Hi All,
I've been designing loudspeakers for a little while and have a decent grounding on how to model, design, built, test and enjoy loudspeakers with boxes. I typically use SoundEasy for much of the process.
I've recently completed the Linkwitz LX521 and really enjoy them. However, having built the Orion's before, I'm always worried about them receiving sub frequencies at loud volumes. Also, sometimes I want a little more bottom octave. So this is for music more than HT but the system is also used for movies sometimes.
Any suggestions or reading material on how to properly design/model open baffle subs or is it just trial and error? I was thinking about a pair of 18" drivers like the ultimax but would like some idea of Excursion vs Freq at various SPLs. I know SoundEasy can model U and H dipole but I can't say I trust what I'm seeing. I can hit up that user forum but hoped someone had some design equations/documentation so that I can understand what I'm getting into before spending money.
I do understand that excursion will limit SPL much faster in open baffle but I really like the quality of open baffle bass🙂
Sent from my iPhone using Tapatalk
I've been designing loudspeakers for a little while and have a decent grounding on how to model, design, built, test and enjoy loudspeakers with boxes. I typically use SoundEasy for much of the process.
I've recently completed the Linkwitz LX521 and really enjoy them. However, having built the Orion's before, I'm always worried about them receiving sub frequencies at loud volumes. Also, sometimes I want a little more bottom octave. So this is for music more than HT but the system is also used for movies sometimes.
Any suggestions or reading material on how to properly design/model open baffle subs or is it just trial and error? I was thinking about a pair of 18" drivers like the ultimax but would like some idea of Excursion vs Freq at various SPLs. I know SoundEasy can model U and H dipole but I can't say I trust what I'm seeing. I can hit up that user forum but hoped someone had some design equations/documentation so that I can understand what I'm getting into before spending money.
I do understand that excursion will limit SPL much faster in open baffle but I really like the quality of open baffle bass🙂
Sent from my iPhone using Tapatalk
MJK's site (quarter-wave.com) has all kinds of useful information on OB, he also has papers showing U and H baffle design. His designs can be used with passive crossovers, which is fairly uncommon. The driver selection is what makes them suitable for passive crossover.
If you have active crossover and dsp capability that really opens up your choices for driver selection.
But for cavity depth you can't go very deep before you get large peaks and nulls in or near the passband, the peaks can be controlled with dsp but the nulls not so much.
OB is ok in the modal region but makes absolutely no sense whatsoever below the room's modal region. So OB "subwoofers" aren't really subs but more like woofers. A true OB sub doesn't make sense because the required baffle size (or cavity depth) to get that low in frequency with any kind of decent sensitivity would be too large to be practical and because there's no room gain from OB. Below the modal range you want some other type of design - sealed, ported, or whatever, just not OB.
If you have active crossover and dsp capability that really opens up your choices for driver selection.
But for cavity depth you can't go very deep before you get large peaks and nulls in or near the passband, the peaks can be controlled with dsp but the nulls not so much.
OB is ok in the modal region but makes absolutely no sense whatsoever below the room's modal region. So OB "subwoofers" aren't really subs but more like woofers. A true OB sub doesn't make sense because the required baffle size (or cavity depth) to get that low in frequency with any kind of decent sensitivity would be too large to be practical and because there's no room gain from OB. Below the modal range you want some other type of design - sealed, ported, or whatever, just not OB.
Another thread to review....
I suggest that you look through this thread:
http://www.diyaudio.com/forums/full-range/271011-rockin-kazba-dipole-k-aperture-z-baffle-dipole.html
Retsel
I suggest that you look through this thread:
http://www.diyaudio.com/forums/full-range/271011-rockin-kazba-dipole-k-aperture-z-baffle-dipole.html
Retsel
Open baffle subwoofers are a dumb idea all around. This is coming from someone who is very pro open baffle. At some point it is better to give up and switch to monopole, perhaps around 60Hz-80Hz, where baffle size becomes quite large and even folded structures like U- and H-frames start to suffer from serious resonance problems.
There is actually a very nice way to transition the pattern from dipole to cardioid and then monopole, and I hope to write up a white paper about it in the next couple of months. But it's not "open baffle" per se.
There is actually a very nice way to transition the pattern from dipole to cardioid and then monopole, and I hope to write up a white paper about it in the next couple of months. But it's not "open baffle" per se.
Last edited:
Sorry for the delay, got crazy busy🙂 Thanks for the info. I read through MJK's site and learned a bit there.
Can you clarify this statement? I'm not sure I understand what you mean:
I've started reading through the KaZba dipole thread but have to spend a bit more time there to get my head around all of it from an engineering and physics perspective.
I have a miniDSP 4x10HD that I can dedicate solely to this project. While I hear the statements about the U and H frame not working at low frequencies, my experience with the orions and LX521's indicate otherwise. I'm not looking for 15Hz. I'd like to get reasonable SPL from 25-80Hz. I'm thinking in the neighborhood of 110dB @ 1 m with 4x 18" subs. Like stereo H frame cabinets....
I guess I don't understand why there is no advantage to a dipole from 25Hz to 50Hz range? I sure seem to hear a huge advantage in bass clarity in my experience. Really just looking for more output compared to what I have? Please feel free to rip me a new one as I'd rather not throw money away🙂
Can you clarify this statement? I'm not sure I understand what you mean:
I've started reading through the KaZba dipole thread but have to spend a bit more time there to get my head around all of it from an engineering and physics perspective.
I have a miniDSP 4x10HD that I can dedicate solely to this project. While I hear the statements about the U and H frame not working at low frequencies, my experience with the orions and LX521's indicate otherwise. I'm not looking for 15Hz. I'd like to get reasonable SPL from 25-80Hz. I'm thinking in the neighborhood of 110dB @ 1 m with 4x 18" subs. Like stereo H frame cabinets....
I guess I don't understand why there is no advantage to a dipole from 25Hz to 50Hz range? I sure seem to hear a huge advantage in bass clarity in my experience. Really just looking for more output compared to what I have? Please feel free to rip me a new one as I'd rather not throw money away🙂
Hi,
Below the lowest room mode of your room the room is essentially
a pressure chamber and a dipole is totally incapable of driving it.
Look at : Linkwitz Lab - Loudspeaker Design
and : Music and Design
rgds, sreten.
Below the lowest room mode of your room the room is essentially
a pressure chamber and a dipole is totally incapable of driving it.
Look at : Linkwitz Lab - Loudspeaker Design
and : Music and Design
rgds, sreten.
basic tenant 101
The very basic tenant of getting a loudspeaker to produce bass frequencies, is by separating the outputs from the front and rear waves of a given driver.
When using a partial baffle board a.k.a "open baffle" (as opposed to a completely sealed enclosure) there will be quite the degree of cancellation of sound, due to the front and back waves meeting. The now very popular open baffle design realizes this, and attempts to achieve a compromise between effective baffle size and the commensurate required opposite equalization needed to "flatten" the response. The lower in frequency the target is, the larger the baffle must be, to the point where it's simply just a dumb idea (as expressed above).
In short, there is no such thing as an effective open baffle "sub".
The very basic tenant of getting a loudspeaker to produce bass frequencies, is by separating the outputs from the front and rear waves of a given driver.
When using a partial baffle board a.k.a "open baffle" (as opposed to a completely sealed enclosure) there will be quite the degree of cancellation of sound, due to the front and back waves meeting. The now very popular open baffle design realizes this, and attempts to achieve a compromise between effective baffle size and the commensurate required opposite equalization needed to "flatten" the response. The lower in frequency the target is, the larger the baffle must be, to the point where it's simply just a dumb idea (as expressed above).
In short, there is no such thing as an effective open baffle "sub".
Well, that depends.
One of my wild and untested ideas was the use of a horizontal open baffle "tent" subwoofer for PA duty, to avoid the annoying effects of low bass exiting the designated area, possibly resulting in the festivities being brought an early close due to complaints. Basically the "sub" or "subs" consist of a fiberglass shallow dome or pyramid you bolt together from panels (which will provide the necessary stiffness), with the driver mounted open baffle at the apex of the dome or pyramid, and you treat this as the "roof" for your festivity by raising it off the ground and using it as a tent. You should get good bass below it, and the nulls will be directed towards the sides, greatly reducing low bass feed to nearby buildings. And 25 feet of canopy should work pretty well for a small tent 🙂
Well, according to theory. I've never tried it 🙂.
If you do some simulations you will find that for practical size baffles and/or U or H frame depths the sensitivity at 25 hz is absolutely awful. And since there's no room gain with OB the room doesn't support anything lower than the lowest room mode. So even a simple sealed sub will be MUCH louder than any practically sized OB at 25 hz.
You would have to do some sims but it's going to take a LOT of OB subs to get anywhere near 110 db at 25 hz. It's going to take a lot of excursion so unless you use a whole lot of subs it's going to be high distortion.
I use MJK's worksheets for OB simulation but I believe John K's ABC Dipole is a free download and should do what you need.
So let me qualify why I'm pursuing this:
1. The bass in outdoor venues with sufficient subwoofer displacement is absolutely fantastic. The reduced room interaction is exactly why I think the dipole subs have a benefit.
2. The Linkwitz Labs dipoles I own create bass as close to the outdoor venue experience as I've heard in any living room (down into the 30'z)
3. I'm chasing bass quality instead of bass quantity. I don't think I even need 110dB at 1m but figured I'd rarely exceed 100dB A weighted so seems like a good target to match. Typically, I listen in the 80dBA range......
4. I want to leave the HPF of the LX521's active crossover engaged and balance the bottom octave without losing the many benefits of the best sounding bass reproduction I've ever had in my room.
5. This design is for a dedicated space so I'm OK with having something stupid like 4x 18" split between left and right cabinets.
6. HT is completely secondary. I have box subwoofers that I can throw in for LFE only if I want to shake the house down.
7. I like that loud music on dipoles doesn't irritate the neighborhood.
Thank you all for the material posted. I've read through John K's site but must have missed the ABC dipole download, that will be very helpful.
I'm going to dig in and start playing with modeling over the next week to see if any design makes much sense. Based on what I get out of 4x10s in the LX521's, I'm still encouraged that increasing displacement capabilities can help achieve my goals...
1. The bass in outdoor venues with sufficient subwoofer displacement is absolutely fantastic. The reduced room interaction is exactly why I think the dipole subs have a benefit.
2. The Linkwitz Labs dipoles I own create bass as close to the outdoor venue experience as I've heard in any living room (down into the 30'z)
3. I'm chasing bass quality instead of bass quantity. I don't think I even need 110dB at 1m but figured I'd rarely exceed 100dB A weighted so seems like a good target to match. Typically, I listen in the 80dBA range......
4. I want to leave the HPF of the LX521's active crossover engaged and balance the bottom octave without losing the many benefits of the best sounding bass reproduction I've ever had in my room.
5. This design is for a dedicated space so I'm OK with having something stupid like 4x 18" split between left and right cabinets.
6. HT is completely secondary. I have box subwoofers that I can throw in for LFE only if I want to shake the house down.
7. I like that loud music on dipoles doesn't irritate the neighborhood.
Does this mean that there is no bass at the listeners ears below the lowest room mode?
Thank you all for the material posted. I've read through John K's site but must have missed the ABC dipole download, that will be very helpful.
I'm going to dig in and start playing with modeling over the next week to see if any design makes much sense. Based on what I get out of 4x10s in the LX521's, I'm still encouraged that increasing displacement capabilities can help achieve my goals...
FWIW, my tests with a dipole sub show that the claimed "reduced room interaction" is likely overblown. See link below.
The Subwoofer DIY Page v1.1 - Projects : An INF10 Dipole Subwoofer
Hi Brian, I would completely expect the frequency response you measured given a long window. There are definitely reflections with dipoles but there is definitely much less energy hitting room boundaries. In my mind the reduced room excitation improves things significantly.
What was your measurement tool, window length and window function?
Sent from my iPhone using Tapatalk
What was your measurement tool, window length and window function?
Sent from my iPhone using Tapatalk
If I'm reading that right it looks like you are 20 db down at 20 hz and showing no room gain (as expected). 20 db down doesn't sound overblown to me, it sounds like a very significant problem. Extremely low frequency sensitivity, as was mentioned.
That measurement was done some time ago and unfortunately I didn't record the details. It likely was a simple RTA measurement using pink noise and 100 sample average.
In interest of full disclosure, I am planning on making another attempt at a dipole bass unit (I wouldn't call it a "subwoofer") when I've got the opportunity to do so. I've got two drivers that are likely to be a much better match for dipole use (Fs~18 Hz, Qts~0.5, Xmax=12mm and a suspension that will not let the cone bottom), and some spare wood lying around.
Ok, here we go, just for fun.
Using MJK's H frame worksheet (which I haven't even looked at for years now) I've changed just a couple of things. I made the H baffle 24 inches total length (not including the divider board width), so 12 inches for the front chamber and 12 inches for the rear. I assumed you wanted it to be fairly compact.
The driver used is the 18 inch Goldwood from PE, it's got 2 mm xmax. It's OB friendly, especially if you want to use a passive crossover and no dsp. It's $100. But really any driver will work if you have active crossover and dsp. In this case, it's the Goldwood's very high 1.1 qts that allows a somewhat natural response without requiring you to boost the everloving snot out of the low frequencies.
http://www.parts-express.com/goldwood-gw-1858-18-pro-woofer--290-386
Pic 1 shows the driver and H frame parameters. (Never mind the extended range driver.)
Pic 2 shows - RMS displacement with 1 watt power input
- top down schematic of H frame in room corner
- frequency response including ONLY floor boundary
- frequency response with floor boundary, side wall 2 feet from front of H frame, rear wall 3 feet from front of H frame, 8 foot floor to ceiling height, all 4 of these boundaries activated
- all results shown with 80 hz 4th order high pass filter
So we can see that the sensitivity at 25 hz is pretty abysmal, even though the high q driver is naturally boosting the response at low frequencies. With only floor bounce calculated sensitivity is around 83 db. With the 4 nearest boundaries calculated the sensitivity is only 75 db. So clearly placement matters. Also note that a much lower qts "normal" driver would have MUCH worse, MUCH lower sensitivity at 25 hz. It's this driver's high qts that boosts the low frequencies up to a somewhat flat(ish) response in this case.
Also note that 1 watt requires almost all the excursion that this driver can give. The graph shows 1.2 mm displacement but that's RMS displacement, which gives a peak displacement of 1.7 mm. This driver has 2 mm xmax, so with one watt it's just about tapped out as shown here.
Now let's look at how much it would take to get 110 db at 25 hz. Let's just pretend that you can turn up this single driver H frame sub until you get your desired spl. We'll use the sim with only floor bounce (83 db sensitivity) for this as a best case scenario and just add more power. Shown at 500 watts.
26 mm RMS displacement is required to hit 110 db at 25 hz. That's 34 mm peak. If we pretend that this driver could provide that.
Another (more realistic) way to look at it would be to add additional subs instead of pretending this driver could do 34 mm xmax. This one sub can do a bit more than 83 db at 25 hz within it's 2 db xmax, so if we add 6 db (best case scenario) for every doubling of cabs and power we get 107 db from 16 of these. Or 113 db from 32 of them. And remember, that's with the best case scenario without a lot of low end suckouts from boundary bounce causing super low sensitivity in the low frequencies. (Placement matters)
Of course you could use different drivers with WAY more xmax and have less cabs, but you wouldn't have the OB friendly super high qts so you would have to boost the living crap out of the low end to achieve something close to the flat(ish) frequency response that comes naturally for the driver shown in this exercise.
Anyway, good luck with all that.
Using MJK's H frame worksheet (which I haven't even looked at for years now) I've changed just a couple of things. I made the H baffle 24 inches total length (not including the divider board width), so 12 inches for the front chamber and 12 inches for the rear. I assumed you wanted it to be fairly compact.
The driver used is the 18 inch Goldwood from PE, it's got 2 mm xmax. It's OB friendly, especially if you want to use a passive crossover and no dsp. It's $100. But really any driver will work if you have active crossover and dsp. In this case, it's the Goldwood's very high 1.1 qts that allows a somewhat natural response without requiring you to boost the everloving snot out of the low frequencies.
http://www.parts-express.com/goldwood-gw-1858-18-pro-woofer--290-386
Pic 1 shows the driver and H frame parameters. (Never mind the extended range driver.)
An externally hosted image should be here but it was not working when we last tested it.
Pic 2 shows - RMS displacement with 1 watt power input
- top down schematic of H frame in room corner
- frequency response including ONLY floor boundary
- frequency response with floor boundary, side wall 2 feet from front of H frame, rear wall 3 feet from front of H frame, 8 foot floor to ceiling height, all 4 of these boundaries activated
- all results shown with 80 hz 4th order high pass filter
An externally hosted image should be here but it was not working when we last tested it.
So we can see that the sensitivity at 25 hz is pretty abysmal, even though the high q driver is naturally boosting the response at low frequencies. With only floor bounce calculated sensitivity is around 83 db. With the 4 nearest boundaries calculated the sensitivity is only 75 db. So clearly placement matters. Also note that a much lower qts "normal" driver would have MUCH worse, MUCH lower sensitivity at 25 hz. It's this driver's high qts that boosts the low frequencies up to a somewhat flat(ish) response in this case.
Also note that 1 watt requires almost all the excursion that this driver can give. The graph shows 1.2 mm displacement but that's RMS displacement, which gives a peak displacement of 1.7 mm. This driver has 2 mm xmax, so with one watt it's just about tapped out as shown here.
Now let's look at how much it would take to get 110 db at 25 hz. Let's just pretend that you can turn up this single driver H frame sub until you get your desired spl. We'll use the sim with only floor bounce (83 db sensitivity) for this as a best case scenario and just add more power. Shown at 500 watts.
An externally hosted image should be here but it was not working when we last tested it.
26 mm RMS displacement is required to hit 110 db at 25 hz. That's 34 mm peak. If we pretend that this driver could provide that.
Another (more realistic) way to look at it would be to add additional subs instead of pretending this driver could do 34 mm xmax. This one sub can do a bit more than 83 db at 25 hz within it's 2 db xmax, so if we add 6 db (best case scenario) for every doubling of cabs and power we get 107 db from 16 of these. Or 113 db from 32 of them. And remember, that's with the best case scenario without a lot of low end suckouts from boundary bounce causing super low sensitivity in the low frequencies. (Placement matters)
Of course you could use different drivers with WAY more xmax and have less cabs, but you wouldn't have the OB friendly super high qts so you would have to boost the living crap out of the low end to achieve something close to the flat(ish) frequency response that comes naturally for the driver shown in this exercise.
Anyway, good luck with all that.
Last edited:
Sounds like that might be good news for me. My next OB test only has to go down to about 35 Hz or so (I've got another sub to handle the rest) and I'll be using an iNuke 3000DSP to provide the power to the two 12" drivers in separate OB frames. I suspect my biggest issue is going to be finding a good location for them in the living room....
Anyway, to see what the H baffle is actually doing to the driver's response, have a look at the measured frequency response of the driver alone (likely measured on a standard IEC baffle but details on not disclosed on this sheet).
http://www.parts-express.com/pedocs/more-info/290-386-goldwood-gw-1858-more-info.pdf
The Goldwood has what looks to be about 97 db midband sensitivity at 2.83V (which is a bit more power than I simulated, I did 1 watt in reference to Re, not in reference to 8 ohms). Also note the big 8 db spike from the high qts right above fs. That's why the simulated H frame design doesn't need to boost the low end much (if at all) whereas a normal q driver would need a lot of boost at low frequencies.
So the H frame took this 97 db driver and changed it into a ~83 db design, and even that 83 db is only possible because of the outrageously high qts (compared to normal drivers).
That's what OB subwoofers are all about. Extremely low sensitivity and no room gain to help out in the low frequencies.
http://www.parts-express.com/pedocs/more-info/290-386-goldwood-gw-1858-more-info.pdf
The Goldwood has what looks to be about 97 db midband sensitivity at 2.83V (which is a bit more power than I simulated, I did 1 watt in reference to Re, not in reference to 8 ohms). Also note the big 8 db spike from the high qts right above fs. That's why the simulated H frame design doesn't need to boost the low end much (if at all) whereas a normal q driver would need a lot of boost at low frequencies.
So the H frame took this 97 db driver and changed it into a ~83 db design, and even that 83 db is only possible because of the outrageously high qts (compared to normal drivers).
That's what OB subwoofers are all about. Extremely low sensitivity and no room gain to help out in the low frequencies.
That's just crazy.
2 mm Xmax is what I would choose for a midrange working >500 Hz
Try another one !
😡
I chose that one for a few reasons.
1. It's the worksheet default so I didn't have to enter new t/s parameters.
2. I didn't want to guess what type of driver OP wants to use.
3. Any driver with a qts under about 1 will need a tremendous amount of boost in the low bass, that can't be shown adequately in this sim. I can't add boost and therefore can't show how much excursion the boost requires to get flat(ish) response.
4. You can extrapolate these results into other situations. If you know how much displacement is required to hit 25 hz at 110 db you should be able to pretty easily figure out how other drivers would do in this same design.
I could do another sim but I'd probably move it into Hornresp next time. I wanted to show the effect of 4 boundaries, that's why I chose this software. But before doing another sim I'd need some idea of OP's desired driver and design dimensions. And it's probably better if OP did his own sims, you learn a lot better that way.
- Status
- Not open for further replies.