Did you see my comment that these speakers will reside in a "nook". By that I mean a recessed area along a wall that runs floor to ceiling, is 20 inches deep and 6 feet wide. The speakers must be placed fully in the nook, should not stick out, and are located against each sidewall of the nook. There just isn't any room for a dipole (or two!) in that space, and is very close to boundaries. So I am left with few good options.
I mostly want to match the woofer pattern to the next highest driver where it crosses over. I can live with widening as frequency decreases, but I would prefer more control if possible.
I might have phrased it a bit vaguely. What I meant is that higher order gradients are impractical, yet that the B&O is worth checking out despite that.
I had read it, but this explains it better for me. A picture would still help though! In line with your initial idea, you could go the infinite-baffle route, with very large drivers to get directivity.
Member
Joined 2003
I've used wool felt...IIRC it was 1/2 or 3/4" F7 contained between two stiff screens around the back of these speakers. My main purpose was to damp resonances rather than quasi-cardioid radiation which was secondary.
I second the IB suggestion.
I second the IB suggestion.
Update:
I am planning to build (at some point, it's about #3 or #4 on my audio to-do list) a system that uses the "larger is better" principle for directivity control. This will have a compression driver + horn up top, then a sealed 8" driver as a "midrange", and finally another different 18" driver as the "woofer". It's an anti-mini monitor! I was initially planning to make a 2-way with an 18, but I decided that I would add in ANOTHER 18 for the midrange and I found this really nice Selenium model that perfectly fits my needs on sale. I really like that low power amps (e.g. 50W) have more power than I need because the sensitivities of these drivers are very high. Class-A anyone?
The drivers are:
horn: SEOS-12
CD: DN-360 1.1kHz-20kHz
midrange: Selenium 18SW600 200Hz-1.1kHz
woofer: Peavey Low Rider 18 below 200Hz
Cabinet volume (it's a lot):
225 liters for the Peavey woofer
~50 liters for the selenium midrange
Using the above volume for the woofer gives me 2 options at the same time that I can switch between:
1) low Q sealed cabinet+LT, Q~0.6 Fb~50Hz; with LT Fb=35Hz, Q=0.7 (+6dB boost)
2) vented cabinet+2order HP filter, Fb~27Hz, F3~29Hz
I like the ability to switch, but I will need to have a customized crossover for each one because of the additional HP filters. The midrange is very small and will result in some response peaking, but this will be below the low edge of the passband.
I'm trying to figure out how I want the build the cabinet. I could make something about 24" wide, 12" deep and 6 feet tall (Gumby configuration, tall and thin) or I could make a 24Wx24Dx40" tall cab (mailbox configuration). I am also thinking about where to put these beasts in my home (yes these will be home speakers). If I use the mailbox configuration I will need to place the speakers out in the room somewhere (not as wife friendly) but with the Gumby configuration I could put the speakers right up against the wall. These speakers will be pretty directional down to 300-400Hz because of the large format midrange. I think that this will reduce or eliminate reflections from the back wall in the Gumby configuration. But BECAUSE the drivers are so directional I will have to be no more than 20 degrees off axis. In the mailbox configuration I could turn/toe the speakers in towards the listener as I see fit. But the speakers cannot be turned in the Gumby configuration, so I might have to put them pretty close together, which might not work out well.
So, should it be Gumby or Mailbox? (or fur Deutscher: Dietschi oder Datschi?) 🙂
.
I am planning to build (at some point, it's about #3 or #4 on my audio to-do list) a system that uses the "larger is better" principle for directivity control. This will have a compression driver + horn up top, then a sealed 8" driver as a "midrange", and finally another different 18" driver as the "woofer". It's an anti-mini monitor! I was initially planning to make a 2-way with an 18, but I decided that I would add in ANOTHER 18 for the midrange and I found this really nice Selenium model that perfectly fits my needs on sale. I really like that low power amps (e.g. 50W) have more power than I need because the sensitivities of these drivers are very high. Class-A anyone?
The drivers are:
horn: SEOS-12
CD: DN-360 1.1kHz-20kHz
midrange: Selenium 18SW600 200Hz-1.1kHz
woofer: Peavey Low Rider 18 below 200Hz
Cabinet volume (it's a lot):
225 liters for the Peavey woofer
~50 liters for the selenium midrange
Using the above volume for the woofer gives me 2 options at the same time that I can switch between:
1) low Q sealed cabinet+LT, Q~0.6 Fb~50Hz; with LT Fb=35Hz, Q=0.7 (+6dB boost)
2) vented cabinet+2order HP filter, Fb~27Hz, F3~29Hz
I like the ability to switch, but I will need to have a customized crossover for each one because of the additional HP filters. The midrange is very small and will result in some response peaking, but this will be below the low edge of the passband.
I'm trying to figure out how I want the build the cabinet. I could make something about 24" wide, 12" deep and 6 feet tall (Gumby configuration, tall and thin) or I could make a 24Wx24Dx40" tall cab (mailbox configuration). I am also thinking about where to put these beasts in my home (yes these will be home speakers). If I use the mailbox configuration I will need to place the speakers out in the room somewhere (not as wife friendly) but with the Gumby configuration I could put the speakers right up against the wall. These speakers will be pretty directional down to 300-400Hz because of the large format midrange. I think that this will reduce or eliminate reflections from the back wall in the Gumby configuration. But BECAUSE the drivers are so directional I will have to be no more than 20 degrees off axis. In the mailbox configuration I could turn/toe the speakers in towards the listener as I see fit. But the speakers cannot be turned in the Gumby configuration, so I might have to put them pretty close together, which might not work out well.
So, should it be Gumby or Mailbox? (or fur Deutscher: Dietschi oder Datschi?) 🙂
.
I haven't looked at Selenium drivers in a while.........I wish all manufacturers were so detailed with their data! If their response graph is to be believed the break-up looks pretty benign, though I wonder if the narrowing then widening from 800-1khz is indication of that. Off the top of my head I can't think of anyone running an 18 that high in a sound quality app.....I'm interested to see how this works out for you.
I always try to do something "different", and this is certainly in that vein.
The polar plots in the datasheet for the 18WS600 are very nice to have, but they don't quite seem to jibe with the frequency response plot... The FR plot shows on axis (solid) and 45 degrees off axis (dashed) lines. At at 1.25kHz the dashed line looks to be down almost 10dB while at 800Hz is down only 5-6dB. Looking at the polar plots for 800Hz and 1.25kHz I get the feeling that these two have been REVERSED! In that case, the driver looks very good up to about 1kHz before beaming increases.
Also at about 1kHz the SEOS horn is losing directivity as frequency decreases and the CD is running out of steam as well. There should be a frequency somewhere around 1kHz, perhaps slightly higher, where the directivity of the Selenium 18 and the SEOS horn match up well.
So, while it's something different, I think the combination of these two could produce a nice smooth directvity pattern that doesn't widen until a relatively low frequency (OK, not 200Hz, but maybe 400 or 500Hz).
I have looked at that, and what I found was that the 18 just didn't handle the HFs as well as a 15. Had to move the crossover lower, OK, but then the waveguide would need to grow to handle the directivity down lower, and maybe the CD is getting a little weak. Jump to a larger CD and then you have a HF problem at 8-10 kHz because the power response of a larger CD is failing. Things start to run away fast for a relatively small gain in LF directivity.
Charlie, I've read some comments by the SEOS designer (here somewhere) where it was mentioned that the fortuitous response was based partly on luck.
OK, nothing wrong with that. I'd be reluctant to use it outside the intended range. My visual assessment (read, guess..) would be not to push the BW without eg. extending the vertical flare and measuring.
OK, nothing wrong with that. I'd be reluctant to use it outside the intended range. My visual assessment (read, guess..) would be not to push the BW without eg. extending the vertical flare and measuring.
The "intended range" meaning the low end of the range that the SEOS horn and CD should be used is about 1kHz, and that is where I am crossing over to the midrange. Measurements on the horn support that this will actually work out very well.
So you won't cross over to the 18" at 1 kHz? (I don't think that there is an 18" that can do that.) Where will the lower crossover be?
I thought I was pretty clear about my intended crossover points in post #24.
If you look you can find 18s that make it to 1kHz. I'm using one that has response up to almost 3kHz (on axis at least). I wouldn't use it up there, but crossing over to it around, or even slightly above, 1kHz is no problem.
Datasheet is available here:
http://www.parts-express.com/pedocs/specs/264-394-selenium-18ws600-specifications-45295.pdf
Charlie - if it is that well behaved above 1kHz, then yes, its quite usable. Non that I ever tested looked that good - but then my real tests always look different than the published ones (some companies the specs are total fiction. others quite good. I don't know Selenium.) What usually happens is a much larger peak at say 2 kHz that is too high to allow a 1 kHz crossover.
Last edited:
I had some Danley SH50s in my house for a few days, and the 28" waveguide made a noticeable improvement in imaging. I'm guessing the crossover made a difference too.
I've listened to Bill Waslo's Synergy horns at his house, and dare I say it, they were superior in some respects to The Real Deal. They're a really REALLY nice design, and probably the best DIY speaker I've ever heard.
The fact that no one but Bill has built a pair is a tragedy.
Hi Earl,
the problem with dipoles/supercardioids/cardioids being hungry for displaced volume at LF lies in the fact, that the dipole pathlength in given desings mostly is not chosen according to wavelengths to be radiated IMO.
There is a frequency feq where e.g. dipole and monopole will radiate equal SPL on axis with same displaced volume of the driver.
If you go one octave below feq, the dipole needs approx double Vd like a monopole: That is quite "tolerable" IMO to still call the design "efficient".
Going even lower, the dipole needs further doubling in Vd for every octave in relation to the monopole: That will be too costly soon for most applications, sure.
In a dipole, you will have the additional problem then, that approx. 1 octave above feq the dipole pattern starts to deteriorate.
So an "efficient" dipole - like described above - for LF might hardly have more than 2 octaves useable bandwith ... that seems not very attractive.
But this not a problem with a real cardioid, if the rear source is low passed properly, and the front source's DI "does the rest" at mid frequencies, to maintain a rather constant radiation pattern. One may still need a midrange unit capable of low crossover frequency.
But having pathlengs long enough in a cardioid woofer to be "efficient", can only be achieved, when using a dedicated subwoofer, that is either a monopole or also a dipole or cardioid, then with appropriate dipole pathlength on it's own.
Using this line of thought, there is no efficiency problem anymore when making use of directional LF systems, having DI around 5dB or so.
We just would have to accept, the cardioid woofer being designed as a satellite consequently, relying on a separated dedicated subwoofer system (Multisub, Dipole Subs, SBAs, DBAs as one pleases) .... which is to be prefered as a low to low-mid bass source anyhow for reasons of balanced mode excitation.
Cheers Oliver
Last edited:
Charlie
Your math is not quite right.
Yes there is a frequency at which the dipole and monopole are equally efficient - when ka = 2.0. For a 15 " woofer this is about 560 Hz. For each o0ctave below this the monopole requires 1/2 as much excursion as the dipole. This is, as you say, a runaway situation.
A cardioid is a mixture of a monopole and dipole and for the sake of simplicity lest assume that they are equal (its not that simple in reality.) This means that the cardiod needs sqrt(2) times as much excursion as the monopole for each octave or basically twice for each two octaves. This is clearly better, but at no point could it ever be considered "not a problem with a real cardioid". At 35 Hz the cardioid still needs four times as much excursion as the monopole. "Doable" - maybe, but not easily.
If you are saying to some how role from a cardioid to a monopole at LFs then this is doable and shown in my book, but requires two drivers with independent amps (possible passively, maybe, but that would need to be shown.) "Doable", but at what cost. "Practical?" Not from where I sit.
Your math is not quite right.
Yes there is a frequency at which the dipole and monopole are equally efficient - when ka = 2.0. For a 15 " woofer this is about 560 Hz. For each o0ctave below this the monopole requires 1/2 as much excursion as the dipole. This is, as you say, a runaway situation.
A cardioid is a mixture of a monopole and dipole and for the sake of simplicity lest assume that they are equal (its not that simple in reality.) This means that the cardiod needs sqrt(2) times as much excursion as the monopole for each octave or basically twice for each two octaves. This is clearly better, but at no point could it ever be considered "not a problem with a real cardioid". At 35 Hz the cardioid still needs four times as much excursion as the monopole. "Doable" - maybe, but not easily.
If you are saying to some how role from a cardioid to a monopole at LFs then this is doable and shown in my book, but requires two drivers with independent amps (possible passively, maybe, but that would need to be shown.) "Doable", but at what cost. "Practical?" Not from where I sit.
Earl
i was just pessimistic to also include dipoles and supercardioids ...
Sure real Cardioids get away with less Vd.
But i feel you completely missed my take on adjusting the dipole pathlength (and the use of a dedicated Subwoofer):
One is not restricted to "driver size" serving as pathlength.
We can design the enclosure to a certain depth and baffle size and thus put feq where we need it, in order to get acceptable efficiency ...
(btw. it's not Charly posting ...)
Last edited:
I think you guys don't actually disagree. At low frequencies you need a large distance between the two poles, otherwise you a lot of displacement is required. No discussion there. Pro Audio cardioid subs are usually about one meter deep, which results in an effective acoustic distance of a little over a meter. The result is that on-axis both drivers have the same phase at approximately 70 hertz (+6 dB relative to one driver) and only below 25 hertz will the combined output be less than that of one individual driver.
Absolutely! Below you see the far field response at different angles of a 15" driver in a passive cardioid enclosure, for which I did exactly that. Only below approximately 130 hertz does its on-axis output start to drop, in comparison to a monopole. The combination of said speaker and an active cardioid subwoofer has both very good directivity and it goes really loud!
Absolutely! Below you see the far field response at different angles of a 15" driver in a passive cardioid enclosure, for which I did exactly that. Only below approximately 130 hertz does its on-axis output start to drop, in comparison to a monopole. The combination of said speaker and an active cardioid subwoofer has both very good directivity and it goes really loud!
@keyser: I didn't notice you are professional manufacturer !
Your product looks well designed for purpose, maybe even looks a bit too precious for touring and PA ...
I will keep your homepage in mind, maybe we could communicate via PM about one thing or two next time.
Your product looks well designed for purpose, maybe even looks a bit too precious for touring and PA ...
I will keep your homepage in mind, maybe we could communicate via PM about one thing or two next time.
Last edited:
- Status
- Not open for further replies.