Hi all. I'm hoping for expert opinion.
Other than the size and building difficulty, is there any reason not to put an Alpair 10.3 in a tractrix front horn? I'm thinking maybe something around Fc~150 Hz or more. I have another tractrix mid-bass horn that I love the sound of.
Profile is attached. I'm planning on a sealed rear compression chamber tuned to horn Fc. Red line marks the horn throat. Does this look about right?
Would the highs suffer from the waveguide?
Grazie.
Other than the size and building difficulty, is there any reason not to put an Alpair 10.3 in a tractrix front horn? I'm thinking maybe something around Fc~150 Hz or more. I have another tractrix mid-bass horn that I love the sound of.
Profile is attached. I'm planning on a sealed rear compression chamber tuned to horn Fc. Red line marks the horn throat. Does this look about right?
Would the highs suffer from the waveguide?
Grazie.
Thanks Dave.
I'm a bit confused. I understand that the horn will boost the SPL of the lower frequencies. I also believed that it would selectively attenuate the higher frequencies, causing a roll-off. Your post makes me think this is incorrect (which would please me) and that HF would be the same as say direct radiation with a sealed box. I'd love to make this work from a few hundred Hz through ~15kHz or so, but maybe this is unrealistic.
Cheers.
I'm a bit confused. I understand that the horn will boost the SPL of the lower frequencies. I also believed that it would selectively attenuate the higher frequencies, causing a roll-off. Your post makes me think this is incorrect (which would please me) and that HF would be the same as say direct radiation with a sealed box. I'd love to make this work from a few hundred Hz through ~15kHz or so, but maybe this is unrealistic.
Cheers.
Okay. Thanks for the input. I'm happy to take your advice and reserve the 10.3s for a pair of FHXL. Out of curiosity, what is it that you think the 10.3s do so well? With the FLH, I was hoping to capture significant low-distortion dynamics over a wide bandwidth. I kind of like the sound of a bit of slightly 'enhanced' mid-bass as well.
Can anybody point me toward a reference for calculating the high-frequency response of a tractrix horn? It'd be cool to learn something from the theory rather than getting the response from compiled software.
Thanks!
Can anybody point me toward a reference for calculating the high-frequency response of a tractrix horn? It'd be cool to learn something from the theory rather than getting the response from compiled software.
Thanks!
Like in direct radiation, the dispersion range is determined by the effective diameter of the source. In a horn, it's the throat. The larger the throat, the narrower it gets.
In addition, a cone driver inevitably breaks up above certain frequency. So the HF waveform it sends into the horn via the big throat is far from ideal. HF reflections within the horn would be messy, thus being attenuated before they leave the horn and into the room.
So in general, above certain frequency, a big throat horn gives no help to the driver, even somewhat harmful instead. A few successful examples come with drivers which tend to have very uptilt response - excessive HF energy to spare. And the horn would be better short to decrease the internal reflection.
A big midrange horn has its charm, but if it can't play up to the top, then it'd be a huge challenge to get it right in the transition to another HF unit.
In addition, a cone driver inevitably breaks up above certain frequency. So the HF waveform it sends into the horn via the big throat is far from ideal. HF reflections within the horn would be messy, thus being attenuated before they leave the horn and into the room.
So in general, above certain frequency, a big throat horn gives no help to the driver, even somewhat harmful instead. A few successful examples come with drivers which tend to have very uptilt response - excessive HF energy to spare. And the horn would be better short to decrease the internal reflection.
A big midrange horn has its charm, but if it can't play up to the top, then it'd be a huge challenge to get it right in the transition to another HF unit.
Last edited:
I have made tractrix for a 3.5 in fullrange driver and the mids are amplified 10dB and it kills everything above 8 to 10 kHz. For vocals though it sounds very nice and lush. Like CLS says, when you stick anything in front of a direct radiator, it will impact the HF production as they rely on cone breakup modes and those end up getting reflected and absorbed by the horn wall.
http://www.diyaudio.com/forums/full-range/223313-foam-core-board-speaker-enclosures-187.html#post3541667
http://www.diyaudio.com/forums/full-range/223313-foam-core-board-speaker-enclosures-187.html#post3541667
A 150 Hz horn will act over 3-4 octaves. 150->300->600->1200->2400.
You will be boosting up to somewhere between 1200-2400 Hz and will need to be EQed down to create a flat response.
dave
diyaudio is such a great place to learn.
CLS: Is it worthwhile to look at large-throated front horns with co-axials or even a full range with whizzer cone? Maybe at some point, the horn would be sufficiently wide and the geometric spreading of the highs from the whizzer sufficiently narrow that you wouldn't get too much reflection. Also, is there some practical limit to of the ability of low-frequency waves to be audibly affected by a whizzer? Or said differently, at what frequency will the wavefront be relatively insensitive to a few-cm wide scatterer (if there even is a limit)?
Your short-horn comment is taken on board.
xrk: the foam-core project is beautiful. Looks MUCH easier to work with than bending wood. Again, comments on cone breakup modes are taken on board.
Dave: Thanks for the correction. I guess I like a bit of mid 'enhancement'. I'm guessing one thing you think the 10.3 does well is the flat response, which is somewhat negated by a boost in relative lows and reduction in relative highs.
So, thanks all. It seems that the answer to my original question, "is there any reason not to do this?", is YES. I'm sure these seem like basic questions to all of you, but at my current state of knowledge, they are relevant to me!
Cheers.
CLS: Is it worthwhile to look at large-throated front horns with co-axials or even a full range with whizzer cone? Maybe at some point, the horn would be sufficiently wide and the geometric spreading of the highs from the whizzer sufficiently narrow that you wouldn't get too much reflection. Also, is there some practical limit to of the ability of low-frequency waves to be audibly affected by a whizzer? Or said differently, at what frequency will the wavefront be relatively insensitive to a few-cm wide scatterer (if there even is a limit)?
Your short-horn comment is taken on board.
xrk: the foam-core project is beautiful. Looks MUCH easier to work with than bending wood. Again, comments on cone breakup modes are taken on board.
Dave: Thanks for the correction. I guess I like a bit of mid 'enhancement'. I'm guessing one thing you think the 10.3 does well is the flat response, which is somewhat negated by a boost in relative lows and reduction in relative highs.
So, thanks all. It seems that the answer to my original question, "is there any reason not to do this?", is YES. I'm sure these seem like basic questions to all of you, but at my current state of knowledge, they are relevant to me!
Cheers.
For more than one source in a single horn, I'll suggest:
http://www.diyaudio.com/forums/multi-way/217298-square-pegs-14.html
http://www.diyaudio.com/forums/multi-way/254468-pse-144-horn-discussions.html
Those are totally different animals.
Whizzer cone driver in a big throat/ big mouth horn is the way of hi-fi 'purist'. I've heard Lowther PM4 in Oris 150 and I like the sound very much. Although the combination is not an all-rounder, it's very pleasing sounding nevertheless. (the cost of driver is a big let down for sure... )
As to a coaxial driver in a big throat/ big mouth horn, in the era with Oris, I myself had thought a lot about it, too. But eventually I gave up the idea because, sorry for saying this, I feel it's kind of a half-baked solution. It seems it brings more problems than it solves.
One of the very important factor is, it seems no coaxial driver is designed for horn-loaded. They tend to fit in compact boxes for their major application.
Tannoy might be the only exception. Its 15" coaxial is both front and rear loaded. But in the front, the 'throat' is so large and the 'horn' is so short and no effect on its HF unit. In this case, it's more on midrange directivity control, rather than horn loading it.
http://www.diyaudio.com/forums/multi-way/217298-square-pegs-14.html
http://www.diyaudio.com/forums/multi-way/254468-pse-144-horn-discussions.html
Those are totally different animals.
Whizzer cone driver in a big throat/ big mouth horn is the way of hi-fi 'purist'. I've heard Lowther PM4 in Oris 150 and I like the sound very much. Although the combination is not an all-rounder, it's very pleasing sounding nevertheless. (the cost of driver is a big let down for sure... )
As to a coaxial driver in a big throat/ big mouth horn, in the era with Oris, I myself had thought a lot about it, too. But eventually I gave up the idea because, sorry for saying this, I feel it's kind of a half-baked solution. It seems it brings more problems than it solves.
One of the very important factor is, it seems no coaxial driver is designed for horn-loaded. They tend to fit in compact boxes for their major application.
Tannoy might be the only exception. Its 15" coaxial is both front and rear loaded. But in the front, the 'throat' is so large and the 'horn' is so short and no effect on its HF unit. In this case, it's more on midrange directivity control, rather than horn loading it.
Quite, although most coax units are by nature mildly horn loaded (at least the tweeter is).
Front horns are one of the great ways forward, but they're rarely ideally suited to wideband drivers for the reasons Dave notes -exceptions being those with a rapidly rising IB response and / or (preferably both) a rapidly narrowing dispersion as frequency rises, where you can (sort of) get away with it, if you're willing to make sacrifices to be 'pure.' Which is rarely a good idea to my mind, since it usually entails major trade-offs that seem rather more significant than the nebulous benefits lumped under the 'purity' banner. YMMV as always of course.
The Tannoy Westminster front horn does provide a mild boost / fill-in between the bass (back) horn rolloff & the XO to the HF unit. Nice speakers. Although the oft-repeated comment in advertising / reviews that it's a 'full sized' back horn are twaddle, as even the most cursory glance at the terminus size shows.
Front horns are one of the great ways forward, but they're rarely ideally suited to wideband drivers for the reasons Dave notes -exceptions being those with a rapidly rising IB response and / or (preferably both) a rapidly narrowing dispersion as frequency rises, where you can (sort of) get away with it, if you're willing to make sacrifices to be 'pure.' Which is rarely a good idea to my mind, since it usually entails major trade-offs that seem rather more significant than the nebulous benefits lumped under the 'purity' banner. YMMV as always of course.
The Tannoy Westminster front horn does provide a mild boost / fill-in between the bass (back) horn rolloff & the XO to the HF unit. Nice speakers. Although the oft-repeated comment in advertising / reviews that it's a 'full sized' back horn are twaddle, as even the most cursory glance at the terminus size shows.
Last edited:
I dunno about that. A full range or wideband driver in a front horn is one of my all time favorite speaker formats -- assuming there's a good woofer unit and supertweeter to go with it.
For reasons already mentioned, picking a driver with at least a mildly rising response is a good idea. However, to some degree the falling HF can be compensated for with a lower crossover to the supertweeter and/or EQ.
To get an idea of the expected bandwidth and efficiency, most people would run a sim in McBean's hornresp program. It's free and relatively easy to use. It doesn't predict the real world HF response of a FR driver in a horn though, it only shows the gain BW. Real world HF typically extends much higher, but it will still be rolled off, particularly in the top octave.
RE the AP 10.3: I strongly prefer paper cones for horn loading. For whatever reason metal cones haven't ever sounded good to me when loaded with front horns. Horns seem to exacerbate any metallic signature, even if it's some quirk that isn't audible when direct radiating. I have the 10.3 and it's been my daily listener lately, but it wouldn't be my choice for horns.
For reasons already mentioned, picking a driver with at least a mildly rising response is a good idea. However, to some degree the falling HF can be compensated for with a lower crossover to the supertweeter and/or EQ.
To get an idea of the expected bandwidth and efficiency, most people would run a sim in McBean's hornresp program. It's free and relatively easy to use. It doesn't predict the real world HF response of a FR driver in a horn though, it only shows the gain BW. Real world HF typically extends much higher, but it will still be rolled off, particularly in the top octave.
RE the AP 10.3: I strongly prefer paper cones for horn loading. For whatever reason metal cones haven't ever sounded good to me when loaded with front horns. Horns seem to exacerbate any metallic signature, even if it's some quirk that isn't audible when direct radiating. I have the 10.3 and it's been my daily listener lately, but it wouldn't be my choice for horns.
FYI, based on published specs, frequency response, the upper limit for loading this driver is ~143 Hz, so you're already 'dead in the water', so to speak, leaving only a really large conical or parabolic WG to 'focus' its polar response to a lower F6.
GM
Bert Doppenberg of BD-Design/the Oris horn told me that a fronthorn will fall off with about 3dB/octave above that. So if you find a driver with roughly the inverse response, you get good linearity. At a certain point above the point you describe, that 3dB rule probably doesn't hold, the part of the cone still vibrating probably doesn't "see"/feel the impedance of the horn anymore.
There are measurements of a BD-Design customer on the BD-Design forum, the customer measured an 8" tangband fullrange driver in one of the Oris horns. The response is essentially linear way up high and especially the off-axis response looks very good. Equally linear, but a few dB down from the on-axis response. The horn increases directivity in the pass-band you describe and when that meets the start of driver beaming above that, it works out really well.
I liked my Lowthers in backloaded horns, but when I finally tried out fronthorns, it was like they used to be broken and I fixed them. I agree not all drivers match well to all horns, but when you design a good combo, it is very easy good sound.
Finally, Voigt really designed the essential fullrange driver for frontloaded horns. I think that if a driver closely resembles Lowther drivers in size, lay-out and parameters, it is likely a decent frontloaded driver. If a driver deviates much, its intended and best application is probably different. The Alpairs are clearly engineered to radiate freely, I wouldn't bother trying out a fronthorn for them.
Sorry, some hard won knowledge I keep for myself, but those 'practiced in the art' [to borrow from the many patents I've gleaned info from] of speaker/horn design can find it 'buried' in one of Harry Olson's horn papers done decades before T/S filter theory became popular.
GM
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.