Well, this is not an exactly correct question., I fear The electrical behavior of a driver doesn't change much in vacuum where the sound output is most certainly null ;-)
Back-EMF is a term I don't like personally because it's blurry, I prefer microphonic voltage, and it is important to note that this is a voltage thing, not a power thing. A speaker driver is also an efficient microphone. The point is now, it doesn't matter if the movement is induced by the current from the amp or any external mechanical exitation, the voltage generated by the coil is the same, it only depends on the actual movement of the cone no matter what caused it.
This means the amp has always to set up it's output voltage close to the back-EMF voltage of the coil in order to inject the proper current required to change the motion of the cone according to the input signal.
Ripoles are a form of folded dipole but I would probably suggest against building a Ripole using 18" drivers. Would you care to guess why that might be?
Ripoles are the most inefficient dipole possible... and have the worst (highest Q) line resonance. About the only thing going for a ripole is that they are relatively small in size. There is only vibration cancellation if you build a variety with two drivers with the cones moving in opposite directions to cancel vibration.
A good way to look at it is to compare Q-mechanical (should look like the impedance curve) with Q-electrical (far, far duller) which more closely represents cone motion when connected to an amp.
With acoustic instrument music, the average electric power into a driver is tiny although there is need for peak voltage swings (which turns out to require substantial power ratings for amps) to drive occasional very brief peaks (esp. women's chorus). Somewhere under 5% appears as sound. But 1 watt of sound will destroy your hearing.
B.
Last edited:
A ripole does not necessarily have to have use two drivers. The "two driver" version cancels vibration and that is certainly an advantage.
Here is an example of the "single driver" version:
https://www.lautsprechershop.de/hifi/ripol_en.htm]Strassacker: Speaker Building, Components[/url]
Here is an example of the two driver setup:
https://www.lautsprechershop.de/hifi/ridtahler_en.htm]Strassacker: Speaker Building, Components[/url]
You are correct in that the ripole is similar to the dipole subwoofer that SL wrote about, only with smaller openings.
I prefer larger openings to keep the Q of the dipole peak low, so it makes more sense to use an H-frame. The H-frame gives you independent control over the width and depth of the "tunnels" that generate the resonance. This means I can choose whatever dipole peak frequency works best for the application, as well as get control over the Q of the line resonance and make is low enough that I can include it in the passband.
In contrast the Ripole makes the resonances as bad as possible. This means you need to cross over well below the resonance peak, and as a result you have to throw away the "higher SPL" region of the dipole response.
Here is an example of the "single driver" version:
https://www.lautsprechershop.de/hifi/ripol_en.htm]Strassacker: Speaker Building, Components[/url]
Here is an example of the two driver setup:
https://www.lautsprechershop.de/hifi/ridtahler_en.htm]Strassacker: Speaker Building, Components[/url]
You are correct in that the ripole is similar to the dipole subwoofer that SL wrote about, only with smaller openings.
I prefer larger openings to keep the Q of the dipole peak low, so it makes more sense to use an H-frame. The H-frame gives you independent control over the width and depth of the "tunnels" that generate the resonance. This means I can choose whatever dipole peak frequency works best for the application, as well as get control over the Q of the line resonance and make is low enough that I can include it in the passband.
In contrast the Ripole makes the resonances as bad as possible. This means you need to cross over well below the resonance peak, and as a result you have to throw away the "higher SPL" region of the dipole response.
Hi Scott,
I am well aware of this, it has been repeated to exhaustion in this and other threads: 6dB/octave loss below the dipole peak. The dipole peak for this narrow construction is indeed quite high so the 6dB/octave shelving LP correction is applied up to 250Hz (so +18dB at ~30Hz). The eventual output is therefore not very loud indeed, but these speakers are meant for a relatively small room where I can't play loud. For what it is worth, I like them a lot anyway (and they were cheap and easy to make!) and looking for more bass output for my other speakers (therefore following this thread).
Best, Erik
I have found the multi subwoofer solution to room interaction issues to be interesting, but, with logic and cost applied, a silly resolution. Basically, it is smoothing.... where one of the subs is fat, another may be skinny. And so on. So many space eating boxes, and expense. The trick with speaker design is to take the playback space out of the equation.
Excite the room with music, with out exciting the room! I have found you can do that, with out breaking the bank, two ways, assuming a typical residential environment.
A really good medium size two or three way sealed (or highly engineered B/R) on stands 3' from the back wall with the woofer center 24" off the floor, or, a well designed OB - Cardioid. Either, if properly designed can take the room out of the equation.
Excite the room with music, with out exciting the room! I have found you can do that, with out breaking the bank, two ways, assuming a typical residential environment.
A really good medium size two or three way sealed (or highly engineered B/R) on stands 3' from the back wall with the woofer center 24" off the floor, or, a well designed OB - Cardioid. Either, if properly designed can take the room out of the equation.
I'd mostly agree.. BUT it's not silly for multiple listeners (at the same time and different positions in the listening room of course).
A good monopole very near the listener in a single-listener setup with the right amount of delay, can be superior.
Stereo dipole's very near the listener are about ideal when properly positioned (with delay) relative to that single listener. Under that condition you can actually achieve substantive lower freq. ILD for the listener. (..along with lower non-linear distortion and a greater tactile sensation ..for a dipole.)
Last edited:
Agreed. H-frame is also best compromise due to symmetry. My U-frames are 12" deep and are crossed over at 100Hz, so are very nearly dipoles?
ErikdeBest ob Treble considerations
Sorry not back sooner… The real work, tricks and or magic with full or near full range speakers is getting the midrange right. The design type, and / or mix of designs does not mater. By mix of designs, that is where, just for example, you may have a cardioid bass loading, OB mid-range loading and mono-pole tweeter. Or the reverse… so many ways around the barn.
After much experimentation, using what I hear, much more than what I measure, I prefer a properly implemented shallow cardioid loading for the midrange in a three way. It is tough to generalize; each potential midrange has its strengths and weakness. And, it is not an island, how it interacts with other drivers is very important.
For me to comment further, I would need to know a WHOLE lot more about the system overall, goals, costs and all the rest. Like all engineering exercises, you make the most of what you have and the interrelationship of all the parts.
Sorry not back sooner… The real work, tricks and or magic with full or near full range speakers is getting the midrange right. The design type, and / or mix of designs does not mater. By mix of designs, that is where, just for example, you may have a cardioid bass loading, OB mid-range loading and mono-pole tweeter. Or the reverse… so many ways around the barn.
After much experimentation, using what I hear, much more than what I measure, I prefer a properly implemented shallow cardioid loading for the midrange in a three way. It is tough to generalize; each potential midrange has its strengths and weakness. And, it is not an island, how it interacts with other drivers is very important.
For me to comment further, I would need to know a WHOLE lot more about the system overall, goals, costs and all the rest. Like all engineering exercises, you make the most of what you have and the interrelationship of all the parts.
Suitable bass divers ?
Well, since this is a relatively new thread, and I feel the best Open baffle design minds have visited here; I want to ask a serious question. I'm considering an open baffle design, but not full 20-20K, so sorry that this post is a bit off-topic. But rather, I need from 60-500Hz. 500 and above will also be open, but handled with different drivers.
Do you guys prefer pro style drivers ?
Would I be better of with 2x12's (vertical), or, 1x15" ?
I already have this driver, but upon examining the response curve, I see that it's already down 10 db @ 60Hz, relative to 200Hz. Looks like measurements were done in an IEC, installed in a closed off room (??)
Should I apply 10db E.Q to this ?
https://www.parts-express.com/pedocs/specs/290-417--eminence-delta-15lfa-specifications.pdf
Well, since this is a relatively new thread, and I feel the best Open baffle design minds have visited here; I want to ask a serious question. I'm considering an open baffle design, but not full 20-20K, so sorry that this post is a bit off-topic. But rather, I need from 60-500Hz. 500 and above will also be open, but handled with different drivers.
Do you guys prefer pro style drivers ?
Would I be better of with 2x12's (vertical), or, 1x15" ?
I already have this driver, but upon examining the response curve, I see that it's already down 10 db @ 60Hz, relative to 200Hz. Looks like measurements were done in an IEC, installed in a closed off room (??)
Should I apply 10db E.Q to this ?
https://www.parts-express.com/pedocs/specs/290-417--eminence-delta-15lfa-specifications.pdf
I do not use the ripole sub term restrictively by the original design; ie the cavity size can be adjusted
and i suggest to cross the ripole nearfield sub around 60hz and use a steep dsp based low pass filter to deal with cavity resonances
The low end response of the driver(s) can be roughly modeled as the combination of the driver's free-air response and the baffle response.
To model the free-air response of the driver, open any box modeler, enter the Thiele Small parameters for the driver (Fs, Qm, Qe, Qts, etc) select a closed box, and then enter a value for the box size that is 100+ times larger (there is no number "too large") than the driver's Vas. If you are not sure just enter 100000 liters or whatever. The box modeler will spit out the response curve, which is as if the driver is in a box of infinite volume with an infinitely large baffle. You should see the response drooping below 100-150 Hz and may be down (as you have seen from the MFG curve) 10dB by 50Hz.
To model the OB, download "the Edge!" modeling program here:
Tolvan Data
That page has some screen shots, but in essence you choose:
- 4 cornered baffle
- 1 speaker
- select Speaker shape = round
- 8-12 speaker source density
- for "Size" enter the equivalent radius of the driver's Sd (cone area) in mm or directly measure across the cone including 1/3rd of the surround on each side
- size the baffle by entering the dims at the top left, in mm, and clicking "Apply"
- click and hold on the driver and then drag it into place on the baffle
- make sure "Open Baffle" is checked
- set the Mic Distance to 3m
- move the mic (red dot) to somewhere close to the driver, e.g. just above the top edge at its center line
You will see a response window that will give you the baffle response, assuming the driver has infinite bandwidth (no low end roll off at all, all the way down to DC). You take that curve and add to it the box model curve, e.g. subtract 10dB at 50Hz, etc. This gives you a rough guesstimate of the open baffle response of that driver in that baffle.
One thing that is useful is to change:
baffle size
location of driver
And to look at:
both on and off axis angles.
What I do is this:
I start with the "mic" at 3m and next to the edge of the driver. Then I click the icon that looks like a blue mop head (I have no idea what it is supposed to be) in the response window to "freeze" the curve (it turns red I think).
Next I move the mic position over horizontally about 0.75m and repeat. I keep moving the mic position and freezing the curves until I get to 3m to the side. This should be 45 deg off axis I think. You want the curves to stay parallel and not cross or have large dips. See how baffle size and driver position influence this. To clear all curves, click the "X" icon in the response window and start again.
Also, to do that you will need to resize the grid where the baffle and driver are drawn so that you see at least 3m on the x-axis (since you have to move the mic over there). To do this, move the cursor over the x axis 3/4 or more towards the right. You should see an arrow like this |--> appear. Drag it back to the left and this will "zoom out" the x-axis. If the cursor is places closer to the middle of the x-axis a double headed arrow will be shown, like this: <---->. That moves the x-axis left and right, e.g. to move the baffle back and forth. I usually keep the baffle at left around x=0 and have the x-axis spanning from 0 to about 3.2m. Similar rescaling is possible on the y-axis if you want to look at what is happening in that direction.
Once you see how large the losses can be, you can guesstimate how much power would be necessary to "lift" the response up. For every 3dB of lift double the power input. If your box modeler shows cone excursion, you can enter that power and look at where excursion exceeds Xmax.
There are more sophisticated ways to model all of this but it is much more complicated than this. This approach is good enough for roughing out the plan, etc. Then I just build a test baffle or 'frame and measure it.
Last edited:
I recall that the Ripole's defining characteristic was the relatively low ratio of the cross sectional areas of the ripole openings to the driver Sd (cone area). I think in a Ripole this is 0.5 or less. If I recall correctly, it the ratio had to be this small to load the driver sufficiently (via resistance to the air moving in and out of the slot) and the result was lower Fs (Qts should also go up I think).
Once you make the opening cross section larger it is no longer like a Ripole. Of course the opening can be as large as you want it and the behavior will change along a continuum.
There is an entire thread on Ripoles in this forum, for anyone who is interested:
Drivers / parameters for ripole subs
Post #3 in that thread is a list of relevant links. Lots to read. Never mind, most links are broken.
Last edited:
Low Mms 15 inch as open baffle
I like the low Mms but dont like the small voice coil and low Xmax...I reckon its a 100Hz to 500 Hz OB driver....Asking it to go down to 50Hz would require a four fold increase (or is it 8 fold increase per lowering of octave?) in driver travel for the same SPL.... So 60Hz will strain the driver and you will lose the OB magic in the 100Hz to 500Hz band.
As a reference I have used the Precision Devices PD 158 http://www.precision-devices.com/file-downloads/PD158-DATASHEET-310114.pdf
in open baffle (as well as ELF sealed) and its beautiful covering 80Hz to 500Hz in both.
I like the low Mms but dont like the small voice coil and low Xmax...I reckon its a 100Hz to 500 Hz OB driver....Asking it to go down to 50Hz would require a four fold increase (or is it 8 fold increase per lowering of octave?) in driver travel for the same SPL.... So 60Hz will strain the driver and you will lose the OB magic in the 100Hz to 500Hz band.
As a reference I have used the Precision Devices PD 158 http://www.precision-devices.com/file-downloads/PD158-DATASHEET-310114.pdf
in open baffle (as well as ELF sealed) and its beautiful covering 80Hz to 500Hz in both.
Low Qts drivers good for ELF and open baffle
Discopete PM'd me asking for some driver suggestions in another thread, I have attached what I sent him as most of the drivers also work really well in open baffle.
Obviously DSP crossovers / Eq are required in both loading methods.
My comments are based on years of trial and error and use of good but not professional level software & mics.
As I run movies and games through the system I prefer 2, 3 or 4 sealed box subs below 45Hz or so.... 12, 15 or 18 inch subs depending on room and SPL requirements.
Cheers
Alex.
Discopete PM'd me asking for some driver suggestions in another thread, I have attached what I sent him as most of the drivers also work really well in open baffle.
Obviously DSP crossovers / Eq are required in both loading methods.
My comments are based on years of trial and error and use of good but not professional level software & mics.
As I run movies and games through the system I prefer 2, 3 or 4 sealed box subs below 45Hz or so.... 12, 15 or 18 inch subs depending on room and SPL requirements.
Cheers
Alex.
I sort of browsed this thread and since part is concerned with low frequency reproduction I though these quotes for a paper by Blackman might be of interest. They are with regard to low frequency response of difference sources below the first mode.
"Below the first mode the sound pressure, especially for the monopole and the cardioid, first increases with the decreasing frequency, and then reaches a flat level. The increase can be explained by the fact that below the first mode the sound pressure gradually starts to obey the adiabatic equation of the state, which predicts the sound pressure to be proportional to the relative change in the enclosed [room] volume...The flattening of the response at the very lowest frequency results from the room losses; in realistic rooms the losses at very low frequencies are not frequency-independent. [This is with regard to the room transfer function which would be superimposed on the sources free field response.]
“The dipole speaker does not produce a net change in the room volume, but due to the direct sound and the first reflections from the boundaries, which are not totally cancelled by later reflections in a lossy room, some sound pressure is still produced. [ I.E. dipoles don't produce much bass below the first room mode.]
“Although the change in room absorption appears to have about equal effect on all the speaker types ... the cardioid appears more immune to even extreme changes in room absorption around the lowest modes and just below them. The dipole, as is expected, has significantly less low-frequency output than the other types.
“The results from source placement simulation indicate that a cardioid speaker is more immune to source position effects in the sparsely modal region than omnidirectional [monopole] or bidirectional [dipole] speakers, confirming the hypothesis of more even modal coupling.
“As a summary of the results presented above for the room-speaker interaction it can be stated that cardioid source has more immunity against changes in source placement or room absorption in sparsely modal range. Below the lowest mode the cardioid speaker does not have any advantage over the monopole source, but both exhibit both higher output and less source position dependence than the dipole speaker. These results indicate that creating a loudspeaker that has a unidirectional polar pattern [cardioid] in the sparsely modal region and omnidirectional below the lowest mode represents a good compromise between low frequency output capability and
avoiding room coloration effects. ”
From:
1) Low-frequency polar pattern control for improved in-room response. Juha Backman, Presented at the 115th Convention 2003
October 10–13
"Below the first mode the sound pressure, especially for the monopole and the cardioid, first increases with the decreasing frequency, and then reaches a flat level. The increase can be explained by the fact that below the first mode the sound pressure gradually starts to obey the adiabatic equation of the state, which predicts the sound pressure to be proportional to the relative change in the enclosed [room] volume...The flattening of the response at the very lowest frequency results from the room losses; in realistic rooms the losses at very low frequencies are not frequency-independent. [This is with regard to the room transfer function which would be superimposed on the sources free field response.]
“The dipole speaker does not produce a net change in the room volume, but due to the direct sound and the first reflections from the boundaries, which are not totally cancelled by later reflections in a lossy room, some sound pressure is still produced. [ I.E. dipoles don't produce much bass below the first room mode.]
“Although the change in room absorption appears to have about equal effect on all the speaker types ... the cardioid appears more immune to even extreme changes in room absorption around the lowest modes and just below them. The dipole, as is expected, has significantly less low-frequency output than the other types.
“The results from source placement simulation indicate that a cardioid speaker is more immune to source position effects in the sparsely modal region than omnidirectional [monopole] or bidirectional [dipole] speakers, confirming the hypothesis of more even modal coupling.
“As a summary of the results presented above for the room-speaker interaction it can be stated that cardioid source has more immunity against changes in source placement or room absorption in sparsely modal range. Below the lowest mode the cardioid speaker does not have any advantage over the monopole source, but both exhibit both higher output and less source position dependence than the dipole speaker. These results indicate that creating a loudspeaker that has a unidirectional polar pattern [cardioid] in the sparsely modal region and omnidirectional below the lowest mode represents a good compromise between low frequency output capability and
avoiding room coloration effects. ”
From:
1) Low-frequency polar pattern control for improved in-room response. Juha Backman, Presented at the 115th Convention 2003
October 10–13
Did you get this question about resonance in U-baffles sorted? I didnt see an answer so did some extrapolations
I guess from first principles
applied to a U-baffle or H-frame there will be two resonances.
1. Quarter wavelength of the depth (in the direction of the driver movement from the driver to the opening of the baffle)
2. Half wavelength of the smallest dimension of height or width ( 90 degrees to the direction of the driver movement)
From that it follows if using a U or H frame there will be a limit on upper extension due to resonance. This limit will occur earlier at lower frequencies for larger drivers and/or increasing depth to allow better low frequency extension.
Depth dimension should be half that of the largest traverse plane (width or height).
H-baffles should have better low frequency output with the same upper resonance free extension then a U-baffle by using the same depth dimension fore and aft.
A U-baffle made like a floor stander box speaker with a base and lid and the back off could easily get resonance way too early. This could be fixed with a horizontal shelf installed above the woofer.
A problem arises trying to use a U/H-frame for bass. To get good low extension output a a large dipole separation is indicted leading to lower resonance frequency. However woofers can output high second harmonics. These second harmonics are likely to extend into the U/H-frame resonant zone and be further enhanced with harmonic distortion swamping the fundamentals.
This all suggests that demanding a dipole be flat to 20Hz is going against the natural order. It might be possible with enough money, power and time but it might not be elegant.
Heres one solution sans baffles to lose the resonance issues but its not good for 20Hz despite 6 x 12" AE woofers per side .
An externally hosted image should be here but it was not working when we last tested it.
Im just made all of that up and dont expect sound waves to comply. Please correct logic failures
^^ Cardioid bass is obviously best in-room. Sadly cardioid has almost as low efficiency as dipole and must also be very large to go low. At least 2x15" with low Fs and huge amount of Xmax AND eq with dsp. This kind of "unit" can't go very high either, reasonable max around 400Hz to get nice xo to midrange.
Here we have some good examples by kimmosto, with in-room measurements.
My understanding and experience with dipoles suggest max 3 octaves passband per unit/way. This means a 4-way, because each way must have some linear overlap too, for example like this: (alternative)
- sub 20-120Hz 2½oct (20-200Hz)
- bass 120-500Hz 2 oct (200-800Hz)
- mid 500- 2000Hz 2 oct (800-3200Hz)
- tweeter 2000- 20 000Hz 3½ oct (dipole physically possible only up to 6-7kHz with transducers available)
And because of physical limitations mentioned earlier in this thread, mid and tweeter must be narrow planar units.
Here we have some good examples by kimmosto, with in-room measurements.
My understanding and experience with dipoles suggest max 3 octaves passband per unit/way. This means a 4-way, because each way must have some linear overlap too, for example like this: (alternative)
- sub 20-120Hz 2½oct (20-200Hz)
- bass 120-500Hz 2 oct (200-800Hz)
- mid 500- 2000Hz 2 oct (800-3200Hz)
- tweeter 2000- 20 000Hz 3½ oct (dipole physically possible only up to 6-7kHz with transducers available)
And because of physical limitations mentioned earlier in this thread, mid and tweeter must be narrow planar units.
Last edited: