of the 'frame systems). The efficiency and Qts are too low, and the low end SPL even with lots of power applied will be weak. Now consider a "pro audio" driver with Qts=1.26, Fs=40Hz, and SPL@1m,1W=97dB. The high Qts helps to compensate for the dipole losses around the resonance point. The efficiency is high (more than 10dB higher than the home sub), which means that even with some losses the output will be reasonably high with only modest input power. The problem is that Fs is a bit high at 40Hz. Not much of a sub, right?

There is a way to decrease the driver's Fs AND increase Qts: by using an H-frame. The air within the H-frame couples to the driver's own mass and changes the free air response by loading the diaphragm. It's as if you added a weight to the driver cone. This driver in a short H-frame will actually have an Fs of about 30Hz. It's been well documented because the driver I am describing is the Eminence Alpha 15A and it was studied by Martin J. King back in 2008. You can read it here:
http://www.quarter-wave.com/OBs/U_and_H_Frames.pdf

This is one reason why I suggest to use an H-frame for low frequencies. But at the same time while helping in the lowest octave, its resonances limit the upper end. This is why I mentioned the 150Hz limit, which is what you end up with for pretty much any non high-Qts driver like the Alpha 15A. It's just a consequence of the other design decisions and the resulting dimensions of the H-frame.
 
Last edited:
here is my experience with OB bass:
- never attempt to push a driver below its Fs (if using DSP, steep high pass at Fs)
- any frame with cavity, lowpass at where the cavity resonance starts (60-80hz for ripol, slot loaded etc.... 100-200hz for U/H/M etc frame)
- no concern for neighbors; single corner sub (sealed/ported box) lowpassed around 40-50hz depending on room size... dipole above 40-50hz
- concerned with neighbors; 1 or 2 18" ripol subs placed nearfield and lowpassed around 60-80hz
 
Limited options

Hello, I recently had an accident and have since been released from the hospital. After decades of making and selling custom home speakers, I haven't yet used up quite all of my back up drivers, in case of failures. I now need to simplify builds, and have come up with a different idea, for me at least.
I'm thinking an OB build, on 90° folded baffle floorstanders. It'll use most of the Mark Audio drivers I have left, per channel a single Alpair 12P vertically centering two Alpair 12PW drivers.
I'm thinking they'll be so close they almost touch, and cross them at a quarerwave of driver to driver centre spacing, about 350 hertz. First order, asymmetric, roughly Butterworth on the lower frequency, a little more than a Solen Split on the highpass, to help flatten a resonance near its lower range.
I think this will ease the bass drag on my 10 watts into 8 ohms class a amp, and match efficiencies, somewhat, into the 15 available watts for a 4 ohm load down low. I may push the crossover to 500 or 600 hertz, as I like a midbass coherence up to there. This keeps my parts count low, maybe a single cap highpass, and a single coil lowpass.
Deepest bass would be to any of a few bass transducers I have already, with their own amifiers, class ab, a, or d, depending on their power handling and efficiencies.
I like this idea in principle, because
A, it uses parts I already have, in a simpler design than one I might do with different parts available.
B, it cuts my driver count from 5 without a sub, to three without a sub and setting my router to just one size instead of two, and
C, trying to get away without the complication of flush mounting them.
D, no power tapering extra wiring with that, and the complication and parts that go with bandpassing 4 out of five tweeters.
Does this idea seem unreasonable?
I think the highish qts of the 12P driver won't matter, as it's kept to outputting fundamentals above 400-600hertz, and the 12PW pairs won't be reproducing deep bass, therefore I won't be straining their suspensions.
Edit: the baffle will be 19mm MDF, phi spaced bracing, but not extensive. 11" wide on the front, 12" deep on one side, 47" tall on both side, round overs everywhere, a shallow wing on the other side of the front, altering the output of one 12PW driver.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.