pattern control below 400Hz

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My current system is a horn >700Hz, dual 15" midbass, dual 15" sub. The main flaw seems to be uneven midbass (in a big open room with lots of glass + hard floors). Better pattern control would be nice.

I made a prototype horn based on the Syncalc spreadsheet:

Coverage angle 90 x 65.52
~60cm wide, horizontal pattern control to 400Hz

I picked an easy build (mostly 45 degree angles, no secondary flare), and easy components, so it worked out OK.

I'd like to maintain some directivity < 400Hz, rather than have an abrupt transition to omni, as a small woofer would do.

Hence I'm considering using the biggest drivers I have (a couple of 18"), in boxes like the "Resistance enclosure" on this site:
Cardioid bass

That seems to have good output and directivity to < 100Hz in a reasonable sized box that doesn't require heroic crafting skill or lots of iterations to build.

Is there a similar / better option I've missed? e.g. would 2 x 15" arranged like the woofers in the small syns work better?
Small Syns

Notes:

the mains won't have to go super low, because subs.
happy to multiamp and use DSP, but don't want it to get super complicated

Limitations:

width <80cm
I'd like to keep everything simple(ish) so I have a decent chance of making the final build look pretty.
I'd like to use up old drivers I already have rather than buy new (one pair 18" EV woofers, multiple pro 15").
 
Last edited:
I've done some info hunting:

- there's more info on cardioid response / resistive enclosures than I thought. Some of it is on this site. Some of it is very pertinent.
- doing a Google search, then choosing results that take me to this forum seems to find better results than the forum’s own search function.

I've done some info experiments:

- I'm staying with a small cone driver at the apex. It sounds better to me than the couple of 1" compression drivers I tried.
- an 18" on OB gives plenty of midbass, and seems to go low enough to work with subs. Sweeps show high distortion ~100Hz, but that's probably because the baffle is cardboard and cable ties :) The raw driver is OK.

I'm now trying to minimax the design:

1) how can I lower diffraction without overly complicating the build?
2) how can I avoid "waistbanding" without overly complicating the build?
3) the test rig has the woofer at floor level. If I raise the system to get the horn to ear height, will I make the bass worse?
- would it be good to use two woofers (vertically) to prevent a floor bounce suckout?

Pix: the directivity image is a Yorkville Unity, taken from the Red Spade blog, to illustrate "waistbanding". The others are from my prototype.

My FR plot has the midbass crossed at 1kHz, no eq. The HF driver is also crossed at 1kHz + has simple eq to address the HF rolloff.
 

Attachments

  • small prototype.jpg
    small prototype.jpg
    132.1 KB · Views: 775
  • small proto in use.jpg
    small proto in use.jpg
    106.5 KB · Views: 775
  • Syn mids and highs 1kHz cross.jpg
    Syn mids and highs 1kHz cross.jpg
    91.3 KB · Views: 1,013
  • Directivity-Yorkville.jpg
    Directivity-Yorkville.jpg
    76.8 KB · Views: 766
Pattern control means a larger horn.. you can't rely on the woofer (even at 18 inches) to do that at the ~500hz region.

Since i built that K402 replica a while ago i have thought of ways to make it simpler to build - one way would be to use foam at the mouth to combat the waistbanding instead of a tractrix curve ala:
118293_18106.jpg


Alternatively, i would just build an even larger conical horn

What driver are you using for your horn, and how sensitive do you estimate it being in the horn? I'm thinking of getting an sb65wbac25-4 for a K402 like horn
 
Last edited:
Radiation pattern control below 400Hz indoors is a real challenge. A dipole or cardiod woofer can do that in principle, and perhaps also practically down to 100hz. But when you put that under your synergy on the floor it doesn't work any more.

Best practice is to just move speakers andl listening spot around the room to find best compromise with reflections and modes.

Go to this page hunecke.de | Loudspeakers Calculator

Give diemnsions of your room and construction materials. Scroll down to the colourful image of a room. Grab a speaker or mic, drag it and see how measured response and modes change! works also vertically. The simulator supposes that loudspeakers have typical radiation pattern, you can change speaker dimensions.
 
This depends on listening distance..

-you can do a horizontally-opposed design with 2x 15" all playing in the same bandwidth.

Say a 40" wide baffle cardioid (resistive side-vent on both sides) with each 15" at either end of the baffle and side venting. You can adjust the upper freq. pattern with how the woofers are "tilted" on the baffle relative to each other as in the Everest below:
 

Attachments

  • 3.png
    3.png
    465.1 KB · Views: 144
What driver are you using for your horn, and how sensitive do you estimate it being in the horn? I'm thinking of getting an sb65wbac25-4 for a K402 like horn

Cone driver NE65w-04.

I compared it to a JBL 2" earlier, to get a guesstimate of sensitivity:

paper cone open frame driver for front horn

Post 39 shows the horn the comparison was done on.

Pattern control means a larger horn
But that's what I'm hoping a more directive midbass would help with. I don't wanna go wider than them horn in that other thread. And I want it to be nicer looking :)
 
A dipole or cardiod woofer can do that in principle, and perhaps also practically down to 100hz. But when you put that under your synergy on the floor it doesn't work any more.
It does if you design it with the floor and top mounted stuff as part of the system. A passive cardioid enclosure is difficult to design. A box with two woofers (separate air spaces) and DSP is quite easy. Just place the box + top mounted stuff on a parking lot, measure both woofers at a certain distance at the angle at which you want maximum cancellation, add processing to make their amplitude and phase responses equal, then flip the polarity of the rear woofer. Then you end up with a cardioid response up to some frequency that is determined by the physical dimensions. A smaller box works up to a higher frequency.
 
Last edited:
Say a 40" wide baffle cardioid (resistive side-vent on both sides) with each 15" at either end of the baffle and side venting. You can adjust the upper freq. pattern with how the woofers are "tilted" on the baffle relative to each other as in the Everest below:

That size and multiplicity of angles is a bit beyond me. I doubt I could fit a complex face seamlessly to the waveguide mouth.

It does if you design it with the floor and top mounted stuff as part of the system.
I was thinking (and have begun experimenting) along those lines.

A passive cardioid enclosure is difficult to design.

That's not good to hear. The Kimmo Saunisto site makes it seem easy.

Since I have a couple of cabinets I can cut up and try things out on, I've started chopping + measuring. So far, it doesn't look good - but it is early days yet.

If the passive cardioid thing doesn't work for me (that is: if it requires a frustrating number of trials), my fallback position is to just package a nicer version of my prototype parts into simple towers - similar layout to this Lambda Unity Tower picture (from Cowan Audio website).

To me, this looks better than an Everest style box.
 

Attachments

  • unity_front.gif
    unity_front.gif
    137.1 KB · Views: 401
To me, this looks better than an Everest style box.

..that's a lot more narrow.

full-on cardioid for that, and the best thread for that (with resistance) is here:

Adventures in cardioid


Note that there is another method for producing a cardioid response (or even hyper-cardioid) than acoustic leakage/resistance, and that's with 2 drivers in their own enclosures effectively "back-to-back" with the front in-phase and the rear out-of-phase (..like a dipole) but with DELAY for the front driver (..that would work well with integrating the upper-freq. horn). ..in addition to delaying the front, there is usually some "shaping" to the rear's response to get the best pattern for a specific directivity result, and of course there is the "boost" on the lower-end (..like you'll need with just about any cardioid setup). Because you are using DSP and you have extra drivers on-hand, this is probably the best method.. but it will require experimentation.
 
Last edited:
Note that there is another method for producing a cardioid response (or even hyper-cardioid) than acoustic leakage/resistance, and that's with 2 drivers in their own enclosures effectively "back-to-back" with the front in-phase and the rear out-of-phase (..like a dipole) but with DELAY for the front driver (..that would work well with integrating the upper-freq. horn). ..in addition to delaying the front, there is usually some "shaping" to the rear's response to get the best pattern for a specific directivity result, and of course there is the "boost" on the lower-end (..like you'll need with just about any cardioid setup). Because you are using DSP and you have extra drivers on-hand, this is probably the best method.. but it will require experimentation.

Does the rear wall cause problems with the rear firing speaker? There would be a secondary bounce off the back wall just like with dipole setup?
 
The Dutch and Dutch 8c speaker appears to have done passive cardioid right.

It uses a forward firing 8" that covers 100-1250Hz. The manual (and patent EP3018915A1) claims that "The passive cardioid cabinet weakens rearward radiation by over 15 dB"

The spec sheet states that the rear facing 8" drivers are crossed over at 100Hz, so it seems like they are simple subs, and do not contribute to the directivity of the forward facing driver. The manual recommends 10-80cm between walls and the rear of the cabinet.

I'd like those characteristics, but for about $15,000 less :)

The image is from their patent, and doesn't appear to require rocket surgery - it looks almost exactly like the boxes on the Kimmo Saunisto site.
 

Attachments

  • imgf0001.png
    imgf0001.png
    52.9 KB · Views: 356
Does the rear wall cause problems with the rear firing speaker? There would be a secondary bounce off the back wall just like with dipole setup?

-it depends on the freq..

If you want it up as high as 1 kHz (including crossover region), then you should have a total width just a bit less than a foot.

Yes, some secondary out-of-phase depending on the cardioid pattern, but it should be lower in level with an appropriate crossover.. but the result won't be dipole'ish, rather it will be like a cardioid or hyper-cardioid.
 
Last edited:
I agree [...]

And, the genesis right here on diyAudio

Thanks for the review link. I like that "there is virtually no speaker boundary interference", and it works in real world positioning. I've got a similar space - hard surfaces, lots of glass - and a partner who is OK with big speakers, but wants them close to the walls, not out in the room.

...and thanks for the thread link. I'd seen parts of that thread before (that's where I saw the Dutch & Dutch patent), but I didn't join the dots / didn't pay enough attention.

While he guards the info about exactly how the 8c has better rear rejection than the early prototypes, the prototypes are pretty good. I'm grateful for the additional good tips for DIY people he's mentioned.

..who do you think the IP behind Dutch & Dutch "is"? :p

Again, I hadn't joined the dots.

What is the link - is he an employee, is his stuff listed as "prior art", was it just a jumping off point..?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.