[danvprod]

This is a new open baffle build I have been working on for my Wild Burro Betsy 8" full-range drivers. I previously had been using them in 24" x 40" flat baffles with alpha 15 woofers, but I decided I wanted something more modular, and also with a narrower baffle to help with imaging. Plus, I was never really happy with the bass response from the Alphas on the flat baffle.

So I started with an H-frame box for the Eminence Alpha based on MJK's design. 16x16" interior cross section, 7.5" deep on both sides.

The betsy now sits on top with a sloping baffle -- 20.5" high, 16" wide on the bottom and 9" on top.

I did get a chance to make some outdoor measurements with the new baffle, based on @gainphile's great article on measurement: http://gainphile.blogspot.com/2011/01/measuring-open-baffle-loudspeakers.html.

See the REW plots for the outdoor, gated measurement. To me, it looks like the dipole peak is about 5 dB @ 500 Hz. Rolloff begins at about 300 Hz, if I am reading it right.
Some pics of the build.

Next step is to decide on a tweeter:

Maybe ESS Heil Air Motion Transformers?
Maybe Bohlender Graebener Neo3W?
Maybe a back-to-back tweeter arrangement?
Or maybe a super tweeter > 8 kHz, just facing forward?

I'm not sure, I would love some opinions and thoughts on the build. Currently bi-amping and using the MiniDSP for the crossover and filters. System is being bi-amped. Using a Jerry's electronics battery-powered class D amp for the betsys and a Rotel 1050 for the alphas. The Alphas seem to need a lot of power. MiniDSP is doing the EQ and crossing over.

[guest42212]

Looks nice !
Not enough highs from the Betsys ?

[danvprod]

So I actually received the Neo 3 tweeters and mounted them up with a simple 6 dB/octave crossover cap. Long story short -- I think it added a layer of un-needed complexity and not that much additional sonic information to the betsy.

I think the betsy does have enough highs as-is and if I were to keep using the Neo 3s, I'd have to tri-amp my system (which would necessitate the purchase of a second MiniDSP and another stereo amp), or use an L-Pad on the tweeter (which would not be ideal because I plan on powering my main full-range with a SET eventually).

A third option would be what poultry does, and amplify the H-frames with Yung Plate amps -- http://www.amazon.com/Yung-SD300-6-Subwoofer-Amplifier-Module/dp/B0070Z88WU, then use the mini-dsp to crossover the full-range to a tweeter.

So in the meantime, I am going to return the tweeters to parts express and re-assess what I have here. I also am considering using the new MarkAudio A10 gen 3s in a slightly larger baffle -- similar to this design: http://techtalk.parts-express.com/showthread.php?225682-Openminded-An-Open-Back-Speaker-Project

The nice thing, again, about this modular approach is that I can dial in the bass response, finish that and then work on the top of the unit.

For now, they are sounding pretty solid -- I still want to tweak my MiniDSP settings a bit more. My short-term plan is to keep working on the MiniDSP settings and go-ahead and finish my H-frames.

I plan on using some trim pieces to work around the edges and also some half-round trim to brace the corners a bit more. Then they will be finished in black textured paint.

I tweaked my x-over a bit to follow closer to what MJK had suggested with a 12 db/octave crossover at 125 Hz. I need to measure the response again this evening, but this should give me 28 Hz - 300 Hz -3 points.

[studiotech]

Not sure where you crossed the neo3, but you've got a large scoop out there btw 4-9khz. I'd sure want to fill that area in a little better if it was my system.

[jimbones]

Just my experience but i've had better results with just a front firing tweeter. Imaging is much more rock solid. YMMV.

[danvprod]

@StudioTech -- yes I noticed that big scoop. Not a big fan of that, and I don't want to really EQ it, because of the off-axis implication. That was kind of my thought of moving to the Mark Audio 10A gen3  http://www.madisoundspeakerstore.com/approx-6-fullrange/markaudio-alpair-10-grey-6-full-range-gen-2/

Could it be something with my baffle dimensions that is cause that dip?
If you look at the frequency response graph on Wild Burro's site, there is a bit of that dip, but it is not nearly as deep as what I have measured:
http://www.wildburroaudio.com/speakers.php

I was crossing the neo3 with a 4 uF cap, so it was coming in at 10k or so. Maybe the better idea would have been use the betsy up to 3k or so and then cross to the tweeter. Either way I would have needed a more complex passive x-over or a bi-amp setup.

@jimbones -- thanks for the insight. Do you have any recommendations for tweeters that you have used?

[Tyson]

Yeah, 10k is way too high.  Try for 3k or 4k.  Even at that lower point, most of the musical info occurs below it. 

[nickd]

The neo cominig lower will help. You will not be able to return them after this, but cutting off the back cup makes a HUGE difference. Go open baffle all the way. BG Radia uses the tweeter with a 12db filter just above 1500hz in a wave guide on the front and open to the back.

They sound really good in that application.

[danvprod]

Thank you, @nickd --

lower the x-over would be an option -- another option might be just switching the betsy drivers to the Mark audio drivers, and mounting them in a wider baffle to lower the dipole roll off.

Something like this --

Either slightly off-set or in the center of a flat u-frame baffle 3" deep, 17.5" wide and 20.5" tall.

[danvprod]

Interesting, the smaller Alpair 7A has a qts of .54.

This may be a good option too?
http://www.madisoundspeakerstore.com/approx-4-fullrange/markaudio-alpair-7-gen3-grey-cone-4-full-range/

Especially because I don't need the LF response really -- the Alpha handles up to 200 Hz easily. This might give me a bit more from 200 Hz - 20,000 Hz.
Just a thought -- this is more similar to MJK's original design with the Fostex FE103E --

which also may be a good idea:
http://www.madisoundspeakerstore.com/approx-4-fullrange/fostex-fe103en-4-full-range/

I'm not sure...

[studiotech]

Don't mistake a more complex filter to be worse than simple.  You use just as much filtering as necessary to hit the desired target.  IMHO, single drivers run too high have far more sonic limitations than using a real crossover handed off to a proper tweeter.

Greg

[JohnR]

Could it be something with my baffle dimensions that is cause that dip?
If you look at the frequency response graph on Wild Burro's site, there is a bit of that dip, but it is not nearly as deep as what I have measured:
http://www.wildburroaudio.com/speakers.php

Hi, I suspect it's because you are not measuring on axis. That's how it seems anyway - your photo seems to show the mic raised up but still not on axis - ?

Edge shows that the dipole peak with that baffle should be a little over 1 kHz. You can sort of see that in your curve. No idea what's happening at 500  Perhaps it might be worth trying a couple other orientations/locations for measurement to see if it's consistent.

I'd also suggest that for high frequency measurements you point the mic straight at the driver - assuming you have a calibration file for it.

PS. before switching drivers, once you can get consistent measurements you might like to try a couple of notches in the miniDSP to even out the response. From the curve on Paul's page 3.5k and 10k would do it.

[danvprod]

Thank you for the thoughts...

I see what you mean about the response in Edge. I've attached the simulation. I will re-try my measurements based on your suggestions to see if I can measure that peak accurately.

I don't know what is going on with 500 Hz either...

The rolloff does look reasonably similar to what I have measured.

[danvprod]

A couple of other ideas:

Maybe I could use one of madisound's two-way kits for the top of the box. I really like what this builder did here:
https://sites.google.com/site/undefinition/diy-sunflowers

He uses two of these 5" woofers -- http://www.parts-express.com/pe/showdetl.cfm?partnumber=295-370 and one of the Vifa tweeters -- http://www.parts-express.com/pe/showdetl.cfm?partnumber=264-1016

Total cost would be $186 for both sides for that system. I could basically build the top part of the sunflower on top of the H-frame and then use his passive x-over design. The dimensions of the sunflower are pretty close to that of my H-frame. I'd could build the mid and high crossover. I wouldn't necessarily need to put together the notch filter (and because my dimensions would be slightly different, it would work anyways).

Other thought would be to use the pre-made Zaph kit:
http://www.madisoundspeakerstore.com/2-way-speaker-kits/zaph|audio-za5.2-tm-2-way-pair/

Or maybe this Seas kit:
http://www.madisoundspeakerstore.com/2-way-speaker-kits/seas-idunn-2-way-speaker-kit-pair-parts-only/

Anyways I appreciate you all brainstorming with me and my warm welcome to this forum. Cheers!
Dan

[JohnR]

A couple of other ideas:

You might be shooting yourself in the foot a bit there. With the box kits the crossover probably includes some baffle step compensation which won't match your baffle, so you'll be eq'ing it back to something else and still losing efficiency. With the OB design you might be able to use it as-is...  although it does look like you'll be losing some efficiency again because the crossover seems to be designed to drop the mid/tweet down to the efficiency of the woofer.

I'm not sure why you don't just go active.

[danvprod]

Yea, active may be the way to go, it just adds a layer of complexity -- I'll have to get a second MiniDSP, more cables and a third amp.

With the Zaph kit -- @Gainphile has had good luck with his S20z project -- http://gainphile.blogspot.com.au/2012/06/s20-z-compact-open-baffle-using-zaph.html

That was kind of my thought with using the Mark Audio driver, no more complexity -- Just mount the driver on a new baffle, upgrade to a SET amp and call it a day.

[JohnR]

There have been some passive designs talked about here as well - look up a level and find Hestia and Manzanita, those seemed to be well thought out designs.

[Tyson]

Sell the miniDSP and get the miniDSP 4x10 HD.  Then you can do it all in one box, plus gain 24/96 playback.

[danvprod]

You are absolutely right -- that would solve my problem.

[Poultrygeist]

I've switched to an Ashly XR-1001 which does a remarkable job while remaining invisible. I know the digital ones are all the rage but this analog unit gets it done for me.

The Neo3's are my favorite tweeters but their lower efficiency hasn't proven to be as great of a match with the highly efficient Beta12lta as I orginally thought. It's much better with the TB W8-1808.

With the TB OB's I drive the Neo's with a t-amp and the TB's with a SET crossing over with the active and using the Yungs to power and crossover the Alphas.

But I spend most of my time listening to the Beta 12LTA tied to a Realistic 40-1310A ( Japanese made Fostex  ) tweeter with a single .47uf Sonicap. The H-frame Alphas ( underneath the Betas ) I drive with the Yung plate with it's crossover maxed and with the Ashly set at 1.50K.

I have enjoyed the Betsy's over the years but they lag behind the phase plug Betas in every category.

I have the 40-1310A ( Japan ), B ( Korea ) and C ( China ). The A model has very different contruction than the other two and it's the only one I'm sold on. I believe it to be the true Fostex tweeter.