If the Altec/GPA 290/390 is this good, that certainly makes it promising.
JohnK: the discussion of the OB is not about EQ, but about controlling IM distortion below the baffle peak. Yes, there are some audiophile drivers with low distortion, but going the prosound + large area route keeps distortion down and also retains system headroom as well. I know what too-small overboosted audiophile drivers sound like - bad, with a very compressed sound and obviously limited headroom.
There were several systems like this at the RMAF. They sounded thin, compressed, and had very poor dynamics. Spacious, yes, of course. But the dynamics were no better than electrostats, but without the see-through transparency, either. I can't see any way horn dynamics can be mimicked by small drivers with lots of EQ and amplifier power.
JohnK: the discussion of the OB is not about EQ, but about controlling IM distortion below the baffle peak. Yes, there are some audiophile drivers with low distortion, but going the prosound + large area route keeps distortion down and also retains system headroom as well. I know what too-small overboosted audiophile drivers sound like - bad, with a very compressed sound and obviously limited headroom.
There were several systems like this at the RMAF. They sounded thin, compressed, and had very poor dynamics. Spacious, yes, of course. But the dynamics were no better than electrostats, but without the see-through transparency, either. I can't see any way horn dynamics can be mimicked by small drivers with lots of EQ and amplifier power.
Attachments
I expect to build several mid and HF assemblies, measure them, and listen at some length. The bass module is going to be heavy enough, with 3 or 4 15-inch drivers, that it will have a shelf on the top so the Mid and HF modules can be skated back and forth and independently adjusted, replaced, and fine-tuned.
No idea which mid configuration will sound better. The double-RAAL will be compared to a good but not insane compression-driver + horn/waveguide. No Gotos, Ales, or TAD, in other words. You gotta draw the line somewhere, and multi-thousand-dollar raw drivers fall in that category.
The overall system, as mentioned in the "Better Crossovers" thread, will be bi-amped with a pro EQ/crossover used for the rough adjustments, and either a prosound or good-quality PP pentode tube amp for the bass module. The Karna amplifier will handle everything above 300~600 Hz, depending on the crossover frequency for the mid/high module.
No idea which mid configuration will sound better. The double-RAAL will be compared to a good but not insane compression-driver + horn/waveguide. No Gotos, Ales, or TAD, in other words. You gotta draw the line somewhere, and multi-thousand-dollar raw drivers fall in that category.
The overall system, as mentioned in the "Better Crossovers" thread, will be bi-amped with a pro EQ/crossover used for the rough adjustments, and either a prosound or good-quality PP pentode tube amp for the bass module. The Karna amplifier will handle everything above 300~600 Hz, depending on the crossover frequency for the mid/high module.
Lynn Olson said:
I'm glad to see you heading this direction. (In my opinion) you will end up with an incredible sounding system going this way. You seem to be skipping to where I have led myself to after approx 8 years of building high efficiency systems.
To perfect the multi driver bass system will get your project much closer to where you'll want to dig in deeper. Even with my cheap 10's it's sooooooo nice and refreshing to hear all the low notes free from the box and compression or horn colorations.
I didn't realize Great Plains is bringing back the 290. I think you'll have a hard side deciding between the raw immediacy of the phenolic compression driver compared to the wonderful tone and deep color saturation of a big motor low mass cone. -- I'm to the point where I crave both - But-----
I really think you can have your cake and eat it too with the *right* cone mid. The 18 Sound 6" may be better than the B&C and there may be others. What's nice about the cones is they won't break the bank experimenting with them.
If you want, I'll share my cone drivers if you share yours......
Regarding:
<<I think you'll have a hard side deciding between the raw immediacy of the phenolic compression driver compared to the wonderful tone and deep color saturation of a big motor low mass cone. -- I'm to the point where I crave both ->>
Why choose? Use a cone for 8 months and a compression driver for 4 months, or visa versa...
The concept of a wife AND a mistress has been around for thousands of years for a reason!
<<I think you'll have a hard side deciding between the raw immediacy of the phenolic compression driver compared to the wonderful tone and deep color saturation of a big motor low mass cone. -- I'm to the point where I crave both ->>
Why choose? Use a cone for 8 months and a compression driver for 4 months, or visa versa...
The concept of a wife AND a mistress has been around for thousands of years for a reason!
Lynn Olson said:
I know this is mega-late, but I just now read this post.
I guess I should start out with what one of my physics professors said: "Everything is an approximation." It is important to note that he didn't say "Everything could be completely wrong." Newton wasn't completely wrong about gravity; his approximation to what we have now observed about gravity is extremely good in many situations. Einstein improved the model of gravity to the extent that we haven't found a disagreement with General Relativity yet, but it is expected that there will be disagreement at higher energies than we have observed. Even then, Einstein will not have been "wrong"; the theory will still work just fine for all but the highest energies.
This worldview agrees with yours about our ideas about the world not really being how the world is, but disagrees with yours when you say that "basically "everything we know" could easily be proven utterly, and completely, wrong in the next several hundred years" because when a theory is tested with experiment (or in the case of something like astronomy, observation) the theories are known to give correct predictions at a certain range of values. The probability of many independent experiments saying that a theory works when it doesn't is about as close to zero as any probability gets.
I don't know. It just irks me when people say that we don't really know anything. Something we know is that the Standard Model and General Relativity both work on every energy scale we have tested them at. These happen to be the most accurate descriptions of anything that mankind has ever produced, and that really is something.
Magnetar said:
To perfect the multi driver bass system will get your project much closer to where you'll want to dig in deeper. Even with my cheap 10's it's sooooooo nice and refreshing to hear all the low notes free from the box and compression or horn colorations.
I didn't realize Great Plains is bringing back the 290. I think you'll have a hard side deciding between the raw immediacy of the phenolic compression driver compared to the wonderful tone and deep color saturation of a big motor low mass cone. -- I'm to the point where I crave both - But-----
I really think you can have your cake and eat it too with the *right* cone mid. The 18 Sound 6" may be better than the B&C and there may be others. What's nice about the cones is they won't break the bank experimenting with them.
If you want, I'll share my cone drivers if you share yours......
Well, as part of research on the 290/390 I saw your many favorable posts about it over in the AA-HE zoo. You've owned and listened to both the 290 and the B&C cone, so I take your comments quite seriously. We may have different sets of subjective impressions, but you're certainly speaking of some of the same things I've heard directly for myself.
I've heard hints of greatness in A4-style systems, and I know there's something of value there. Never was a fan of the A7, 604 Duplex, JBL L100 or L200, but the big dog with the large-format multicell horn is another league altogether. I'm very curious just what a modern Tractrix or LeCleac'h horn can do with a large-format compression driver or a professional-grade cone midrange. Now that I've settled on a 3-way with crossovers in the 400~640 Hz and 3~5 kHz range, it comes down to selecting the right drivers, extensive measuring, and listening at length. JLH is out of the horn-building business, so I'll have to look elsewhere for the Tractrix or LeCleac'h confection to work with these drivers.
Regarding jerko's comment (I like your location, by the way), since I was raised in Japanese and Chinese culture, I take a very long historical view. If the Galactic Federation ever chooses to visit our scenic little tourist attraction, I am confident that most, if not all, of our physics would end up in the junk pile. The scary part is what would be left of our ethics and religions - it's a good question what would survive contact with civilizations thousands, or millions, of years more advanced. This is a very big Galaxy that's been around for many billions of years - I have no doubt somebody is out there.
The only question is when - tomorrow, or 10 million years from now. The radio-wave sphere now has a radius of 90 light-years - good question if there are any monitoring systems in place or not. Guess we'll find out, won't we?
The record of contact of pre-literature cultures when they met a more advanced literate culture is not a happy one - and all of humanity would then be in the same position. The hidden subtext behind the alien-invasion SF genre is the very real record of what happened to native cultures when they disintegrated on contact with more advanced cultures - they basically lost everything, their world-view, their language, their religion, and the physical places where they lived.
Our pint-sized little civilization is only about 10,000 years old, and most of humanity is preoccupied in the same sort of tribal struggles that go back to our prehistory. Only the labels have changed - and the fact that tribalism can now be promoted on TV and the Internet, so wars can engulf the entire planet, not just a scruffy bunch of smelly and greedy warriors armed with bone clubs. In terms of mental age, our civilization is in early adolescence, and is heavily armed with guns, tanks, bombers, missiles, and 20,000 nuclear weapons.
We all have our fingers crossed - we're going to need it, since a rescue by the Galactics, or your favorite semi-friendly deity, seems like a long shot. We'll have to figure it out on our own, as we've done many times before. This century should be interesting.
Andy Graddon said:
Executive Summary:
Physical condition still precludes building - getting up and down from the floor is still a hassle, and carrying things downstairs is still a somewhat nervous and awkward business requiring a cane and a nearby railing. Upstairs is much easier. Will focus on physical training in the months to come - mobility still isn't 100% there yet. Snow and ice certainly make for extremely careful ventures outdoors - canes are a little weird on bad surfaces.
Actual construction start seems probable in the Jan, Feb, March timeframe. RAAL tweeters already in hand, MLSSA works fine, SoundEasy yet to be bought. New PC will host SoundEasy, and old PC will host MLSSA. ACO Pacific 1/2" calibrated condenser mike on low-diffraction stand will perform measurements.
System will be 3-way, active crossover in the 400~640 Hz range, and implemented with a prosound crossover/parametric EQ unit. Initial setup will use old transistor amp for the bass, and the Karna amplifier for mid and HF. Mid/HF crossover passive in the 3~5 kHz range.
Bass Module: 3 or 4 15-inch drivers, two or three Eminence Delta-15LFA's, and one Great Plains Audio 416 or 515, Alnico preferred.
Mid: Three choices mentioned previously. Will measure and listen to all three, no idea at present which will be the favorite. I do not expect them to sound the same at all.
HF: Double-high RAAL in hand. Magnetar's choice of HF compression driver and horn combo sound like a good, simple starting point for comparison.
If you're expecting a kit, uh, well, it ain't gonna be like that. What will appear will be progress reports, with a number of U-turns and revisions going on for several months. My usual rate of progress on a new speaker or amplifier is 6 to 18 months, which is pretty much the high-end industry average.
The Ariel was buildable by anyone with moderate carpentry skills.
This looks like it will be a major PITA for the average DIY guy !! More a personal thing , which I think is good, btw !
3-4 x15" drivers . ouch, my wallet hurts !!
It really does sound like some sort of PA speaker that you are aiming at, which is, I suppose, the outcome expected from the thread.
Hope you get to speaker building capability soon .... !! Its is a necessary part of life after all !
This looks like it will be a major PITA for the average DIY guy !! More a personal thing , which I think is good, btw !
3-4 x15" drivers . ouch, my wallet hurts !!
It really does sound like some sort of PA speaker that you are aiming at, which is, I suppose, the outcome expected from the thread.
Hope you get to speaker building capability soon .... !! Its is a necessary part of life after all !
This is the speaker equivalent of the Amity and Karna, projects for the most advanced (and adventurous) builders only. There are no commercial equivalents of the Amity and Karna, and I don't expect any commercial equivalents of this speaker. I design products that fill empty niches in the audio-ecosphere.
It doesn't fall into a standard marketing demographic, and the Big Two magazines are not going to promote it. (You want promotion, demographics, and ultra-high WAF, build a Magico Mini.) The real money is in building speakers that are trivial variants of existing styles and market categories, with a strong emphasis on HT and decorator-friendly systems. What's a little amusing about the Ariel it was one of the very first tall and narrow MTM systems - now the format is so common it's become a cliche.
It would be nice to license this new system to a manufacturer, but that's for the future.
It doesn't fall into a standard marketing demographic, and the Big Two magazines are not going to promote it. (You want promotion, demographics, and ultra-high WAF, build a Magico Mini.) The real money is in building speakers that are trivial variants of existing styles and market categories, with a strong emphasis on HT and decorator-friendly systems. What's a little amusing about the Ariel it was one of the very first tall and narrow MTM systems - now the format is so common it's become a cliche.
It would be nice to license this new system to a manufacturer, but that's for the future.
Lynn Olson said:
Lynn, I understand your position and objectives. However, I think you are missing my point. On several occasions you have indicated that the way you will approach the dipole roll off at low frequency is to add area. I agree that this will reduce the required excursion but it is unlikely a simple solution to the dipole roll off. Eq for the dipole roll off isn't a function of driver size. It's a function of baffle size/shape.
If I can express this in another way, it seems to me that you are thinking about dealing with the dipole roll off in a manner similar to dealing with the baffle step in a conventional box speaker by designing it as a 2 1/2 way system. That works because the additional 1/2 way driver(s) bring in an additional 6dB which compensated for the 2Pi to 4Pi transition. But with a dipole system we are dealing with a 6dB/octave roll off that continues to DC. Rolling in a second driver just places a zero and a pole in the dipole roll off that puts a kink in the roll off and shifts the dipole asymptote an octave lower (assuming the drivers are all identical)as shown below, where both drivers are assumed to be flat to DC. From left to right the second driver is bought in at higher and higher frequency.
An externally hosted image should be here but it was not working when we last tested it.
The red is one driver and the green is the same driver augmented by a second driver with a 1st order LP filter. This has noting to do with the type of drivers, system sensitivity or dynamics. It is only a function the dipole delay and where the second driver rolls in. This response ultimately needs tp be equalized to achieve flat (or desired) low frequency response.
If you start considering drivers with different T/S parameters, as you suggested some time ago, then as the frequency drops off the output form the lower Q drivers will fall off faster than the high Q drivers and the dipole asymptote will turn down again comming back towards the red trace.
I might also suggest that the thinness you are referring to may well have been due to the lack of, or improper equalization applied in the mid bass region. I don't think you would find my NaO system sounding thin. However, I am sure you would not be satisfied with the dynamics. NaO panels were designed for a max SPL of 110 dB@ 1M given appropriate amplifier power which would seem to fall short of your requirement.
Using the drivers he's been talking (low QTS 12's and 15's) about will surely require massive eq even with the low pass on some of them -
Using much lower cost high QTS mid sensitivity drivers (around 92 db/w) in multiple clusters low to the ground sounds very good and requires little to no eq (provided he has a subwoofer that will keep up below 50-60 cycles!) and will meet the high output criteria. It's really pretty simple - there are others out there that are building systems like this and I look forward to seeing their end results.
I think what's happening is people get caught up on perceived gains by the sensitivity ratings of the bass drivers and MOSTLY the illusion that cost equals quality. In this application it seems cost equals more cost.
A 101 db sensitive 360 dollar .25 QTS woofer will be far less sensitive than six 92 db sensitive 60 dollar mid QTS (.6 up) woofers in a bass dipole - the multi driver high QTS system also will have higher output in the bass and require less to no eq and doesn't require the multiple low pass crossovers - I also feel the distortion per dollar will be far less with the 60 dollar woofers due to the lower excursion.
The savings alone in using this type of bass system (less complicated crossovers, amplifiers and driver costs) will free up more cash for a better mid (probably buy the mid horns or more) and will keep the design simple and more accessible to more people -
Using much lower cost high QTS mid sensitivity drivers (around 92 db/w) in multiple clusters low to the ground sounds very good and requires little to no eq (provided he has a subwoofer that will keep up below 50-60 cycles!) and will meet the high output criteria. It's really pretty simple - there are others out there that are building systems like this and I look forward to seeing their end results.
I think what's happening is people get caught up on perceived gains by the sensitivity ratings of the bass drivers and MOSTLY the illusion that cost equals quality. In this application it seems cost equals more cost.
A 101 db sensitive 360 dollar .25 QTS woofer will be far less sensitive than six 92 db sensitive 60 dollar mid QTS (.6 up) woofers in a bass dipole - the multi driver high QTS system also will have higher output in the bass and require less to no eq and doesn't require the multiple low pass crossovers - I also feel the distortion per dollar will be far less with the 60 dollar woofers due to the lower excursion. The savings alone in using this type of bass system (less complicated crossovers, amplifiers and driver costs) will free up more cash for a better mid (probably buy the mid horns or more) and will keep the design simple and more accessible to more people -
Hi Mag,
The insertion loss isn’t there but the phase shift is. I figure the insertion loss can be limited to a fraction of a dB, so no big deal.
The real bottom line is that once the baffle shape/size and driver compliment is set the system sensitivity at the low frequency cut off is fixed. What remains is to equalize the response to be flat (or match the target) above the system cut off in the dipole region and beyond. If you start with 4 identical drivers with 96dB/W 2Pi sensitivity then with a series/parallel configuration you end up with 102dB/W 2Pi sensitivity. Dipole would drop to the 4Pi sensitivity but near floor operation would restore 2Pi (102). Dipole roll off of 6dB an octave below the dipole = monopole point would drop that back to 96dB/W. Now assume four 12" drivers on a 1M x 1M baffle with a max excursion limitation of 5 mm. The dipole = monopole point will be at about 100 Hz so at 50 Hz the system sensitivity is pegged at 96dB/W. Max SPL with 5 mm is about 110dB requiring 25 watts. To get 120 dB would require an excursion of 15 mm+ and a power of 251 W. Pushing the cut off up to 72 Hz would limit the excursion to 5 mm and reduce the required power (about 3 dB) to 125w.
I'm assuming 500 cm^2 per driver here. So a total of 2000 cm^2 area. We can look for more sensitive drivers which would reduce the required power but the excursion requirements won't change. Four 15" drivers would reduce the excursion (at 50 Hz) to about 9 mm. I'm having a hard time finding prosound drivers which have these kinds of Xmax.
Now if you start mixing high Qts and low Qts drivers we also have to consider that the drivers won't be working together at low frequency. For example considering drivers with the same Fs but one pair with Qts = 0.33 and the other with Qts = 1 the low Qts driver would produce almost 10dB less output at fs, assuming the same midband sensitivity. So mixing Qts driver means we would basically have just the two high Qts drivers producing the low frequency output near Fs. so there goes another 6dB.
Going back to the sensitivity issue, if you want to target the system nominal sensitivity at sensitivity of the midbad drivers, around 102 dB, then you have to provide the low frequency eq at the line level. but the low frequency sensitivity will still be at 96dB. Placing the eq at the line level (passive or active) will maximize the potential of the amplifiers. If you use passive eq at the speaker side then the sensitivity is fixed at 96 dB and all the added midband driver sensitivity goes up in heat in the crossover and reduces the amplifiers potential to drive the system.
I guess the way I see it is that if the target is 120dB with a 25 watt SET amp or what ever the system sensitivity needs to be 106dB/W. That seems to be a pretty daunting task with any kind of dipole system with acceptable low frequency response (50 to 75 Hz cut off) assuming a conventional subwoofer is also present.
The insertion loss isn’t there but the phase shift is. I figure the insertion loss can be limited to a fraction of a dB, so no big deal.
The real bottom line is that once the baffle shape/size and driver compliment is set the system sensitivity at the low frequency cut off is fixed. What remains is to equalize the response to be flat (or match the target) above the system cut off in the dipole region and beyond. If you start with 4 identical drivers with 96dB/W 2Pi sensitivity then with a series/parallel configuration you end up with 102dB/W 2Pi sensitivity. Dipole would drop to the 4Pi sensitivity but near floor operation would restore 2Pi (102). Dipole roll off of 6dB an octave below the dipole = monopole point would drop that back to 96dB/W. Now assume four 12" drivers on a 1M x 1M baffle with a max excursion limitation of 5 mm. The dipole = monopole point will be at about 100 Hz so at 50 Hz the system sensitivity is pegged at 96dB/W. Max SPL with 5 mm is about 110dB requiring 25 watts. To get 120 dB would require an excursion of 15 mm+ and a power of 251 W. Pushing the cut off up to 72 Hz would limit the excursion to 5 mm and reduce the required power (about 3 dB) to 125w.
I'm assuming 500 cm^2 per driver here. So a total of 2000 cm^2 area. We can look for more sensitive drivers which would reduce the required power but the excursion requirements won't change. Four 15" drivers would reduce the excursion (at 50 Hz) to about 9 mm. I'm having a hard time finding prosound drivers which have these kinds of Xmax.
Now if you start mixing high Qts and low Qts drivers we also have to consider that the drivers won't be working together at low frequency. For example considering drivers with the same Fs but one pair with Qts = 0.33 and the other with Qts = 1 the low Qts driver would produce almost 10dB less output at fs, assuming the same midband sensitivity. So mixing Qts driver means we would basically have just the two high Qts drivers producing the low frequency output near Fs. so there goes another 6dB.
Going back to the sensitivity issue, if you want to target the system nominal sensitivity at sensitivity of the midbad drivers, around 102 dB, then you have to provide the low frequency eq at the line level. but the low frequency sensitivity will still be at 96dB. Placing the eq at the line level (passive or active) will maximize the potential of the amplifiers. If you use passive eq at the speaker side then the sensitivity is fixed at 96 dB and all the added midband driver sensitivity goes up in heat in the crossover and reduces the amplifiers potential to drive the system.
I guess the way I see it is that if the target is 120dB with a 25 watt SET amp or what ever the system sensitivity needs to be 106dB/W. That seems to be a pretty daunting task with any kind of dipole system with acceptable low frequency response (50 to 75 Hz cut off) assuming a conventional subwoofer is also present.
Hi
For an "easy to play around a little" with Xmax and SPL max and Sd required :
http://www.linkwitzlab.com/spl_max1.xls
Maybe also worth to have a look at to see how simple things can work also :
http://www.diyaudio.com/forums/showthread.php?threadid=110583&perpage=25&pagenumber=1
----------------------------
JohnK, couldn't you show us a simulation where you double Sd each octave downwards - lets say three times in an optimised way what would make 1 + 1 + 2 = 4 identical drivers in total ?
Or even four times doubling 1 + 1 + 2 + 4 = 8 identical drivers in total ?
Lets say the baffle peak is at 500 Hz let them come in at 300 Hz / 150 Hz / 75 Hz or so...
Greetings
Michael
For an "easy to play around a little" with Xmax and SPL max and Sd required :
http://www.linkwitzlab.com/spl_max1.xls
Maybe also worth to have a look at to see how simple things can work also :
http://www.diyaudio.com/forums/showthread.php?threadid=110583&perpage=25&pagenumber=1
----------------------------
JohnK, couldn't you show us a simulation where you double Sd each octave downwards - lets say three times in an optimised way what would make 1 + 1 + 2 = 4 identical drivers in total ?
Or even four times doubling 1 + 1 + 2 + 4 = 8 identical drivers in total ?
Lets say the baffle peak is at 500 Hz let them come in at 300 Hz / 150 Hz / 75 Hz or so...
Greetings
Michael