I’m looking to build a simple three-way open baffle using this 15 inch driver and matching it to a GRS 10 inch planar mid range. I’d like to have the mid range hi pass filter start around four or 500hz. The front baffle of the subwoofer is drawn at 21 inches in my CAD but can be adjusted somewhat. I’d like to have a U Frame rather than an H frame for aesthetic reasons but I’m having trouble calculating the size. Someone told me I should cross the subwoofer at 100 Hz and the mid range of 500 Hz due to open baffle effects. So does the U frame get calculated on 100 Hz? The MJK paper on U frame subwoofers says they should be crossed at 200 Hz, so should I go with that?
Also, the midrange/tweeter baffle shape has to stay somewhat similar to my CAD due to it being a WAF-approved design, but should that upper section be a U frame as well? Can it be (for matching aesthetics)?
An example of a similar design to my own:
15" High Excursion Paper Cone Subwoofer Driver 4 Ohm
Thiele-Small Parameters
Resonant Frequency (Fs) 34.5Hz
DC Resistance (Re) 3.79Ω
Voice Coil Inductance (Le) 1.96mH
Mechanical Q (Qms) 8.03
Electromagnetic Q (Qes) 0.86
Total Q (Qts) 0.78
Compliance Equivalent Volume (Vas) 3.46ft³
Mechanical Compliance of Suspension (Cms) 0.121mm/N
BL Product (BL) 12.96T·m
Maximum Linear Excursion (Xmax) 7.5mm
Surface Area of Cone (Sd) 755cm²
Also, the midrange/tweeter baffle shape has to stay somewhat similar to my CAD due to it being a WAF-approved design, but should that upper section be a U frame as well? Can it be (for matching aesthetics)?
An example of a similar design to my own:

An externally hosted image should be here but it was not working when we last tested it.
15" High Excursion Paper Cone Subwoofer Driver 4 Ohm
Thiele-Small Parameters
Resonant Frequency (Fs) 34.5Hz
DC Resistance (Re) 3.79Ω
Voice Coil Inductance (Le) 1.96mH
Mechanical Q (Qms) 8.03
Electromagnetic Q (Qes) 0.86
Total Q (Qts) 0.78
Compliance Equivalent Volume (Vas) 3.46ft³
Mechanical Compliance of Suspension (Cms) 0.121mm/N
BL Product (BL) 12.96T·m
Maximum Linear Excursion (Xmax) 7.5mm
Surface Area of Cone (Sd) 755cm²
Last edited:
Regarding the crossover point to the subwoofer:
It's a design choice and related to both how much power you have to driver the woofer and how loud you need it to play at its lowest frequency.
I like to start by calculating what on-axis frequency response I might expect if a given driver is used in an open baffle of a certain dimensions. The free field FR can be calculated by combining the infinite closed box response of the driver PLUS the open baffle gain/loss curve (it's like the diffraction or baffle-step curve for a closed box speaker). You calculate the infinite baffle closed box response using any box modeling program and using Vbox of 100 or 1000 times Vb. You calculate the gain/loss curve for the OB using a modeling program. For ease I use "The Edge" by Tolvan Data. This lets you draw the baffle shape and position a "mic" at some distance away from some position on or off the baffle. This will give you the response curve that the driver would produce relative to the nominal driver level, if that driver had infinite bandwidth (e.g. response down to 0Hz and to infinity). By combining (adding the SPL levels) of these two models you will get an idea of how much SPL loss you will have at a given frequency, e.g. as the frequency falls below 100Hz.
At the same time you will know that you have X Watts of power available, either because that is the power of the amp you can use, or it is the max rated power for the driver. Given this power, find the frequency where the driver will start to exceed Xmax. Using less power means less total SPL output, but you will be able to operate it to a lower frequency before Xmax is exceeded. Conversely, higher input power means more SPL output, but Xmax will be exceeded at a higher frequency. Take a look at this tradeoff when doing the closed box modeling and then decide on what power you will be using.
You have to balance all of these points to answer your question about "where should I cross over to my subwoofer". Only you can answer this question, so do the modeling work first, then come back here and post screen shots of the results. That will make for some good discussion.
It's a design choice and related to both how much power you have to driver the woofer and how loud you need it to play at its lowest frequency.
I like to start by calculating what on-axis frequency response I might expect if a given driver is used in an open baffle of a certain dimensions. The free field FR can be calculated by combining the infinite closed box response of the driver PLUS the open baffle gain/loss curve (it's like the diffraction or baffle-step curve for a closed box speaker). You calculate the infinite baffle closed box response using any box modeling program and using Vbox of 100 or 1000 times Vb. You calculate the gain/loss curve for the OB using a modeling program. For ease I use "The Edge" by Tolvan Data. This lets you draw the baffle shape and position a "mic" at some distance away from some position on or off the baffle. This will give you the response curve that the driver would produce relative to the nominal driver level, if that driver had infinite bandwidth (e.g. response down to 0Hz and to infinity). By combining (adding the SPL levels) of these two models you will get an idea of how much SPL loss you will have at a given frequency, e.g. as the frequency falls below 100Hz.
At the same time you will know that you have X Watts of power available, either because that is the power of the amp you can use, or it is the max rated power for the driver. Given this power, find the frequency where the driver will start to exceed Xmax. Using less power means less total SPL output, but you will be able to operate it to a lower frequency before Xmax is exceeded. Conversely, higher input power means more SPL output, but Xmax will be exceeded at a higher frequency. Take a look at this tradeoff when doing the closed box modeling and then decide on what power you will be using.
You have to balance all of these points to answer your question about "where should I cross over to my subwoofer". Only you can answer this question, so do the modeling work first, then come back here and post screen shots of the results. That will make for some good discussion.
Then regarding the subwoofer size:
So once you have decided the frequency at which you will cross over the woofer, you can size your subwoofer.
The issue here is that you want to make your U-frame subwoofer as large as possible to reduce the dipole losses that will happen at low frequencies. As you make the depth of the U-frame larger/longer the dipole peak and rear resonance will move down in frequency. You want to operate the sub below this frequency (unless you will be using an active crossover, then you can operate right up to the dipole peak). In MJK's study, he uses the U-/H- frames up to about 150Hz or 200Hz. But if you are crossing over the woofer at 100Hz or 80Hz, you are undersizing the subwoofer and it should be larger.
So figure out the woofer crossover point first, and then we can work out the sub design.
So once you have decided the frequency at which you will cross over the woofer, you can size your subwoofer.
The issue here is that you want to make your U-frame subwoofer as large as possible to reduce the dipole losses that will happen at low frequencies. As you make the depth of the U-frame larger/longer the dipole peak and rear resonance will move down in frequency. You want to operate the sub below this frequency (unless you will be using an active crossover, then you can operate right up to the dipole peak). In MJK's study, he uses the U-/H- frames up to about 150Hz or 200Hz. But if you are crossing over the woofer at 100Hz or 80Hz, you are undersizing the subwoofer and it should be larger.
So figure out the woofer crossover point first, and then we can work out the sub design.
Hold on, is the subwoofer part of the 3-way? E.g. Subwoofer-midrange-tweeter? Sorry, I didn't look at the pics before posting (this is after lots of coffee!!!).
I would strongly advise using a U-frame that way for the "subwoofer" driver (kind of a misnomer in this case). It would be better to use a planar baffle for all three drivers. That's because the U-frame will need to be very short to move the tunnel resonance well above the crossover frequency, and so it will be rather useless for extending the bass response, which is typically the entire purpose of the U-frame construction.
Just because someone built a speaker that way (grey painted speaker pic from your first post) doesn't mean that is a good design approach that you should copy. There are multiple design flaws.
I would strongly advise using a U-frame that way for the "subwoofer" driver (kind of a misnomer in this case). It would be better to use a planar baffle for all three drivers. That's because the U-frame will need to be very short to move the tunnel resonance well above the crossover frequency, and so it will be rather useless for extending the bass response, which is typically the entire purpose of the U-frame construction.
Just because someone built a speaker that way (grey painted speaker pic from your first post) doesn't mean that is a good design approach that you should copy. There are multiple design flaws.
Last edited:
Thank you for all the information [mention]CharlieLaub [/mention]. I’ll download the edge tonight and see if I can figure it out.
In general: it’ll be used for a lot of low-bass and bass heavy music, and I’ll have a 300wpc@4ohm amp powering them. I have a separate ported subwoofer already, but if these are able to stand on their own without assistance, it would be nice to hear the OB sound alone.
What do you mean by “they should be in a planar baffle” since it’s a 3 way? Is that just a flat baffle for all the drivers? If the front dimensions can stay somewhat similar, that would work too.
In general: it’ll be used for a lot of low-bass and bass heavy music, and I’ll have a 300wpc@4ohm amp powering them. I have a separate ported subwoofer already, but if these are able to stand on their own without assistance, it would be nice to hear the OB sound alone.
What do you mean by “they should be in a planar baffle” since it’s a 3 way? Is that just a flat baffle for all the drivers? If the front dimensions can stay somewhat similar, that would work too.
It's because you indicated that you want to (and rightly so) cross the woofer over to the midrange at 500Hz or so.What do you mean by “they should be in a planar baffle” since it’s a 3 way?
A U-frame is made by forming a "tunnel" behind the driver. Even if this tunnel is relatively short, it forms a 1/4 wave resonator. This cannot really be "fixed" in the U-frame, so it is operated below the frequency of the tunnel resonance. To make the frequency of the tunnel resonance high, you make the tunnel short. But if you want to operate the driver up to 500Hz the tunnel will be quite short. Since it is also relatively wide (the width of your baffle) it's not all that different from a planar baffle.
The advantage of a planar baffle is that there is absolutely no resonance associated with it (like the U-frame has). Of course there is the dipole response, but that is not a "resonance" per se. The disadvantage is that as you make the front-to-back pathlength shorter the bass losses grow larger. But at the same time you can use the planar baffle driver as high in frequency as you like, and it is really only the driver itself that is limiting how high you cross over to the midrange.
Finally, you mention that you will be listening to bass heavy music. Even with a short U-frame you will not really get strong deep bass. Maybe down to 30Hz after lots of EQ. 300W is a good amount of power, but you will find that the combination of the dipole losses AND the drivers own drooping response will add up to a lot of low frequency SPL loss, and even 300W is not going to be all that much power, probably not enough to get you the "lots of low bass" you crave.
Also you should understand that in a modest sized room a dipole bass system is limited by the room size, and the room itself will be attenuating the bass output to some degree. This is because the dipole is a velocity source and in small spaces the confinement tends to exaggerate the front+back cancellation from the dipole.
If you can use the system in a large room, with a dimension well over 20 feet, then you can get some good LF from a dipole. But since you already have one right there I would just stick with your existing subwoofer (not sure exactly what it is) for the lowest frequencies and use the OB above 80Hz or 100Hz, crossing over to the sub below that point. This relaxes the need to prioritize low frequency SPL, and you can use a planar baffle, and the baffle does not need to be large. A large baffle makes an acoustic shadow of the front wall reflection, which is needed to make the OB have the best sound stage, etc.
I recently made some measurements of the GRS 5010 planar (nude, no baffle) that you can see here, first the frequency response and then the distortion:
The PT5010 is a great driver, but it has rising distortion and a falling frequency response below 700Hz. This is only somewhat moderated by using it in a baffle with open back. So you need to make sure your woofer can operate that high, and so the U-frame is really not well suited to do that, or is essentially just like an OB because it has to be very short.
Hi,
Unfortunately Edge can't model tunnel resonance that happens in U-frame. Hornresp can do that, but it is pretty tricky to start using...
John Kreskowsky has done some modelling and IRL tests that you should read
Tech
And then comes the room and positioning of dipole woofers... What will the floor and front wall do? Floor will extend the baffle and front wall will work as reflector and project the backwave towards the listener louder and delayed, making also many more interference nullings as frequency goes up.
Good luck with your speakers, be prepared to many surprises! Measuring, analyzing and making conclusions followed by design choices will give you many sleepless nights.
Please remember that Linkwitz, Kreskowsky, Laub and myself have ended up with 4-way .
Unfortunately Edge can't model tunnel resonance that happens in U-frame. Hornresp can do that, but it is pretty tricky to start using...
John Kreskowsky has done some modelling and IRL tests that you should read
Tech
And then comes the room and positioning of dipole woofers... What will the floor and front wall do? Floor will extend the baffle and front wall will work as reflector and project the backwave towards the listener louder and delayed, making also many more interference nullings as frequency goes up.
Good luck with your speakers, be prepared to many surprises! Measuring, analyzing and making conclusions followed by design choices will give you many sleepless nights.
Please remember that Linkwitz, Kreskowsky, Laub and myself have ended up with 4-way .
Last edited:
Yes, that's because you all mostly want dipole, open baffle requirements aren't quite as stringent. 😉
I did a quick model of your baffle in The Edge. This is for a planar baffle, on axis, 3m distant from the position labeled "mic" in the attached pic.
This is the effect of the open baffle only. You have to add this on top of the infinite closed box response of the driver. I calculated that, shown in the second (frequency response) and third (excursion) attachments, for 300W input power.
The closed box modeling shows the driver has extension to around 35Hz, where it is rolling off as usual. But adding on top of that the dipole response from The Edge, the SPL will have the additional "down tilt" and an SPL loss of around 8dB there. Also, you can see by the excursion plot that Xmax is exceeded with 300W at 55Hz. One way to look at it is that when operated "unloaded" (in free air) it takes much less power to get the driver to reach Xmax. There is no air-spring in a closed box to stop it. The red line in the SPL response plot shows the Xmax limited SPL. At 55 Hz this is 113dB plus the -3.5dB from The Edge baffle response. That is still pretty high, but 55Hz isn't "low bass" at all. Below about 50Hz the situation gets increasingly worse, so I would not expect this to be a "low bass machine" and you would easily be able to driver the sub into its mechanical limits.
So you COULD still try to use this driver, in this baffle, to low frequency. You mentioned that you would be using a passive crossover. The only way for a passive crossover to EQ the frequency response is to "cut" power, that is to divert it away from the driver. Because of the response shape from the OB, you choose the LF corner and then cut power at 6dB/oct above that point in order to flatten out the baffle response shape. The lower in frequency this is done, the larger an inductor ($$$) you will need. This also means that the effective woofer sensitivity is set by the low corner frequency, since you will be cutting power above that point. When the low corner is low in frequency, the woofer sensitivity will be made very low. In this case, even if you could operate the woofer to 40Hz or so, the effective sensitivity will be the woofer's voltage sensitivity 92dB@2.83V minus the baffle loss in dB, which is about 4-5dB. So the woofer is sitting at around 87-88dB sensitivity, which is barely average.
On the other hand, as soon as you give up trying to take the driver that low, things look pretty good and the design of the crossover also gets much easier. Let's say you cross over at 100Hz and still use a passive crossover. You will still need to cut the power above 100Hz, but this time the 100Hz sensitivity is 113dB plus almost 2dB from the open baffle response - yes the OB response is actually GREATER than the closed box response on axis at this frequency. So the woofer's effective sensitivity rises to 94dB. That's pretty good! Since the woofer typically has the most losses of all the drivers in the OB system, this is setting the total system sensitivity and you can enjoy very dynamic transients because your amp will have lots of headroom, which is not used up trying to bring up the low frequency response. Also, the woofer will only need about 2-3mm of excursion at 100Hz to reach that 113 dB SPL, so it stays more linear and distortion will be lower. Similar things happen if your crossover point is set to 80Hz, with sensitivity a bit lower, Xmax requirements a bit higher, and so on.
As you can see, there are many benefits for not trying to take a woofer too low in an OB if it also has to play "high" in frequency. You COULD add another driver to the OB to make a 4-way so that the woofer only needs to play UP to 150-200Hz, and then using a U-frame would make more sense but that would also make for a more complicated passive crossover design (but this is a valid approach for an all active system). You COULD make the baffle much larger to reduce low frequency losses, but this can impact the response around 500HZ and you end up with a pair of monster walls as speakers that act more like an infinite baffle than an OB. So that is why, in my own systems, I have opted to go with a 3-way OB plus a single large central subwoofer, either H-frame or closed box depending on how large the room is.
This is the effect of the open baffle only. You have to add this on top of the infinite closed box response of the driver. I calculated that, shown in the second (frequency response) and third (excursion) attachments, for 300W input power.
The closed box modeling shows the driver has extension to around 35Hz, where it is rolling off as usual. But adding on top of that the dipole response from The Edge, the SPL will have the additional "down tilt" and an SPL loss of around 8dB there. Also, you can see by the excursion plot that Xmax is exceeded with 300W at 55Hz. One way to look at it is that when operated "unloaded" (in free air) it takes much less power to get the driver to reach Xmax. There is no air-spring in a closed box to stop it. The red line in the SPL response plot shows the Xmax limited SPL. At 55 Hz this is 113dB plus the -3.5dB from The Edge baffle response. That is still pretty high, but 55Hz isn't "low bass" at all. Below about 50Hz the situation gets increasingly worse, so I would not expect this to be a "low bass machine" and you would easily be able to driver the sub into its mechanical limits.
So you COULD still try to use this driver, in this baffle, to low frequency. You mentioned that you would be using a passive crossover. The only way for a passive crossover to EQ the frequency response is to "cut" power, that is to divert it away from the driver. Because of the response shape from the OB, you choose the LF corner and then cut power at 6dB/oct above that point in order to flatten out the baffle response shape. The lower in frequency this is done, the larger an inductor ($$$) you will need. This also means that the effective woofer sensitivity is set by the low corner frequency, since you will be cutting power above that point. When the low corner is low in frequency, the woofer sensitivity will be made very low. In this case, even if you could operate the woofer to 40Hz or so, the effective sensitivity will be the woofer's voltage sensitivity 92dB@2.83V minus the baffle loss in dB, which is about 4-5dB. So the woofer is sitting at around 87-88dB sensitivity, which is barely average.
On the other hand, as soon as you give up trying to take the driver that low, things look pretty good and the design of the crossover also gets much easier. Let's say you cross over at 100Hz and still use a passive crossover. You will still need to cut the power above 100Hz, but this time the 100Hz sensitivity is 113dB plus almost 2dB from the open baffle response - yes the OB response is actually GREATER than the closed box response on axis at this frequency. So the woofer's effective sensitivity rises to 94dB. That's pretty good! Since the woofer typically has the most losses of all the drivers in the OB system, this is setting the total system sensitivity and you can enjoy very dynamic transients because your amp will have lots of headroom, which is not used up trying to bring up the low frequency response. Also, the woofer will only need about 2-3mm of excursion at 100Hz to reach that 113 dB SPL, so it stays more linear and distortion will be lower. Similar things happen if your crossover point is set to 80Hz, with sensitivity a bit lower, Xmax requirements a bit higher, and so on.
As you can see, there are many benefits for not trying to take a woofer too low in an OB if it also has to play "high" in frequency. You COULD add another driver to the OB to make a 4-way so that the woofer only needs to play UP to 150-200Hz, and then using a U-frame would make more sense but that would also make for a more complicated passive crossover design (but this is a valid approach for an all active system). You COULD make the baffle much larger to reduce low frequency losses, but this can impact the response around 500HZ and you end up with a pair of monster walls as speakers that act more like an infinite baffle than an OB. So that is why, in my own systems, I have opted to go with a 3-way OB plus a single large central subwoofer, either H-frame or closed box depending on how large the room is.
Attachments
Last edited:
CharlieLaub
So it appears that my best option is to make them with a totally flat baffle and then use my sub to support them. The sub is a 2x12" dayton titanic w/ a 1000w amp, should be plenty 😀.
As for room size, its 22' wide and 16' deep, 10ft ceilings. Current setup shown. Speakers are about 12' apart and I sit maybe 12' from them with the back wall right behind me.
I was hoping to make these speakers as a learning tool without a mind-numbing amount of research and design put into them. They're mostly an experiment with OB to see if I want to incorporate it into a later design (that may end up being a 4 way, as it just so happens). So if I can follow some general rules of thumb and they get a thumbs up from the guru's, I'll be happy. I'll probably move these to a less used room someday.
Currently I'm leaning towards the suggested flat planar baffle and my powered sub for low end support. Does the 21" wide circular baffle in my cad model seem sufficient? I could go up to 24" or so without much issue, but maybe it doesn't matter if the separate sub is going to be filling in.
So it appears that my best option is to make them with a totally flat baffle and then use my sub to support them. The sub is a 2x12" dayton titanic w/ a 1000w amp, should be plenty 😀.
I'll have a read on that site later. Looks like lots of good info.John Kreskowsky has done some modelling and IRL tests that you should read
Tech
As for room size, its 22' wide and 16' deep, 10ft ceilings. Current setup shown. Speakers are about 12' apart and I sit maybe 12' from them with the back wall right behind me.
I was hoping to make these speakers as a learning tool without a mind-numbing amount of research and design put into them. They're mostly an experiment with OB to see if I want to incorporate it into a later design (that may end up being a 4 way, as it just so happens). So if I can follow some general rules of thumb and they get a thumbs up from the guru's, I'll be happy. I'll probably move these to a less used room someday.
Currently I'm leaning towards the suggested flat planar baffle and my powered sub for low end support. Does the 21" wide circular baffle in my cad model seem sufficient? I could go up to 24" or so without much issue, but maybe it doesn't matter if the separate sub is going to be filling in.
Attachments
Last edited by a moderator:
The rational way to decide on the baffle width is to first decide at what frequency you think you want to cross over to your subwoofer and then make the baffle only as wide as needed so that the woofer will be able to operate down to the crossover frequency and meet the SPL target. Remember there are two contributions: the "free air" or "infinite baffle boxed" response of the driver AND the baffle gain/loss curve, and it is only the baffle gain/loss curve that is changing with the width/size of the baffle.
As a first approximation, the shape of the baffle gain/loss curve will remain the same but will move up or down in frequency depending on the baffle width (this is valid for low frequencies below the dipole peak). Think of the baffle width as moving the baffle gain/loss curve "left and right" on the plot. Narrower/smaller baffles will increase low frequency losses and will move the curve to the right. Also, the baffle width will determine the system sensitivity because of where you will start to cut power to flatten the woofer frequency response. Look at the example I gave a couple of posts ago for the 100Hz crossover point that results in a 94dB@2.83 sensitivity, and then decide if you want more extension (a lower crossover point to the sub) at the cost of lower overall sensitivity or not. 12" vs 24" baffle width is not going to make a whole lot of difference, maybe 0.5dB, and circular or more linear is not going to influence the woofer's response below 500Hz much at all.
As a first approximation, the shape of the baffle gain/loss curve will remain the same but will move up or down in frequency depending on the baffle width (this is valid for low frequencies below the dipole peak). Think of the baffle width as moving the baffle gain/loss curve "left and right" on the plot. Narrower/smaller baffles will increase low frequency losses and will move the curve to the right. Also, the baffle width will determine the system sensitivity because of where you will start to cut power to flatten the woofer frequency response. Look at the example I gave a couple of posts ago for the 100Hz crossover point that results in a 94dB@2.83 sensitivity, and then decide if you want more extension (a lower crossover point to the sub) at the cost of lower overall sensitivity or not. 12" vs 24" baffle width is not going to make a whole lot of difference, maybe 0.5dB, and circular or more linear is not going to influence the woofer's response below 500Hz much at all.
Thanks again Charlie. I just read your post previous to this one with the graphs. If moderate changes in the baffle width aren't going to change the response much, think I'll go with essentially what I drew but in your suggested flat baffle.
I'll look into how that woofer sounds up to around 500hz, as a midwoofer may be needed. Maybe for simplicity's sake I'll lose the tweeter and have the GRS 10" play all the way up and put a notch filter on that bad 8k peak it has.
The subwoofer I have now can have a 120hz down filter so I could high pass the 15" above that to help with excursion problems.
I'll look into how that woofer sounds up to around 500hz, as a midwoofer may be needed. Maybe for simplicity's sake I'll lose the tweeter and have the GRS 10" play all the way up and put a notch filter on that bad 8k peak it has.
The subwoofer I have now can have a 120hz down filter so I could high pass the 15" above that to help with excursion problems.
Ahem, Oveboard,
You haven't told us if you are making the speaker passive or active, and if the xo will be analog or dsp. So, do we need to consider different efficiencies of drivers and ways of making dipole loss eq?
Scottjoplin - "open baffle" speakers typically have disastereously bad 3D dispersion through spectrum, because the designers don't have a clue about directivity basics and it's importance. Simply making the woofer and perhaps the 2" cone (instead of a monopole twweeter/horn) to haphazardly radiate similar amount of energy backwards and not having cabinet resonances is not what I would call modern and good loudspeaker design. But when you have smooth dipole radiation over wider range, and most challengingly between 300-5khz, and can locate these speakers in a big enough room like Overboard has, you can hear angels sing!
Here is a directivity sim of a "dipole" which sharply becomes monopole at 1kHz, presented by kimmosto in another thread is a good example of bad OB design IMO. I guess it would be very difficult to make it sound good in a normal room at home. https://www.diyaudio.com/forums/multi-way/371217-baffle-speakers-irkiosan-3.html#post6631499
You haven't told us if you are making the speaker passive or active, and if the xo will be analog or dsp. So, do we need to consider different efficiencies of drivers and ways of making dipole loss eq?
Scottjoplin - "open baffle" speakers typically have disastereously bad 3D dispersion through spectrum, because the designers don't have a clue about directivity basics and it's importance. Simply making the woofer and perhaps the 2" cone (instead of a monopole twweeter/horn) to haphazardly radiate similar amount of energy backwards and not having cabinet resonances is not what I would call modern and good loudspeaker design. But when you have smooth dipole radiation over wider range, and most challengingly between 300-5khz, and can locate these speakers in a big enough room like Overboard has, you can hear angels sing!
Here is a directivity sim of a "dipole" which sharply becomes monopole at 1kHz, presented by kimmosto in another thread is a good example of bad OB design IMO. I guess it would be very difficult to make it sound good in a normal room at home. https://www.diyaudio.com/forums/multi-way/371217-baffle-speakers-irkiosan-3.html#post6631499
Last edited:
Powered passively and analog XO. I won’t be surprised if I need an Lpad for the planars once I get to the xo designing stage.
It’s too bad that kimmosto’s design may have some serious flaws, it’s a sharp looking tower.
It’s too bad that kimmosto’s design may have some serious flaws, it’s a sharp looking tower.
Last edited by a moderator:
I didn't say "serious flaws", but 1kHz is in sweet spot for hearing differencies, and there this speaker suddenly stops sending soundwaves backwards. Late arrival reflections are needed for spaciousness, which is just the basic difference of dipole/OB to normal speakers. The diffrence can be measured at listening spot by room decay or wavelets.
^Who will listen to these tall multiway floorstanders in nearfield?
I think it is fair to suppose listening in normal size livingroom stereo setup, distance roughly 2,5- 3,5m ( 8-12'). Distance to walls is a major factor with OB/dipoles, to minimize problems with first reflections and maximize late reflections - setting speakers along the long wall is recommended.
Another good setup is what Linkwitz recommends, a very long room and speakers at least 2m (7') from front wall and lots of open space behind the listener as well.
Each setup changes bass boost, first reflections and timing and relative spl of backside and lateral late reflections. These changes apply for every type of speakers, horns and cardioids the least and dipoles the most. And no doubt the differences are easy to hear, even if spl response at spot is equalized!
Here impulse responses at spot of AINOgradient vs. normal on-stand 6,5"/1" 2-way with a tiny waveguide. Reflections at 1-3ms obviously come from the 50" TV next to the speaker, which is slightly aimed at the spot (at 2m distance in this test). The dipole is more angled so that lateral null is towards the screen.
The normalized horizontal directivity of AINOgradient is in my avatar!
I think it is fair to suppose listening in normal size livingroom stereo setup, distance roughly 2,5- 3,5m ( 8-12'). Distance to walls is a major factor with OB/dipoles, to minimize problems with first reflections and maximize late reflections - setting speakers along the long wall is recommended.
Another good setup is what Linkwitz recommends, a very long room and speakers at least 2m (7') from front wall and lots of open space behind the listener as well.
Each setup changes bass boost, first reflections and timing and relative spl of backside and lateral late reflections. These changes apply for every type of speakers, horns and cardioids the least and dipoles the most. And no doubt the differences are easy to hear, even if spl response at spot is equalized!
Here impulse responses at spot of AINOgradient vs. normal on-stand 6,5"/1" 2-way with a tiny waveguide. Reflections at 1-3ms obviously come from the 50" TV next to the speaker, which is slightly aimed at the spot (at 2m distance in this test). The dipole is more angled so that lateral null is towards the screen.
The normalized horizontal directivity of AINOgradient is in my avatar!
Attachments
Last edited:
OK. So the behaviour of the speaker is starting to have monopole radiation past 1-1.5k and that´s bad because it´s within a sensitive range of frequencies, right?Here is a directivity sim of a "dipole" which sharply becomes monopole at 1kHz, presented by kimmosto in another thread is a good example of bad OB design IMO.
I always assumed that it´s hard to have pure dipole behaviour >2k anyway.
What about your AINOgradient? When I look at your avatar it has no back radiation at lower frequencies in a certain band if I read that graph correct (~blue at 180°). Is that a crossover point? It seems pretty broad.
- Home
- Loudspeakers
- Multi-Way
- Some open baffle subwoofer design help needed