I was simulating closer what real conditions would be, but I see your point. Thanks for taking the time to provide a thorough answer.
You mean low passing the 10", correct? I was thinking lower, like 300Hz, but your comment is making me think it would maybe be too demanding on the midrange. 600Hz would reduce by 8x the excursion demanded at 300Hz so no small difference. Eventually I will check how music sounds in that region played by the 8" vs. the twin 10".
What's your view on the value of making these measurements indoors? With speakers standing on the floor? How should the be made to make them as valuable as possible?
Yes concidering where to cross and what order is complicated, especially with dipoles. With dsp it is very easy to try different settings and minidsp can save four on presets. Then listening tests and distortion wiht high spl!
Measuring indoors below say 800Hz is conmplcated always. Speaker diagonally at least 2m from sidewalls and on a stand measured at 1m gives lowest reflections and is ok for mid up and setting crossovers and timing. But even then below 400Hz or so is a mess.
Ground plane with long gating 60-100ms will help with bass, combined with nearfield longer gate. Nearfield misses baffle effect and most importantly dipole summation, so that must be kept in mind. Guide to ground plane outdoors but indoor at 1-2m is better than nothing too.
Listening position measurement of both speakers individually with 500ms gating is needed to find gain for each driver/way. Many people eq the lowest room mode too, but not me.
Measuring indoors below say 800Hz is conmplcated always. Speaker diagonally at least 2m from sidewalls and on a stand measured at 1m gives lowest reflections and is ok for mid up and setting crossovers and timing. But even then below 400Hz or so is a mess.
Ground plane with long gating 60-100ms will help with bass, combined with nearfield longer gate. Nearfield misses baffle effect and most importantly dipole summation, so that must be kept in mind. Guide to ground plane outdoors but indoor at 1-2m is better than nothing too.
Listening position measurement of both speakers individually with 500ms gating is needed to find gain for each driver/way. Many people eq the lowest room mode too, but not me.
Thanks for the links. Will look into them tonight.
Trying different crossovers will be rather easy in my set up: I use Acourate to derive digital xo and room correction, so can test different xo points/types/slopes and try them until my heart's contempt
. Acourate will also take care of matching gains for each driver, and those measurements are taken at listening position indeed.
With boxed speakers one would usually linearize driver response with Acourate by taking very nearfiled measurements, but with OB it doesn't make sense to do that, as you pointed. Then one would run the final measurement from the listening position at ear-height. Acourate reduces the response down to a target curve defined by me. I wonder how applying 15dB reduction (or more) will sound. I guess the way to know is to try it.
All of the above will happen when I have the system reconfigured. For now I'm testing drivers individually to understand if this has any chance of succeeding. Seems like 2x10" up to about 600Hz and 8" baffleless up to 1500Hz could work. Correct?
I need to look into incorporating the TPL-150H. StigErik concluded TPL-150 were too large for proper dipole operation and moved to a smaller Mundorf and eventually RD75. I'm sticking to TPL for now. I'm leaning to keeping the waveguide on. With factory back chamber, enlarged closed chamber, open back? Opinions very welcomed!
Trying different crossovers will be rather easy in my set up: I use Acourate to derive digital xo and room correction, so can test different xo points/types/slopes and try them until my heart's contempt
. Acourate will also take care of matching gains for each driver, and those measurements are taken at listening position indeed.With boxed speakers one would usually linearize driver response with Acourate by taking very nearfiled measurements, but with OB it doesn't make sense to do that, as you pointed. Then one would run the final measurement from the listening position at ear-height. Acourate reduces the response down to a target curve defined by me. I wonder how applying 15dB reduction (or more) will sound. I guess the way to know is to try it.
All of the above will happen when I have the system reconfigured. For now I'm testing drivers individually to understand if this has any chance of succeeding. Seems like 2x10" up to about 600Hz and 8" baffleless up to 1500Hz could work. Correct?
I need to look into incorporating the TPL-150H. StigErik concluded TPL-150 were too large for proper dipole operation and moved to a smaller Mundorf and eventually RD75. I'm sticking to TPL for now. I'm leaning to keeping the waveguide on. With factory back chamber, enlarged closed chamber, open back? Opinions very welcomed!
A practical question: if the front NF is measured close to the apex of the dustcap (assuming no phase plugs etc), then at which position do you measure the rear NF? What if it is a U-frame with full-depth stuffing/damping?
Made some progress. I had an issue with my soundcard which took a fair amount of time to figure out it was a hardware issue only affecting two channels, so I was able to circumvent the issue and measure using other channels.
Anyway, here's where I stand: built one speaker, playing mono.
Two 12" sealed subs play up to 70Hz
Two 10" OB from 70 to 500Hz powered by the Hypex UcD amp (the low gain in these amps forced me to re-route soundcard channels as not all of them have the same gain)
Swinging 8" bafleless from 500 to 1500Hz powered by one channel in MC275
TPL-150H with back cover as sold by Beyma from 1500Hz up, powered by MC275.
XO: NT 2nd-order. Digital XO and filters generated in Acourate.
Sounding good. Many things to try out, though. I turned the speaker around 90 degrees and opened the dividing panels so the room is now 10m long and speakers are 3m away from the [now] front wall and 7m from the back wall.
Fun!
Anyway, here's where I stand: built one speaker, playing mono.
Two 12" sealed subs play up to 70Hz
Two 10" OB from 70 to 500Hz powered by the Hypex UcD amp (the low gain in these amps forced me to re-route soundcard channels as not all of them have the same gain)
Swinging 8" bafleless from 500 to 1500Hz powered by one channel in MC275
TPL-150H with back cover as sold by Beyma from 1500Hz up, powered by MC275.
XO: NT 2nd-order. Digital XO and filters generated in Acourate.
Sounding good. Many things to try out, though. I turned the speaker around 90 degrees and opened the dividing panels so the room is now 10m long and speakers are 3m away from the [now] front wall and 7m from the back wall.
Fun!
The Beyma TPL has no backside radiation at all and high directivity. This is a huge difference to having dipole pattern 70 1500Hz in a room!
In a room we hear also reflected sound as part of total sound pressure/room response. An additional backside tweeter will help a lot to get sound balance right and to give similar "3D presentation" of stereophonic imaging.
Here my horn tweeter version of AINOgradients with and without backside tweeter - at listening spot/on-axis. Crossed around 3500Hz and it was very easy to hear the difference! On-axis response was not equal in these measurements, but I had numerous settings tested and the backside tweeter stayed until I purchased B&G Neo3 planar dipole tweeters.
In a room we hear also reflected sound as part of total sound pressure/room response. An additional backside tweeter will help a lot to get sound balance right and to give similar "3D presentation" of stereophonic imaging.
Here my horn tweeter version of AINOgradients with and without backside tweeter - at listening spot/on-axis. Crossed around 3500Hz and it was very easy to hear the difference! On-axis response was not equal in these measurements, but I had numerous settings tested and the backside tweeter stayed until I purchased B&G Neo3 planar dipole tweeters.
Attachments
Last edited:
It might not be so bad. The TPL has plenty of directivity with the back on.
I'm not convinced that the cross to the B&C won't work because with the size of the magnet what is going to be happening at 1k5? Theoretically the DI would be out by 6dB but in reality I'd want to see measurements. What about felt behind the mid to gradually blend it toward forward radiation only at its top end?
I'm not convinced that the cross to the B&C won't work because with the size of the magnet what is going to be happening at 1k5? Theoretically the DI would be out by 6dB but in reality I'd want to see measurements. What about felt behind the mid to gradually blend it toward forward radiation only at its top end?
One of my next steps is to take the back cover off in the TPL and use it as dipole. Long time ago StigErik was using the TPL in this fashion, although without the horn, and spoke highly about it. Eventually he concluded the TPL was too large and moved on to B&G Neo 3 dipoles too.
I'm not sure what to take away from the plots you attached. The top graph is with back tweeter and longer decay?
Which measurements would you like to see? I'm limited to in-room measurements (no outdoors option).
If I were in your position, I would start with 180 degree polars, probably just horizontal to begin with to see if it can be done. Both tweeter and mid. In room is ok if you can get clean down into the mid hundreds or so.
I'd be looking to see where and how the B&C struggles with the highs from the back. Whether it would be possible to use this, and to blend somehow to meet the tweeter.
Finally I guess, if there is a remaining imbalance, you could tweak by using an absorber on the wall behind the speaker which is thin enough only to catch some of the upper midrange.
I'd be looking to see where and how the B&C struggles with the highs from the back. Whether it would be possible to use this, and to blend somehow to meet the tweeter.
Finally I guess, if there is a remaining imbalance, you could tweak by using an absorber on the wall behind the speaker which is thin enough only to catch some of the upper midrange.
Sorry - I didn't see this question until now.
I had my H-frames in mind when I wrote this, and in that case you are measuring at the plane of the opening at both front and rear. It's pretty easy to do.
For a U-frame, you measure in front of the cone/dustcap (front) and at the plane of the rear opening. Also easy.
For a driver in a planar baffle, the front is measured close to the dustcap, and the rear you can measure next to the basket, behind the cone. It's not optimal, but the magnet is in the way so you have to choose another spot nearby that is still close to the rear of the cone.
These measurements only need to be accurate up to a few hundred Hertz, so its not as important to place the mic extremely close to the cone.
Lewinski, yes upper was with backside dome tweeter and Fountek Neo 3.5H on the front (version NeoB). More measurements from that time in my AINOgradient thread.
Yes the B&C midrange obviously doesn't have perfect dipole radiation around 1500hz, but in a rom total radiated power front/back seems to be more important than the perfectness of backside radiation pattern.
Linkwitz and Kreskowsky have as well used backside tweeters in many versions. Kreskowsky's last Note II RS doesn't have it but tweeter is crossed around 6kHz. Linkwitz LX521 uses pretty high xo too.
Listening tests are important too! I want the backside radiation to get reflected from the front wall and then spread 'n bounce around the room! This gives the pleasant sound for me. If one is looking for sharp stereo image and separation of every click and whine, dipoles are a bad choice.
Yes the B&C midrange obviously doesn't have perfect dipole radiation around 1500hz, but in a rom total radiated power front/back seems to be more important than the perfectness of backside radiation pattern.
Linkwitz and Kreskowsky have as well used backside tweeters in many versions. Kreskowsky's last Note II RS doesn't have it but tweeter is crossed around 6kHz. Linkwitz LX521 uses pretty high xo too.
Listening tests are important too! I want the backside radiation to get reflected from the front wall and then spread 'n bounce around the room! This gives the pleasant sound for me. If one is looking for sharp stereo image and separation of every click and whine, dipoles are a bad choice.
It's important to keep in mind that just plopping on a rear tweeter is not a good recipe for "dipole" sound.
For example, SL in his LX521 wanted to have a dipole tweeter, and he liked the response of domes (for several good reasons) so he attempted to create a dipole using the front+rear combo. Except it didn't work well. During development he had to keep raising and raising the crossover point for the because at lower frequencies the dipole pattern was poor. Finally, when he moved the point all the way up to 7kHz he could get acceptable dipole radiation. But this was only because the dome tweeters were beaming at that point, and they no longer "saw" each other and did not interact.
Quoting SL from here:
LX521 Description
The reason that SL could not get the two tweeters to act as a dipole at lower frequencies is because
- the baffle they are mounted in is too wide
- the domes are offset vertically on the front and back sides of the baffle
- a 1" dome tweeter is an omnidirectional source up to almost 5kHz
What is needed is either a radiator that is naturally a dipole (e.g. Neo3PDR), or some combination of two domes that does not suffer from these problems. The only way I have been able to solve this issue using domes is when I used two dome tweeters in waveguides, positioned very close back-to-back, and used without a baffle. The particular driver I used has a very small motor, and this also helped to keep them more closely aligned vertically. As a result I could almost position the motors above one another, which also improves the resulting dipole pattern. The waveguide increases the directivity down to 3kHz and below that the front and rear output seems to add in phase, due to some luck with the pathlength formed by the waveguides.
It's my opinion that SL could have omitted the rear dome in the LX521 and it would have been about the same, because of the high crossover point. Anecdotal evidence from Kreskovsky and others points to about a 5kHz crossover point (but not lower) above which a dipole tweeter is no longer enhancing the sound of the speaker. Up to that point using dipole radiation is helpful in terms of soundstage, etc. It might have to do with the fact that many room materials (like drywall) become increasingly absorbent to sound at high audio frequencies, so the rear radiation does not return to the listener as strongly as at lower frequencies.
So, related to your project LewinkiH01, if you use the TPL tweeter crossed over at 1.5kHz you will eliminating the rear (dipole) radiation above that and below the 5kHz or so cutoff, and this will likely worsen the "ambience" of the dipole system. Also, crossing over a dipole to a monopole is a bit trickier than when crossing over two drivers having the same type of radiation. Some kind of cardiod pattern will be formed, but since your tweeter is horn loaded exactly what will happen is not clear. It's a bit of a gamble, and it may work or may not. Once you have the system playing, I would listen carefully.
Last edited:
Thanks for the great answer! I learnt quite a bit.
Yesterday I went ahead and removed the back chamber in the TPL and assembled a quick foamboard box around the back with an open back, yet I glued a plastic fencing piece to prevent large objects to get close to the magnet/membrane.
So now I have a waveguide on one side of the AMT and open on the other. Still not symmetric dipole.
Been doing some listening and I like what I hear. I have not remeasured and adjusted the filters.
It's fun to hear treble from the back of the speaker. Has taken me by surprise a couple times while doing something else on that part of the room.
More listening needed
Midbass
I'm lacking a bit of sense of powerful bass in the midbass. I'm thinking the twin 10" might be shy on the total surface they offer. Which might be why I usually see twin 12" or twin 15" in many OB designs (of course some of those designs aren't meant to run with subs).
Since I will continue to run my sealed subs at the bottom, I was wondering if a single 18" per side to run from about 50Hz (vs the twin 10" starting at 70Hz) would do better.
The 10G40 have Sd of 380cm2, so 760cm2 per speaker. One Eminence Alpha 15A, often seen in OB has 856cm2, and a Beyma 18LX60 has 1380cm2. So 1x18" has about twice the Sd of my 2x10".
On Edge I simulated 2x10" on a 50x700mm baffle (red line) vs 1x18" on same baffle (green) and basically see the same response, with the dipole peak around 350Hz:
The 10G40 have Qts of 0.3, which is on the low side for OB from what I see. Eminence 15A has 1.26 and AE Dipole 15 has 0.94 and Dipole 18 has 0.81. AE drivers aren't available here and logistics would be too complicated, so looking around local options I see these:
Are these good candidates to use in OB for 50-300Hz? Or should I go for 15"?
Since I will continue to run my sealed subs at the bottom, I was wondering if a single 18" per side to run from about 50Hz (vs the twin 10" starting at 70Hz) would do better.
The 10G40 have Sd of 380cm2, so 760cm2 per speaker. One Eminence Alpha 15A, often seen in OB has 856cm2, and a Beyma 18LX60 has 1380cm2. So 1x18" has about twice the Sd of my 2x10".
On Edge I simulated 2x10" on a 50x700mm baffle (red line) vs 1x18" on same baffle (green) and basically see the same response, with the dipole peak around 350Hz:
The 10G40 have Qts of 0.3, which is on the low side for OB from what I see. Eminence 15A has 1.26 and AE Dipole 15 has 0.94 and Dipole 18 has 0.81. AE drivers aren't available here and logistics would be too complicated, so looking around local options I see these:
- Beyma 18LX60: Qts: 0.48, Xmax: 8mm, fs: 35Hz, Sd: 1320cm2, Sensitivity@50Hz: about 89-90dB, impedance@50Hz: about 16 ohm
- Faital 18HP1010: Qts: 0.40, Xmax: 9.75mm, fs: 35Hz, Sd: 1207cm2, Sensitivity@50Hz: about 93dB, impedance@50Hz: about 27 ohm
Are these good candidates to use in OB for 50-300Hz? Or should I go for 15"?
Last edited:
Measuring outside
I've been listening mono to the speaker and I'm really liking what I hear. TPL tweeter is open back (per post #57), the 8" nude on the swing, and the twin 10" on the panel as above pictures show. Really liking the tweeter open back, and the bass definition. Was wanting more sense of impact from low midbass.
Went ahead and purchased two 18" for midbass (one per side), increasing midbass Sd by 60% vs the twin 10". I'm hoping to be allowed to stay in the house . Faital 18HP1010. I did a fair amount of reading, including Charlie's pursuit of 20-20k dipole thread (great learning!), and think these Faitals should work well and have a nice open back to make the rear wave as similar to front wave as possible.
The plan is to swing them, much like the 8" in above posts. And also run it nude as starting point. Might also experiment with light, non resonating baffle materials later.
Tomorrow I plan to take a trip to where I can measure outside. Would like to get input to insure my measurements turn out useful. The goal is to have measurements to help make informed decisions about xo points, and suitability of these drivers. In particular, I'm guessing the big magnet behind the 8" must be an issue, and wondering how the TPL with a horn + open back + extended frame behaves.
Although I'm stating the plan below, it's really a bunch of questions for confirmation/rectification:
Guidance from the experienced will be highly appreciated!!
I've been listening mono to the speaker and I'm really liking what I hear. TPL tweeter is open back (per post #57), the 8" nude on the swing, and the twin 10" on the panel as above pictures show. Really liking the tweeter open back, and the bass definition. Was wanting more sense of impact from low midbass.
Went ahead and purchased two 18" for midbass (one per side), increasing midbass Sd by 60% vs the twin 10". I'm hoping to be allowed to stay in the house
. Faital 18HP1010. I did a fair amount of reading, including Charlie's pursuit of 20-20k dipole thread (great learning!), and think these Faitals should work well and have a nice open back to make the rear wave as similar to front wave as possible.The plan is to swing them, much like the 8" in above posts. And also run it nude as starting point. Might also experiment with light, non resonating baffle materials later.
Tomorrow I plan to take a trip to where I can measure outside. Would like to get input to insure my measurements turn out useful. The goal is to have measurements to help make informed decisions about xo points, and suitability of these drivers. In particular, I'm guessing the big magnet behind the 8" must be an issue, and wondering how the TPL with a horn + open back + extended frame behaves.
Although I'm stating the plan below, it's really a bunch of questions for confirmation/rectification:
- Measure each driver separately, from 0 to 180 degrees horizontally in 15 degree steps far away from any boundaries. Mic to stay in place and rotate speaker on a rotation table.
- Rotation table will be on top of a regular table, so about 70cm above ground. That will place the tweeter at 1.7m above ground which seems good, but the midbass center will be at about 1m above ground. Note the midbass in the room will be playing on the floor (basket in vertical position almost touching the floor)
- No measurements of vertical dispersions planned for now.
- Mic to be placed on vertical axis for each driver measurement, 1 meter away. Or 2m?
- What REW gating settings should I use? I'm used to measuring indoors and gating to avoid reflections, which I shouldn't have much tomorrow.
- FWIW, I will bring an absorption panel to place on the floor reflection point.
- Am I missing something important?
Guidance from the experienced will be highly appreciated!!