Just go back to pages 10-12. If you're still asking, you haven't read them.
Yes, but the pressure is the same if the cones work one back and one forward, in the same direction from a listener on the couch, then they are out of phase. Then the pressure would be the same between the cones, isobaric. Put a pressure gauge in there and it will read the same as it would sitting on the couch.
If the cones move one forward and the other forward, in opposite directions relative to a listener on the couch, then they are in phase because one driver is turned around. They open the cavity and close it, then it's not isobaric and the pressure goes up and down. This is what I think the design is, the picture is incorrect and it confused me.
Last edited:
I read pages 11-12 I thought I knew what was going on but then I saw this and it confused me, sorry. The arrows on the dipole drawing on the right are wrong.
That's a sealed bandpass enclosure. It is none of these: open baffle, H-frame, U-frame, or clam shell.
You are right, the arrows on the right picture are indeed wrong. The figure of 8 is drawn right, but all arrows should point in the same direction if one feels inclined to draw it this way.
Back to my questions: How does one align these with the midrange for time delay?
If the woofers are making in phase sound from the front and the back of the clam shell and cancel sound in the middle then how do you integrate an open baffle midrange that is making in phase out of the front and out of phase from the back? The clam shell is something a typical single woofer system doesn't do. Does a steep crossover filter handle the phase issues?
If the woofers are making in phase sound from the front and the back of the clam shell and cancel sound in the middle then how do you integrate an open baffle midrange that is making in phase out of the front and out of phase from the back? The clam shell is something a typical single woofer system doesn't do. Does a steep crossover filter handle the phase issues?
If woofers would be moving towards each other, it would not work.
Unless there would be big box and they were oposite, it would be bipole.
You are talking about bipole, but image shows dipole. You seem to confuse bipole with dipole. I do have web page explaining this. If interested, go to my blog, find open baffle sub page.
Unless there would be big box and they were oposite, it would be bipole.
You are talking about bipole, but image shows dipole. You seem to confuse bipole with dipole. I do have web page explaining this. If interested, go to my blog, find open baffle sub page.
Sorry for the confusion re diagram. It was just supposed to show how the open clamshell configuration has a narrower spread pattern compared to monopole. This might clarify (or confuse more?) 😕
Thats a good question. Technically, the midrange driver should be in phase with the driver closest to the listener. Cant do this by eyeballing because the midrange excursion are hard to see, and anyway, in my case, they are cut off @ 60hz. To further confuse the phasing, my sub amps are wired in reverse compared to the active speakers I use,. In addition you have the polarity reversed on the rear driver. Jeez.....has driven me nuts many times!
Luckily, I found a solution. I noticed when taking a full range frequency scan that sometimes* I got a notch at the crossover frequency. At first I thought this was some anomaly in the room. I later discovered if I reversed the woofer polarity, the notch disappeared and I got a smooth transition.
By sometimes I mean early on in the process of dialling in the subs (which has gone through many permutations over the years) I would test different wiring configurations, positions and even different drivers.....I would get the polarities mixed up. I would wonder why my bass would disappear, or sound weird. There are many ways the wiring could go wrong.....and I have done them all! Nowadays I label all the connections so if I get the urge to try something different, I can put it back together the right way.
The speaker on the right is dipole too, because it is open and rearside radiation spreads to same space. A box that closes the neg. polarity radiation wouldl make that speaker monopole
Please study these diagrams https://www.acousticsciences.com/newsletter/lets-talk-hifi-speakers-pt-1/
https://en.wikipedia.org/wiki/Dipole_speaker
https://musicanddesign.speakerdesign.net/u_frame.html
Clamshell and ripole are variations of the basic single open baffle dipole. It is very important how polarities are connected! Sound frequency/wavelength and baffle dimensions determine how well radiation pattern works. At axial response dipole null frequency pattern is omni (balloon). H and U frame are ways to extend baffle width with small outer dimensions, but at expense of cavity resonances.
Please study these diagrams https://www.acousticsciences.com/newsletter/lets-talk-hifi-speakers-pt-1/
https://en.wikipedia.org/wiki/Dipole_speaker
https://musicanddesign.speakerdesign.net/u_frame.html
Clamshell and ripole are variations of the basic single open baffle dipole. It is very important how polarities are connected! Sound frequency/wavelength and baffle dimensions determine how well radiation pattern works. At axial response dipole null frequency pattern is omni (balloon). H and U frame are ways to extend baffle width with small outer dimensions, but at expense of cavity resonances.
There is nothing special about the "clamshell" configuration, as long as there is some minimal separation in between the drivers of at least a few inches that might be related to some fraction of the driver diameter. When widely spaced, the two drivers act as two independent dipole sources. In the low frequency limit, they should have about the same output whether side by side or in-front and behind one-another. In the far-field limit each is a dipole source and the total SPL should be their sum.
There are some details that can make the summed SPL slightly different, which depend on the relative mic position and frequency. For example for the front-back clamshell configuration, you cannot place the microphone at a position that is equi-distant from each driver when "on axis" because one driver is behind the other and therefore farther away. So if for example the drivers are separated by 0.35m (14") and you measure at 1m "in front" of the front driver, the rear driver is farther away and its SPL will be lower by 1dB or more, which will reduce the sum by the same amount. At higher frequencies, the pathlength difference in the clamshell configuration causes a phase difference in the wavefronts reaching the mic and eventually this starts to be more than 90 degrees of phase and with increasing frequency there can be interference that causes a dip.
When the drivers are very close together (e.g. 1 inch separation) and in the clamshell setup, the source from the inner surfaces of the diaphragms is mostly cancelled. But because there are two motors pushing on the cones the SPL is +3 dB compared to a single driver (assuming either a parallel connection, or two identical amp channels powering each driver). When you open up the separation more and more, the SPL should approach +6dB WRT a single driver in the far-field. SL's analysis of the compound dipole "clamshell" is based on point sources does not include any effect of the physical driver itself. It is really only accurate in the limit of large driver spacing and observation distances that are very large compared to the driver spacing.
See my measurements earlier in this thread:
https://www.diyaudio.com/community/...eless-full-range-speakers.374043/post-6729463
At low frequencies there is about a 3dB difference between a small spacing and a large one. It is a little more complicated because I kept the mic position fixed WRT the center position between the two drivers while changing the separation distance between the drivers. This makes the near driver move closer and closer to the mic, which was not all that far away, and this will increase its SPL a bit for the measurements at the largest separation distances.
Last edited:
Ok. The original diagram I adapted was corrupted in translation. Here is the corrected version:
The single (non-clamshell) driver also has some baffle around the driver that probably makes the most difference.
k9vap, what makes you think that a clamshell would have more directivity than a sigle dipole driver?
Please study dipole and loudspeaker radiation theory more:
https://musicanddesign.speakerdesign.net/Dipoles_and_open_baffles.html
https://musicanddesign.speakerdesign.net/tech.html
https://www.linkwitzlab.com/models.htm
http://www.dipolplus.de/index.html
Please study dipole and loudspeaker radiation theory more:
https://musicanddesign.speakerdesign.net/Dipoles_and_open_baffles.html
https://musicanddesign.speakerdesign.net/tech.html
https://www.linkwitzlab.com/models.htm
http://www.dipolplus.de/index.html
Given the frequency range involved, how much different is it than the inherent imperfections in midrange drivers due to the driver face not being completely flat (or better yet slightly concave with respect to a spheres surface with a center at the listenting position which is likely very different than the concave shape of the driver cone). Some of that sound from the mid range driver is inherently coming from the near center of the cone that is furthest away from the listener, and some of it is coming from the edges of the cone which is relatively closer to the listening position, etc.
ie, there is probably just a level of error that we always inherently live with when it comes to (multiple) driver time delays.
Last edited:
A curved radiator acts like a planar when wavelegth is longer than diameter/2. Acoustic center Z (depth) is roughly at edges of a cone. Woofer-mid xo is quite low freq so depth separation is negligible. Mid-tweeter xo is more sensitive.
https://sound-au.com/articles/ac-offset.htm
https://www.klippel.de/fileadmin/_migrated/content_uploads/KLIPPEL_Sound_Radiation_Poster_01.pdf
https://sound-au.com/articles/ac-offset.htm
https://www.klippel.de/fileadmin/_migrated/content_uploads/KLIPPEL_Sound_Radiation_Poster_01.pdf
looked into more details here and the main thing i was not noticing is the distance between these drivers clamshell mounted. They indeed act as two dipole sources. i assumed zero separation distance, but actually from photos i see it is more then 10cm probably between acoustic planes. So in example if i simulate 20cm diameter driver and add 10cm additional path difference i get 5.3db SPL increase at 50hz 1 meter away.
This clamshell design is an interesting way of simulating H frame or other baffle to some extent. The same as two drivers at the ends of very short pipes. Interesting aproach. But if the gap gets bigger between clamshell drivers this virtual baffle should loose its magic. But still it is interesting if somebody needs a small baffle and don't want to do any physical baffle. Didn't try clamshells, but yes should work to some limited extent.
This clamshell design is an interesting way of simulating H frame or other baffle to some extent. The same as two drivers at the ends of very short pipes. Interesting aproach. But if the gap gets bigger between clamshell drivers this virtual baffle should loose its magic. But still it is interesting if somebody needs a small baffle and don't want to do any physical baffle. Didn't try clamshells, but yes should work to some limited extent.