[nullspace]

You're more than welcome, Dan, but I didn't point out anything you wouldn't have figured out eventually. You seem to have an earnest desire to learn about this stuff and that's half the battle, as they say.

Generally, I think the curves look very good and if you're increasingly happy with them then you're certainly headed in the right direction. If those were mine, I'd like to know more about what's going on in the 1.8k-3khz range but at this point I think that's being picayune.

Best of the luck with the box-building. I'm terrible at working with wood and don't much look forward to that part of projects.

John

[DanTheMan]

Yea, that's my biggest area of concern.  It's no longer audible as an error, but I do know it's there.  My bet it once the tweeter is boxed, it'll be OK.  We'll see.  Otherwise I don't really care enough to try to improve it.  Crazy thing is that it is right in the middle of where the ear is most sensitive.  Strange to me that it's not glaringly obvious to my ear.  The other thing is that maybe it's just do to past experience that this sounds so neutral and good.

The way the previous crossover was, the "S" curve was audibly obvious.  It could also be that a sin of omission is just less obvious.  One other thing is that I've adjusted my toe in to be less severe.  IOW, my listening axis is a bit more toward the center of the speaker than it was.

thanks again!

Dan

[DanTheMan]

John, I've been looking at your build and thinking about you saying the no bass below 50Hz.

I made an OB once not long ago:



Here's the 11.25 off axis graph:



22.5 off axis (what I was hoping to be my listening axis):



Let's move to 45 degrees:



and at 67.5 off axis:



and finally the listening position response:



 
If that last graph was taken at the time of arrival from the speaker to the loudspeaker, it would look something like the 45 degree angle graph.
You've certainly solved a lot of these problems with the waveguide, but the baffle's woofers and MW should behave in a similar manner.  You can see in my last graph that there is a lot of reflection coming off the bump in the power response and it was definitely audible and reduced the image clarity.  It gave me the sense that there was too little bass and too much midrange as well.  Even though I'm not getting as deep of bass now without the boost, it sounds like I have a better balance either way.  I was thinking, mastering engineers do not use OB speakers and therefor OBs w/o a boosted low end will just sound bass light.  I hate to say it b/c I love the idea of OBs, but I can't go back now unless I could bring the power response back to more like a cardioid.  That's what made me want to build a box again.  I tried to reasonably get the balance to align with an OB, but it just seemed a daunting challenge.  Everything started looking like a small midrange, a WG tweeter and a number of large woofers with a lot of custom filters.  There are more options if you're interested in thinking out loud here or starting a new thread to consider. At about 1.3 kHz, it looks like your baffle may behave the same way as my smaller baffle did in the 1-4 kHz range.  You may want to look there for the lack of bass character.

Hopefully that may shed some light on what you might be hearing,

Dan

[nullspace]

Thanks for your thoughts, Dan, but I think it's pretty clear that the lack of low end from my speaker is both a function of the woofers I used and a miscalculation on my part in the low-end boost the woofer filter would provide.

The midbass, ToneTubby 12", is very efficient -- the mfg. says 97.3dB/1W and I think that's low; book values for Qes and Fs are 0.86 and 80.3Hz and mine measure 0.762 and 89.9Hz, so unless Vas is wicked low (I don't have a decent setup to measure Vas) my pair is somewhere in the neighborhood of 98.5dB/1W. To put together a woofer section, even if there's four in parallel, that'll keep up sensitivity-wise @ <50Hz is a tall, tall order.

As well, the Eminence Kappa-12A I'm using have a resonant freq. of 45Hz, and there's just no way you're going to get open baffle bass below Fs. Also, I had expected the series resistor + series inductor to provide some peaking in the reponse at Fs -- instead, the peaking is at 55Hz which turns out to be the effective end of bass reproduction.

I'm actually pretty happy with it -- I'd have to be, since I've been listening in mono for almost four months now. But there is a distinct absence of energy in the bottom octave and a half that makes some music less enjoyable to listen to.

The ~1.3kHz bump you point out (it's actually at 1.2kHz, but who's quibbling?) isn't a baffle artifact; it's a combination of cone breakup for the ToneTubby and the widening pattern of the horn. Since the crossover is right at 1.2kHz, you get a little of both for the price of just one...

John

[DanTheMan]

I listened to mono OB for 2 years!   Was absolutely please with the price/performance.

I didn't realize that your crossover was so low.   Judging by the response i had assumed higher.  One should never assume.

I think if I were you I'd just add a sub then, I understand the desire to keep it simple.

good luck, and keep us informed.  I'd love to talk more about why you chose all the different aspects of you design, but perhaps this isn't the place.

Good stuff,
 
Dan

[DanTheMan]

Update on some crossover work.  I figured out what's going on in the 2-3k department. It boils down to Eminence's graph being well off.
Here's the polar graph of the woofer:



So I built a notch filter with some spare parts:



Then looking at the tweeter's response:



It was clear that is wasn't the major offender.

The woofer sans notch filter looks like this:



Then add in the notch:



Much closer to ideal.

The whole response prior to notch:



Then with notch added:



Notice the bass doesn't drop off on this b/c I wasn't running a fullrange signal on the rest of the graphs to save time.

Finally a little different notch filter:



Just an update,

Dan

BTW, I think the first notch is a bit better.









 

[nullspace]

Hi Dan,

I'm sure you realize this by now, but you'd have a significantly easier time if you had either:
A) a horn/compression driver combo with a bit more BW on the low end, or
B) a better-behaved mid driver.

At 2kHz, the mid is well into breakup and the treble is at the end of it's line, making blending the two together very difficult. I will say, given all that, your simulations look pretty good. From an aural standpoint, I'd be listening for the mid's cone breakup at 2kHz.

Thanks for posting the ongoing evolution.

John

[DanTheMan]

I wish that midbass was as well behaved as the Eminence graph!  I do have a couple tricks up the sleeve yet.  I hope to make it work.

Yea, the midbass break up is audible.  Not as noticeable as you may think, but certainly not ideal.

Thanks for the thoughts,

Dan

[*Scotty*]

Dan,what are we seeing in the graph at the 6.9k cursor mark. The response starting at this point and continuing on out to where the highs roll off looks kind of ragged. Is this a measurement artifact or is it real.  The response on the tweeter looks great below the 6.95k point.
Scotty

[DanTheMan]

Yea, I was just marking the "sweet zone" so to speak.  I believe the ragged response that is real as I don't have it when I measure other speakers.  Not really sure what's causing it.  Right now my biggest concern is the crossover.  That raggedness will be my next battle though and I've got some ideas of how to fix it.  My best guess it that it's the roughness in the throat of the Dayton waveguides.
Here's what a different and cheaper driver looks like on it:



Has anyone mentioned that perfecting a speaker is quite a lot of work!

Dan

[DanTheMan]

I think I'm getting a bit closer:


old best:

[AudioVoice]

lots of info's about horn speakers at my forum :

http://www.audiovoice-acoustics.com/forum/index.php

55 pages of different systems from all around the world, which i update frequently.

http://www.audiovoice-acoustics.com/...read.php?t=168

and more galleries here :

http://www.flickr.com/photos/7881544@N05/
http://www.flickr.com/photos/8010632@N05/
http://www.flickr.com/photos/12312992@N07/page1/

have fun : =))

[DanTheMan]

Here's my latest configuration and I gotta say that I am done for a while with this project.



Does that look a bit better John?   I don't want to calculate another notch filter for a long time only to have it not work out and be forced to do more fiddling and measuring. I still need to improve my last one, but this sounds ridiculously good at this point.

Dan

[doug s.]

hey dan, ever play w/something like a deqx dsp x-over/eq?  you can get amazingly flat response w/these...

doug s.

[DanTheMan]

hey dan, ever play w/something like a deqx dsp x-over/eq?  you can get amazingly flat response w/these...

doug s.

You can't EQ a polar response unless it's the same deviation at every angle.  Making this crossover is sort of like squeezing a baloon as Zilch says.    You can flatten it at one angle, but a bubble will pop up somewhere else.  The only thing a device like that would help me with is the 9 and 12k dips.  I think I may well know what the problem is there, but it isn't offensive so I'm not too worried about it at present.  Right now I just want to spend a lot of time listening to music.  It's been a week since my stereo has been fully operational.

Dan

[nullspace]

Does that look a bit better John? ... this sounds ridiculously good at this point.

Looks great, Dan. The last word is always 'How does it sound?" And, if it sounds 'ridiculously good' then nothing more needs to be done. Nice job and congratulations.

John

[DanTheMan]

Thanks John!  Much appreciated.  Too bad I find it so hard to say "I'm done.  It's good enough." OCD.  But it does sound incredible.  That I'm happy with.  For some reason I want to perfect the stinking graph.  By the sound of it I think it's done, but the ripples in the graph are plain eating me.   :nono:Gotta learn to relax.

Really my next battle has to be good performance for the money subwoofing.

BTW, nice website.   

Dan

[macrojack]

I read somewhere that anyone can paint but it takes an artist to know when the painting is finished. Be an artist - not a technician.

[DanTheMan]

I read somewhere that anyone can paint but it takes an artist to know when the painting is finished. Be an artist - not a technician.

Wise words Mr. Jack.  I think part of what is eating me is b/c I just had surgery and I have to take time off work.  So, I am just lounging around trying not to do anything.  It's hard to do nothing, but that's what I'm doing. 

Surfing the web while listening to music,

Dan

[macrojack]

Dan - I bet the doctor told you to take it easy. Do as you are told. Busy is not necessarily a virtue. It can be a debilitating distraction. We are human beings - not human doings. Meditate!