SimontY said:
Hi,
Stuff such a Linkwitz/Riley crossover values for zobelled drivers.
I perfectly understand the electrical theory and know that in reality
these values are meaningless, what you want in reality is an acoustic
L/R response and the only way to get this depends on the drivers
response in situ, the actual crossover values cannot be calculated.
🙂/sreten.
Thanks Streten. I see where you're coming from and am surprised the book doesn't go as far as real world design in such depth.
I still intend to buy it though 🙂
Simon
I still intend to buy it though 🙂
Simon
Thanks for that Sreten, some things I really have trouble with still, 9.5dB is not to be sneezed at is it??
I was joking about the "declaration of war" speaker I hope every one realised that
I was joking about the "declaration of war" speaker I hope every one realised that
SimontY said:
Hi,
As I understand you really do need to know your loudspeaker onions
to not get the wrong end of the stick from some parts of that book.
🙂/sreten.
sreten said:
... until the enclosures are built, the drivers are in place, and precise phase and frequency response plots are obtained for each driver.
Sreten pretty much summed it up, except for the difference in sound. I build exclusively sealed box speakers. I do this because when I started, I had read a lot of articles about how much better sealed boxes were than ported boxes. I've never spent time comparing sealed vs. ported in listening tests, so I can't really offer an opinion what sounds better. I might be making things harder than they have to be by building only sealed boxes.
I am currently forced to listen to a pair of store bought ported speakers, and I don't particularly care for the bass. It sounds "hummy" to me.
If I assume 1.5 inch thick walls all the way around, that comes out to about 4.8 cubic feet. The L21's only needed 1 cu ft each of sealed box volume. I believe it is the same for the L22's. The 8" speakers that Moondog found (good job!) are perfect for sealed boxes and need even less volume. So that size box is just about right.
Actually, crossover pro calcuates the acoustic phase response. If your Thiele small paramaters are pretty accurate, the mid tweeter crossover phase comes out very close to what is modelled, and can easily be verified by inverting the tweeter polarity and measuring the dip at the crossover frequency. I've had little success measuring a dip between the woofer and mid. I think it is due to room reflections, but it could be do to poor modelling of phase by the simulator.
I find it much harder to model relative driver levels than the phase response.
It can't predict how it will sound. That is a matter of build it and listen. Change crossover, and listen some more.
JJ
If you do build these I would like very much to come around and listen one day; while my post about running 10 drivers in the bass was meant as a joke, the idea of di-pole bass perhaps has serious merit, thee are othersin the forum who are much more experienced and qualified than I who could explain the need for baffle step compensation.
Just remember that it all has to balance, PHL have some good mid-ranges avaliable but seriously expensive, but excellent SPL
Regards Ted
Just remember that it all has to balance, PHL have some good mid-ranges avaliable but seriously expensive, but excellent SPL
Regards Ted
sreten said:
Hmmm....... I haven’t worked through to the topic of baffle step correction yet, so I’m still not entirely sure what that is, but I think an infinite baffle enclosure with four 8” drivers with sufficiently low sensitivity and fs/Qts<50 is a near certainty.
This will be quite achievable within my cabinet dimension limits (especially if I take advantage of the space behind the MTM enclosure instead of sealing off above the woofer compartment and increase the box compliance with 100% fibre fill) and it is just so much easier to get right than a ported enclosure.
sreten said:
Because I can 🙂
SimontY said:
Yes, for a sealed enclosure it is still only +6dB, f3 and Q remain the same and the box needs to be twice as big (over just two drivers in parallel).
However, the input impedance is almost unchanged over a single woofer (the series-parallel 4 woofer option has a slightly lower resonant frequency), the load is distributed over double the cone area giving less distortion at a given SPL and power handling is much improved.
A sealed box comparison between a single driver, two in parallel and 4 in series parallel is made in the cookbook. I can scan it for you if you like.
Cheers,
Glen
Moondog55 said:
No worries. The only problem with that though is I'm a bit far away, down south of the sleepy city of Adeliade.
I'll take note to check out the PHL mid range drivers. Thanks for the tips so far!
Cheers,
Glen
My wife and I could do with a second honeymoon but it's a long way to drive for a week-end, I do like the "City of Churches " tho, some great jazz venues in and around North Terrace if I remember correctly.
PHL drivers avaliable in Oz from Audio Marketplace
http://www.audiomarketplace.com.au/...acturer_id,0/option,com_virtuemart/Itemid,49/
They are also Australian agents for Shiva and Maelstrom sub-woofer drivers so we are unable to deal directly with US manufacturer/supplier.
I think Soundlabs gives a discount if you purchase 10 drivers or more ( usual practice ) use 6 and parcel off the other 4.
Ted
PHL drivers avaliable in Oz from Audio Marketplace
http://www.audiomarketplace.com.au/...acturer_id,0/option,com_virtuemart/Itemid,49/
They are also Australian agents for Shiva and Maelstrom sub-woofer drivers so we are unable to deal directly with US manufacturer/supplier.
I think Soundlabs gives a discount if you purchase 10 drivers or more ( usual practice ) use 6 and parcel off the other 4.
Ted
jupiterjune said:
Ok, I understand this graph now (I think 😕 ) I don’t have access to my Cookbook where I am ATM, but some musings anyway.
The midrange and tweeter will have +6dB sensitivity as they operate above the baffle step (speed of sound / enclosure width), so you need and extra +6dB from the woofers.
This would make good use of half the +12dB gained from the quad woofer array. Yes?
Suppose I use four of the TIW 200 XS sub woofer drivers for the WWWW in a 100L sealed box:
http://www.soundlabsgroup.com.au/p/V-1340-TIW200XS/TIW+200+XS+-+8+Ohm
These are rated at 84dB (1W/1m).
4 in parallel gives +12dB, so 96dB total sensitivity.
So, assuming a baffle step of 6dB, the midrange and tweeter sensitivities would have to be 90dB (1W/1m)
This seems acceptable. For instance the midrange linked to by Moondog55 in the post above.........
http://www.audiomarketplace.com.au/...d,0/option,com_virtuemart/Itemid,49/vmcchk,1/
….. is rated at 97dB half space (1W/1m). It also has a sufficient power rating. It is expensive, but not so bad if I can get away with just one midrange driver.
A couple of quick Qs:
If my enclosure has a width of 0.3m the baffle step will be at 343/.3 = 1140Hz.
This is significantly higher than the previously suggested bass crossover (~500Hz). Is this what I want?
The midrange driver sensitivity of 97dB is specified at "half space". The woofer sensitivity at 84dB is not. Are the two figures directly comparable? If so, is 7dB an excessive amount to pad down the midrange, making a lower sensitivity unit a better option?
Cheers,
Glen
I hope some-one else can answer the querys about baffle step, my feelings are to avoid padding down the midrange if at all possible;-( why pay for all that sensitivity and then throw it away with 50cent resistors??
You might derive some benefit from this builder's experience.
Here is a link to his HTGuide profile which has links to his projects and a photo of the speakers in question.
http://htguide.com/forum/member.php4?u=7788
He titled these speakers "Project Overkill"
http://www.htguide.com/forum/showthread.php4?t=20872
Check the photos attached to post #24 and #35.
While it takes him a while to narrow it down, the original concept was a 4-way system with a W(Mb)MTM(Mb)W, where (Mb) is MidBass.
In post #35, you see his concept is 12"+10"+7"+1.125"+7"+10"+12"
But I think in his final design, starting from the top, he went with 12"+12"+7"+1.125"+7"+10"+10"
Eventually, he decided to cut the speakers apart and have them in two separate cabinets (If I understand correctly).
So, now he has one cabinet with his subwoofer (12"+12") and a second cabinet with his (Mb)(Mb)MTM (10"+10"+7"+1.125"+7").
http://www.htguide.com/forum/showthread.php4?t=27020
Though there might be something in his experience that would benefit you.
Steve/bluewizard
Here is a link to his HTGuide profile which has links to his projects and a photo of the speakers in question.
http://htguide.com/forum/member.php4?u=7788
He titled these speakers "Project Overkill"
http://www.htguide.com/forum/showthread.php4?t=20872
Check the photos attached to post #24 and #35.
While it takes him a while to narrow it down, the original concept was a 4-way system with a W(Mb)MTM(Mb)W, where (Mb) is MidBass.
In post #35, you see his concept is 12"+10"+7"+1.125"+7"+10"+12"
But I think in his final design, starting from the top, he went with 12"+12"+7"+1.125"+7"+10"+10"
Eventually, he decided to cut the speakers apart and have them in two separate cabinets (If I understand correctly).
So, now he has one cabinet with his subwoofer (12"+12") and a second cabinet with his (Mb)(Mb)MTM (10"+10"+7"+1.125"+7").
http://www.htguide.com/forum/showthread.php4?t=27020
Though there might be something in his experience that would benefit you.
Steve/bluewizard
the 1660 phl midrange linked to above is the one I use.
has tons of speed and real bite.
however, it cannot be crossed low (I cross mine at 200 hz, but with around 70 db slopes)
has tons of speed and real bite.
however, it cannot be crossed low (I cross mine at 200 hz, but with around 70 db slopes)
G.Kleinschmidt said:
Thanks Glen, it does sound interesting, but I do actually intend to buy this book now ASAP, as I'm embarking on another speaker project and could do with it for reference purposes!
G.Kleinschmidt said:
It sounds very high, is it halved afterwards or something?? I recall when I modified a crossover from 2 way to 2.5 way (to warm the sound up with BSC) I brought my second woofer in at about 350hz or so. And that was a narrow (8" or so) baffle, and those speakers do sound good!
I also calculated it should be higher (for my speaker), I remember 400hz or so coming to mind, but in the end I copied the inductor choice from Tony Gee's Proteus speaker, figuring such a speaker guru would know some reason to do it that way.
I expect listening tests and some measurement might be needed for fine-tuning this value.
Simon
BlueWizard said:
Thanks for the link steve. That guy makes me feel quite sane 🙂
SimontY said:
Hi Simon.
The baffle step or the crossover frequency?
I haven’t read up on baffle step compensation yet so I am likely talking out of my depth here, but the TIW 200 XS 8” woofer, despite its healthily low fs of 29Hz, has excellent bandwidth, being 10dB down at 7.5kHz.
I suppose that with a –12dB/octave crossover it can be crossed over pretty high. That would conveniently relax the bandwidth requirement for the mid range driver(s) I presume, which would be fortunate as the mid-range drivers with +90dB efficiency are likely to require a high crossover.
Cheers,
Glen
Hi,
All speaker senstivities are quoted into halfspace.
Baffle step for a 0.3m width will be all over by 1kHz.
(Seas data sheets are worth looking at. They show bass drivers
measured in real boxes, the baffle step is very evident, as is
the bass vs midband sensitivities)
An approximation for the BS midpoint is 1/3 of a wavelength, but
this does to a degree depend on trhe form factor of the baffle.
~ 90dB/2.83V is a perfectly sensible target, the numbers do stack up.
Now for the bad news ........
4 8" drivers in a vertical array crossed over high will not sound good at all.
You will have lobing issues and more importantly inconsistent
dispersion or power response into the room. The array at higher
frequencies will tend towards a line source where level falls
off inversely to distance, whilst at lower and higher frequencies
dispersion will tend towards a point source where level falls
off inversely to distance squared.
In other words it will sound different at different distances.
(I had a related problem with a 200mm bass + 500mm ribbon.)
IMO you cannot go there, i.e. implement all BSC in the bass array.
🙂/sreten.
Attached some quick modelling of 4 Visatons.
Note the green vented rolls off faster than the sealed yellow.
The red vented rolls of slower than sealed and should match room gain.
All speaker senstivities are quoted into halfspace.
Baffle step for a 0.3m width will be all over by 1kHz.
(Seas data sheets are worth looking at. They show bass drivers
measured in real boxes, the baffle step is very evident, as is
the bass vs midband sensitivities)
An approximation for the BS midpoint is 1/3 of a wavelength, but
this does to a degree depend on trhe form factor of the baffle.
~ 90dB/2.83V is a perfectly sensible target, the numbers do stack up.
Now for the bad news ........
4 8" drivers in a vertical array crossed over high will not sound good at all.
You will have lobing issues and more importantly inconsistent
dispersion or power response into the room. The array at higher
frequencies will tend towards a line source where level falls
off inversely to distance, whilst at lower and higher frequencies
dispersion will tend towards a point source where level falls
off inversely to distance squared.
In other words it will sound different at different distances.
(I had a related problem with a 200mm bass + 500mm ribbon.)
IMO you cannot go there, i.e. implement all BSC in the bass array.
🙂/sreten.
Attached some quick modelling of 4 Visatons.
Note the green vented rolls off faster than the sealed yellow.
The red vented rolls of slower than sealed and should match room gain.
Attachments
Sreten would you say then that WWMTMWW is an option??
and does this mean that crossing over lower is the best option??
Ted
and does this mean that crossing over lower is the best option??
Ted
sreten said:
OK, this is getting back to one of my original questions, namely, is the concept of a vertical line array intrinsically flawed?
But what you say here WRT to on-axis beaming is exactly true for any single woofer. The horizontal polar response with frequency is related to the diameter of the woofer. How does four woofers in a vertical array complicate matters and would four 8" drivers crossed over at, say 500Hz sound any worse than just one crossed over at 500Hz?
The Cookbook gives -6dB attenuation at +/-45 degrees as a acceptable level for on-axis beaming and -3dB as a more stringent standard.
For an 8" woofer this equates to crossover frequencies of 1302Hz (-3dB) and 2055Hz (-6dB).
These are rather higher crossovers than ~500Hz and an 8" woofer would seldom be crossed over that high other than in a 2-way system.
Nothing is said about vertical polar response though, which I assume is what you are talking about?????
An alternative direction - two 4 ohm 10" or 12" woofers in push-pull with a D'Appolito MTM top?
Cheers,
Glen
- Status
- Not open for further replies.