Some inspiration for open boxes like discussed in #8720 and the following posts http://www.diyaudio.com/forums/multi-way/192737-2-way-waveguide-cardioid-like.html
What is used for the treble? Also any XO info would be appreciated.
If I interpret the dimensions correctly, it's a horizontal cab (a really big one for the Beta 8 at that). I'm guessing that means a seperate horn on top?
Sorry all the questions; despite all the gear detail on Gary's site, there's nary a peep about the speakers.
Thanks!
-- Mark
Well, until recently Gary was using a Chinese ribbon (don't know which one) with a computer-based crossover and multi-amping, Beta 8 for the mids, and four 15" woofers in two open-baffle W boxes for the stereo-pair subs. (Four 15-inchers in all.) Very heavy EQ for the subs, pretty much no EQ for the Beta 8's, and not much for the ribbon tweeter.
I talked to Gary on the phone just a couple of days ago, and he's built a custom LeCleac'h wood horn for the Radian 475 1" exit compression driver, with a T=0.707 and entrance angle of 5.5 degrees to match the 5.5 degree exit angle of the Radian 475. The horn is a 1 kHz horn, and he's using a 2 kHz 4th-order highpass filter with a bit of in-band EQ.
He was previously using a commercial Tractrix horn with the 475, and he's much much happier with the LeCleac'h, reporting pretty much the same things I noticed with the AH425 ... very deep images, very natural spatial qualities, and vivid dynamics, with much less horn coloration than the Tractrix horn. He's really thrilled with the combination of a wood horn, the LeCleac'h T ratio of 0.707, and exact matching of exit and entrance angle between compression driver and horn ... very low measured diffraction, smooth response, and much more dynamic sound than the ribbon.
Since Gary has an electronic crossover, all he has to do is turn a virtual knob with the mouse at the listening position, using his big LCD TV screen as a monitor, to evaluate the sound of different crossover frequencies and (acoustic) slopes. With the 475, 1.4 kHz was a bit too low, and 2 kHz sounded great.
His next plan is for a bigger LeCleac'h horn, 425 Hz, or maybe even 360 Hz, since the maximum size of the AH425 was constrained by needing to fit into an Australian Post box for economical shipping from Australia. (FedEx would be about US$500 or more, while AusPost is about US$120 a pair) For the larger horn, Gary's considering either the Radian 745Neo or possibly the really big one favored by JMMLC himself, the 950Neo. One advantage of the 950 is you can get a 4" beryllium diaphragm for it right now, although they're not cheap.
I talked to Gary on the phone just a couple of days ago, and he's built a custom LeCleac'h wood horn for the Radian 475 1" exit compression driver, with a T=0.707 and entrance angle of 5.5 degrees to match the 5.5 degree exit angle of the Radian 475. The horn is a 1 kHz horn, and he's using a 2 kHz 4th-order highpass filter with a bit of in-band EQ.
He was previously using a commercial Tractrix horn with the 475, and he's much much happier with the LeCleac'h, reporting pretty much the same things I noticed with the AH425 ... very deep images, very natural spatial qualities, and vivid dynamics, with much less horn coloration than the Tractrix horn. He's really thrilled with the combination of a wood horn, the LeCleac'h T ratio of 0.707, and exact matching of exit and entrance angle between compression driver and horn ... very low measured diffraction, smooth response, and much more dynamic sound than the ribbon.
Since Gary has an electronic crossover, all he has to do is turn a virtual knob with the mouse at the listening position, using his big LCD TV screen as a monitor, to evaluate the sound of different crossover frequencies and (acoustic) slopes. With the 475, 1.4 kHz was a bit too low, and 2 kHz sounded great.
His next plan is for a bigger LeCleac'h horn, 425 Hz, or maybe even 360 Hz, since the maximum size of the AH425 was constrained by needing to fit into an Australian Post box for economical shipping from Australia. (FedEx would be about US$500 or more, while AusPost is about US$120 a pair) For the larger horn, Gary's considering either the Radian 745Neo or possibly the really big one favored by JMMLC himself, the 950Neo. One advantage of the 950 is you can get a 4" beryllium diaphragm for it right now, although they're not cheap.
Last edited:
Hi-Vi planar. A sort of quasi-ribbon. 1st order high pass circa 5K, IIRC. It mated well with the Beta 8. Beta 8 had a small inductor in series to calm the slight rising response.
There are some online courier services now that give better prices, which perhaps makes the 340Hz JMLC azurahorns worth looking at . eg a carton 52 x 52 x 40cm to USA (from Oz) is $175. 61 x61 x 49cm is $295. (interparcel.com.au via ups). A lot less to NZ or SE Asia. Still expensive but in terms of overall cost might make sense. BTW just a hobby for me.
martin
Is there any real advantage to a 340 Hz Horn compared to the 425 Hz version when using a 15" / 16" woofer?
the 340 is larger, heaver, more expensive, harder to ship, presumably has more beaming at higher frequencies.
In return, group delay and pattern control are lowered another 1/3 octave.
the 340 is larger, heaver, more expensive, harder to ship, presumably has more beaming at higher frequencies.
In return, group delay and pattern control are lowered another 1/3 octave.
The exponential nature of sound really confronts you with horns. Perhaps the size difference is the best intuitive measure. Certainly it doesnt buy much on the linear scale of things. The 425 horn is a great landing spot for many reasons, but in the end it is the bass / mid reproducer that determines what you need. At least with Jean-Michel's curve the character of presentation is the same in all sizes. Add construction rigidity to advantages of smaller horns.
Martin brings up important points about the size of the horn. Bigger is not always better; what you win in some areas is taken away in others.
What's a little unusual about the AH425 is that it's fairly compact (16.5" overall diameter) for a large-format compression driver. It's not a horn designed for a 500 Hz crossover; that would be about 45% larger, 24" or so.
I did a bit of research on the performance of direct-radiator paper-cone 12" and 15" drivers, and the versions that are not designed for PA use have pretty good performance through 1 kHz, with a broad, low-Q peak around 1.5 kHz. The pro drivers intended for high-power use have heavy, rigid cones, with complex breakups in the 1.2 kHz and higher regions.
I don't want to rely on digital brickwall filters to prevent power from going to complex breakup regions; I'd rather select drivers with are controlled and well-behaved from 1 to 5 kHz. The Altec/GPA 416 and 515 fit the bill, but there are other pro drivers designed for studio-monitor use that are suitable as well. I expected the JBL 2226 to be suitable, but it didn't audition that well compared to the GPA 416 Alnico.
Even though these drivers are being used in their piston band (flat-response region), there are pretty noticeable subjective differences between them. Some of the subjective differences may be the result of out-of-band artifacts, while others may come down to differences in surround, suspension, and magnetic-field geometry in the gap.
What I've learned about horns is that "stretching" bandwidth does not work well; a given driver and horn has a preferred working range, and trying to squeeze out a few more Hz at the bottom of the range isn't a good idea. The Radian 745 has a stiffer suspension (stretched Mylar) than the crimped-aluminum tangential suspension of the Altec/GPA 288, and this affects how low the compression driver can go before noticeable IM distortion starts creeping in. So I'm not sure that a 500 Hz crossover, along with a larger horn, is even a good idea for the Radian 745. The Radian 950 is probably a better fit for that setup.
It's the builders using folded bass horns that have to make some hard choices: many folded bass horns are in pretty bad shape by 500 Hz, never mind 700 Hz, where the response is a rumpled mess. Horns that have flat response (and good power-handling) down to 300~350 Hz are really big, in the 32~40" region, and the best driver is probably a cone, not a compression driver. Time alignment gets tricky with horn that deep ... it's either digital time correction and crossover, which implies multi-amping, or the loudspeaker itself starts to get quite large and quite deep. So there's a whole collection of tradeoffs when HF horn has to meet a folded bass horn.
The alternative is direct-radiator bass, with a choice between two 12" drivers (high efficiency but with a higher F3 frequency), a single 15" driver (good overall compromise), and two 15" drivers (cabinet starts to get large and heavy). An interesting option with a vertical array of two bass drivers is separate lowpass filters for each driver, which gives a provision for a moderate degree of time-aligning between upper and lower driver.
The two-drivers in a vertical array, with the lower driver almost touching the floor, has some interesting characteristics when the floor image is considered. At the lowest frequencies, there are effectively four drivers in a vertical array, with the lower two comprised of the floor images (carpet absorbs less than 1 dB in the 50 Hz to 500 Hz range). Yes, we're back to the line array, in a way, but it's lowpassed at 700 Hz.
Last edited:
I keep hearing from the experts about how conventional dome tweeters are smoother than compression drivers and better in the home environment. So, I started experimenting with the Vifa XT25 ring radiator in a SEOS-12 waveguide. Results are promising. No problems with distortion or perceived dynamics, as long as you watch the low end on the tweeter. And you get that last octave in the top end, which is missing in most 1" CDs, and is non-existent in 2" CDs. It really makes a difference in terms of the ambiance of the recording.
The EQ required to flatten the response is substantial though, and I haven't gotten the overall balance right yet. But it does sound better defined than a CD. Instruments have distinct signatures, and notes have better precision/definition.
Has anybody else made similar comparisons? Are the response curves from CDs truly smooth? I guess we need high-res measurements (Toole recommends a 2 Hz resolution) that cannot be obtained through gating.
The EQ required to flatten the response is substantial though, and I haven't gotten the overall balance right yet. But it does sound better defined than a CD. Instruments have distinct signatures, and notes have better precision/definition.
Has anybody else made similar comparisons? Are the response curves from CDs truly smooth? I guess we need high-res measurements (Toole recommends a 2 Hz resolution) that cannot be obtained through gating.
ra7,
I have also mounted dome tweeters on horn lenses and have had some success with this approach. The differences with a compression driver are many though. For one you can not load the dome tweeter as you can with a compression driver as this will cause many problems with a typical soft dome as they can not handle any of the typical loading of the diaphragm as a true compression driver is made to handle. So the efficiency will never approach a CD and that is one of the real limitations with this approach. Another factor is that while you can get increased output in the lower registry of the dome tweeter as you have found you will then cause a real decreasing output with increasing frequency and this will require much eq which is something of a vicious circle of attempting to get back to what you had before adding the horn or waveguide. Typically you will see very short waveguides used in this type of application as you are using a shallow waveguide only for directivity and not for any gain. But even a short waveguide will cause some tilting of the response curve and this will again require eq to correct. Since the dome tweeter does not create a waveform in the same manner as a compression driver you will create some unusual interactions between the dome tweeters normal directivity and any horn or waveguide. It really is a complex situation and this is one of the reasons this type of application is usually only seen in powered monitors that have much eq to correct some of the problems created by the mismatch between the two disparate devices. Don't get me wrong, it can be made to work but it is not a simple substitution of a CD for a dome tweeter, it is much more complicated than that in reality.
I have also mounted dome tweeters on horn lenses and have had some success with this approach. The differences with a compression driver are many though. For one you can not load the dome tweeter as you can with a compression driver as this will cause many problems with a typical soft dome as they can not handle any of the typical loading of the diaphragm as a true compression driver is made to handle. So the efficiency will never approach a CD and that is one of the real limitations with this approach. Another factor is that while you can get increased output in the lower registry of the dome tweeter as you have found you will then cause a real decreasing output with increasing frequency and this will require much eq which is something of a vicious circle of attempting to get back to what you had before adding the horn or waveguide. Typically you will see very short waveguides used in this type of application as you are using a shallow waveguide only for directivity and not for any gain. But even a short waveguide will cause some tilting of the response curve and this will again require eq to correct. Since the dome tweeter does not create a waveform in the same manner as a compression driver you will create some unusual interactions between the dome tweeters normal directivity and any horn or waveguide. It really is a complex situation and this is one of the reasons this type of application is usually only seen in powered monitors that have much eq to correct some of the problems created by the mismatch between the two disparate devices. Don't get me wrong, it can be made to work but it is not a simple substitution of a CD for a dome tweeter, it is much more complicated than that in reality.
Thanks Kindhornman. The EQ is indeed hard to master. I haven't observed what's happening off-axis in the ring-radiator and SEOS-12 combo, but it will be interesting.
The most important thing though is whether it results in a smoother response compared to CDs. I will be looking into it, but was wondering if anyone else has seen it before.
Smooth means the small wiggles and resonances are absent. These are only visible in un-smoothed high resolution measurements.
The most important thing though is whether it results in a smoother response compared to CDs. I will be looking into it, but was wondering if anyone else has seen it before.
Smooth means the small wiggles and resonances are absent. These are only visible in un-smoothed high resolution measurements.
Here's the XT25 in a WG-300 waveguide: Test Vifa XT-300/K4 (XT25TG-30/04) am Waveguide WG-300
Some data for you to compare it to the SEOS waveguide...
Some data for you to compare it to the SEOS waveguide...
Last edited:
My experience with domes loaded in waveguides is not nearly as complex as yours Kindhornman. As you mention, energy decreases with frequency. This usually allows the use of a simple one cap network, or a single cap plus contour circuit, which is generally as simple as, or even simpler than your average dome tweeter network.
Here are some examples of well designed dome waveguides: acoustic waveguides Measurements: https://plus.google.com/photos/101632266659473725850/albums?banner=pwa
Here are some examples of well designed dome waveguides: acoustic waveguides Measurements: https://plus.google.com/photos/101632266659473725850/albums?banner=pwa
I wonder about matched termination to the driver. The LeCleac'h (and most likely, other horns) are sensitive to the exit angle of the compression driver matching the entrance angle of the horn. If they don't match, you get reflections, which shows up as ripples in the FR and in the time-domain response.
Dome tweeters are intended to be accurately flush-mounted on a flat surface; even mismatches of 1mm are measurable, and edge diffraction from the boundaries of the front baffle create substantial reflections that are measurable and audible. In the development of the Ariel, there were measurable and audible differences between 0.75" and 1.5" radii on the cabinet edges, with the larger radius obviously superior in terms of image depth and freedom from coloration on vocals.
Mounting a dome tweeter in a waveguide or horn is nowhere close to what the original designer intended, so it's not surprising that a lot of equalization is required ... if the dome is designed to be flat-response radiating into a hemisphere, it certainly won't be flat radiating into a waveguide or horn.
EQ is all well and good, but what about time response? That's what I'd be concerned about. What kind of reflections are created when the air-load, as seen from the diaphragm, is completely different than the intended design condition?
At an intuitive level, I'd expect the LF response of the dome tweeter mounted in a waveguide or horn to be substantially boosted, with an increase in power-handling. I'd also expect to see a series of ripples in the FR and time domain, as a result of mismatch, that are not present when the dome-tweeter is mounted on a large flat baffle.
Ra7, you mentioned a need for equalization when a dome tweeter is mounted in a waveguide. Is this HF boost equalization, or do ripples need to be notched out as well?
Last edited:
When experimenting with a SEOS HA-25 waveguide in front of an RAAL 70-20XR(OEM), I noticed that an airtight seal was needed between the WG and ribbon face place, or there were ripples and other nasties present. AutoTech HA-25 - AutoTech Waveguides - Waveguides DIY Sound Group
Thanks Wesayo, Face, Lynn. Face, I hadn't seen those waveguides before. Thanks. Will order some to play with.
Lynn, the low end of the XT25 gets a huge boost, about 10db in the 1-4 kHz range. Pretty useful. Further up, the response evens out. But as you mention, there are many dips and peaks, apart from the low end boost. These, I'm guessing, are reflections near the mouth. The radiating area of the XT25 is really quite large, definitely larger than the 1" entry of the SEOS-12.
Apart from real measurements, I guess I was posing a more fundamental question about whether given the right matching between dome/ring radiator and waveguide, can it be superior to a CD/horn combo? Is the response of a traditional tweeter inherently smoother than that from a CD? If all we are craving is directivity, then it might be a worthwhile solution.
I doubt any tweeter can compete with the low end of a 2" CD, or even a 1.4" CD, but the XT25/SEOS-12 combo easily does 1.2 kHz, LR4.
Lynn, the low end of the XT25 gets a huge boost, about 10db in the 1-4 kHz range. Pretty useful. Further up, the response evens out. But as you mention, there are many dips and peaks, apart from the low end boost. These, I'm guessing, are reflections near the mouth. The radiating area of the XT25 is really quite large, definitely larger than the 1" entry of the SEOS-12.
Apart from real measurements, I guess I was posing a more fundamental question about whether given the right matching between dome/ring radiator and waveguide, can it be superior to a CD/horn combo? Is the response of a traditional tweeter inherently smoother than that from a CD? If all we are craving is directivity, then it might be a worthwhile solution.
I doubt any tweeter can compete with the low end of a 2" CD, or even a 1.4" CD, but the XT25/SEOS-12 combo easily does 1.2 kHz, LR4.