"Music Room" Open Baffle Project

[ErinH]

I’ve been reading up on OB projects and works/literature by MJK, SL’s info on his Orion setup, and using programs such as Edge to generate baffle response curves, among reading many other posts/threads here. Here’s a few links I’ve gathered in case anyone is ever searching and comes across my thread:
http://www.quarter-wave.com/OBs/OB_Theory.html
http://www.linkwitzlab.com/orion_challenge.htm
http://fullrangedriver.com/singledriver/baffle.html
http://www.tolvan.com/edge/
http://www.troelsgravesen.dk/OBS.htm


Right now, my main concern is in regards to the baffle shape itself. I’ve been reading up on h-frame and standard OB construction. H-frame actually is getting the nod due to it’s aesthetics combined with the boosted low end. However, since this room is upstairs and dedicated to my music, I can make the baffle as large as I like. I have a test sheet of ¾” MDF with dimensions of 5x4ft sitting in my garage right now. After modeling some simulations, I’m considering putting a hole in this for my sub and seeing what the response looks like. I’d like to make about 3-4 test baffles to see where the point of diminishing returns is, and I may wind up having to go this far.
In addition to the sub, I want the TB 8” to get to about 200-250hz and let it play the rest. As it stands, I have it crossed at 300hz 6dB/oct and it sounds pretty darn nice. I hate to put it in such a large baffle, so I may even go with an asymmetrical h-frame (a la the Orion setup). Needless to say, I have quite a bit of testing to do.

I’ve modeled the response for the low end with Edge and came up with the following options, which have been normalized, and also included the graph which corresponds with some of the models:

The 1600x1600mm results are a ‘best case’ scenario, which is a 63x63” baffle. Very friggin’ large. Though, it helps me to know what I could achieve if I needed to, as I do have this kind of room in the listening room. It seems that 1200x1200mm (~48x48”) will lend good response allowing only a drop of 13dB from 200hz to 30hz, where a 1000x1000mm baffle (40x40”) gives a drop of 14.5dB from 200 to 30hz (see chart and figure below). 800x800mm (31x31”) is roughly 3dB down from the best case scenario at 30hz, with an overall drop of 15.4dB, compared to 12dB for the best case 1600mm^2 baffle. This might be a good ‘stopping point’ as the rule of double power for 3dB is starting to weigh in here which is telling me I could be running out of gas at this point.
With that said, I’ll likely build a 3ftx3ft baffle, place the driver in the center (as predicted for nominal output) and measure it as it is.
According to MJK’s quarter wavelength site, I can use a side piece of ~ 17 inches before I get to the quarter length of 200hz (some fudging here, but alas) and run into reflective issues with the sub. So, I could build a side panel and toy with that to see how the response changes.
*by the way, does anyone happen to know if there’s a way to model this up without sacrificing a lot of wood on the experiment?*

 

[ErinH]

Here's a few pictures of today's build. The idea was to build a fairly large baffle and then start adding pieces to the picture to see how it shapes the response.
The size is 32x48" (my cheesy self is in the picture for reference).

For this purpose, I'm using the AE IB15 just to get an idea of how things will work. In fact, I may just wind up purchasing the IB15 instead of the OB15 due to its better low end response since I'm bi-amping it all anyway.

I'll start doing some experimenting tomorrow with things and try to post results Friday.

 

[ErinH]

anyone out there???....
if I'm wasting my time typing this stuff and posting here I think I'll stop. LMK if anyone is watching because I won't keep wasting my time if not.

..........................


After spending the last 30 minutes modeling on Edge, the one thing I’ve found is that as long as the baffle is of decent size in some dimension, the low end response isn’t hugely effected.
Take the below two pictures for example. Starting with the outer dimensions of 31.5x48 inches, I’ve drawn a standard square (picture 2 & 3) and a more complex shape (picture 1).


I’ve shown the response of the woofer location (bottom) and the fullrange at the top.
***Note: In the simulation you can only choose a single size for modeling multiple woofers. I’ve verified that modeling the location/size of the woofer as a smaller 8 inch driver is essentially the same below 200hz. Therefore, I’ve modeled the 15 in driver as an 8 inch so that I can model both drivers together. If you’re going to do the same, I ask that you verify the smaller size driver will effectively simulate the larger size up to the point at which you plan to cross it over. If I were crossing higher than 200hz, this would not have worked well at all.***

Note the difference in response at 30hz for both baffle shapes is not terribly out of sync, whereas the hump at 200hz is brought down a bit with the ‘coffin’ design. I think I almost prefer this design as it lends itself to a good low end response, without a peaky top end, which serves me no purpose as I’m crossing quite a bit below 200hz anyway.
I’ve also removed the data above 2000hz as it’s essentially the same in the grand scheme of things.

The pictures below correspond with the graphs shown overlayed.
Curve 1, Picture 1 = Coffin Shape
Curve 2, Picture 2 = Rectangle with offset full range at the top
Curve 3, Picture 3 = Current Rectangular design (the piece I made yesterday)

The coffin shape was chosen somewhat arbitrarily to show the differences in a complex shape having smaller woofer-to-edge parameters. The fullrange driver offset in the coffin shape is a ‘best case’, which saves me from cluttering up the thread with needless designs.
At 30hz, the coffin shape is about 2dB off my current rectangular shape. At 200hz the difference is 1.5dB.

So, what does this mean? It means, simply, that the large distance between the edge of the baffle and the woofer is important, though not as immensely as I had though. Also, it shows that a more subtle phase transition is attainable with a design like this as it allows me to make the fullrange driver’s baffle area smaller, while still giving large surface area to the woofer. Is this the final design I’m going with? Heck no. The coffin is ugly. But, for illustration purposes, it serves well. Also, I can base further designs off this, knowing I’m able to live with this kind of compromise.

I’m doing a lot of this work so that the design is at least thought out to some degree. However, the benefit of the DSP will likely wind up playing more of a role than the design itself in regards to the low end response. Sometimes I feel like I may be putting the trailer before the horse, given my potential use of DSP.

I’m curious to see how my modeled results in picture #3 compare to the actual results of this design I built yesterday.

Aaaaannndddd….
FTMFW!
MathCad Computer Models : Upgraded Versions

$25 fee, but access to some AWESOME mathcad programs. Very detailed, too, if you check out his sample worksheets (ie: Full Range + Woofer OB worksheet). I use mathcad regularly at work so I'm already familiar with the program.

I'll have to sell some things, lol... me has no monies.

 

[dkane360]

I'm following the thread

I don't really have much to say about it, but I find it interesting. Building something like this popped into my head a while ago because I didnt feel like building the whole enclosure lol. Now it seems a little out of my league But I would like to comment on the 4 car garage that your neighbor has

Oh and if you need to sell some things, I'll take an older receiver off your hands if you have one. My friends brother is just getting into this obsession and needs a starter receiver.

 

[ErinH]

But I would like to comment on the 4 car garage that your neighbor has

and the best part?
He has 3 corvettes in there:
65 vette
c5
c6 z06 (grey... beautiful)

His wife drives a toyota.

furthermore, he's retired and she still works.

ahhhh, livin the dream.

 

[STRONGBADF1]

I'm following...barely It's cool that you are experimenting with this. Keep it up as long as you're having a good time doing it.

 

[adwilk]

I'm following- but most (read "all") is over my head so I'm staying out of it. I'll start posting in here but now that Doug is here too, expect this thread to get derailed quick fast.

Boobies!

 

[ErinH]

I did some testing last night.
Below you’ll see RTA (smoothed at 1/3 octave) for the tang band by itself and the AE IB15 by itself. However, for the AE, I’ve shown 3 different designs: standard OB, u-frame, and h-frame.
The ‘frame’ was simply 2 sheets of 41”x28” mdf I had laying around. With the H-frame, I laid the piece on its side and positioned it so there was ~15-16” on either side of the baffle. For the u-frame, I stood the pieces up so the wings were 41” tall and 28” wide. It’s not an exact science, but it’s close. I mainly wanted to see what I was dealing with.

These measurements are not gated. They were generated via an mls sweep and then smoothed. I’m farily new to holmImpulse so didn’t really try to get into windowing. I just ran the sweep and saved the results. I’ll keep plugging away at it, but wanted to share my early results.

TB w8-1772 response:
Zoomed out:

Zoomed in:

At approximately 225hz, the effect of the baffle takes place and the rolloff slope is 18dB/oct. If you look at the two graphs and piece them together, you’ll see that overall, the response really isn’t too bad. But, you can clearly see the rising response centered at about 10khz, which explains the top end brightness I spoke of before.


AE IB15 response:
Blue = OB
Red = U-Frame
Green = H-Frame

You can see at 45hz, the subs take a nosedive at approximately -15dB/octave. All three responses show this, so it’s likely the baffle influence. I’m actually pretty happy that I managed to get this low before the rolloff kicked in, though.
The u-frame design seems to look the best, given that it has a lower amplitude above 100hz (better for blending) and a decent response on the low end relative to the other two designs.



I’ll go back and study up on the holmimpulse tutorial tonight and walk through that to make sure all my settings are correct and maybe even window the response. If things measure drastically different, I’ll post the results back up. I also have a few mathcad sheets to play around with for further simulations.

Got some car stuff to get done this weekend, so probably won’t be able to pick back up until next week.

Hope everyone’s having fun.

- Erin

 

[annunaki]

Erin,

You should have started this in the DIY section if you wanted more responses . Let me know if you want it moved.

Very interesting design & experimentation. Lot's of work involved here. More than I would ever have time for so kudos to you!

 

[highfigh]

anyone out there???....
if I'm wasting my time typing this stuff and posting here I think I'll stop. LMK if anyone is watching because I won't keep wasting my time if not.

..........................


After spending the last 30 minutes modeling on Edge, the one thing I’ve found is that as long as the baffle is of decent size in some dimension, the low end response isn’t hugely effected.
Take the below two pictures for example. Starting with the outer dimensions of 31.5x48 inches, I’ve drawn a standard square (picture 2 & 3) and a more complex shape (picture 1).


I’ve shown the response of the woofer location (bottom) and the fullrange at the top.
***Note: In the simulation you can only choose a single size for modeling multiple woofers. I’ve verified that modeling the location/size of the woofer as a smaller 8 inch driver is essentially the same below 200hz. Therefore, I’ve modeled the 15 in driver as an 8 inch so that I can model both drivers together. If you’re going to do the same, I ask that you verify the smaller size driver will effectively simulate the larger size up to the point at which you plan to cross it over. If I were crossing higher than 200hz, this would not have worked well at all.***

Note the difference in response at 30hz for both baffle shapes is not terribly out of sync, whereas the hump at 200hz is brought down a bit with the ‘coffin’ design. I think I almost prefer this design as it lends itself to a good low end response, without a peaky top end, which serves me no purpose as I’m crossing quite a bit below 200hz anyway.
I’ve also removed the data above 2000hz as it’s essentially the same in the grand scheme of things.

The pictures below correspond with the graphs shown overlayed.
Curve 1, Picture 1 = Coffin Shape
Curve 2, Picture 2 = Rectangle with offset full range at the top
Curve 3, Picture 3 = Current Rectangular design (the piece I made yesterday)

The coffin shape was chosen somewhat arbitrarily to show the differences in a complex shape having smaller woofer-to-edge parameters. The fullrange driver offset in the coffin shape is a ‘best case’, which saves me from cluttering up the thread with needless designs.
At 30hz, the coffin shape is about 2dB off my current rectangular shape. At 200hz the difference is 1.5dB.

So, what does this mean? It means, simply, that the large distance between the edge of the baffle and the woofer is important, though not as immensely as I had though. Also, it shows that a more subtle phase transition is attainable with a design like this as it allows me to make the fullrange driver’s baffle area smaller, while still giving large surface area to the woofer. Is this the final design I’m going with? Heck no. The coffin is ugly. But, for illustration purposes, it serves well. Also, I can base further designs off this, knowing I’m able to live with this kind of compromise.

I’m doing a lot of this work so that the design is at least thought out to some degree. However, the benefit of the DSP will likely wind up playing more of a role than the design itself in regards to the low end response. Sometimes I feel like I may be putting the trailer before the horse, given my potential use of DSP.

I’m curious to see how my modeled results in picture #3 compare to the actual results of this design I built yesterday.

Aaaaannndddd….
FTMFW!
MathCad Computer Models : Upgraded Versions

$25 fee, but access to some AWESOME mathcad programs. Very detailed, too, if you check out his sample worksheets (ie: Full Range + Woofer OB worksheet). I use mathcad regularly at work so I'm already familiar with the program.

I'll have to sell some things, lol... me has no monies.

Keep posting. You need to remember that we don't live here, we come here when we have time.

 

[90gstman]

Keep posting, I'm interested. I agree with annunaki, the reason your not getting many posts is because the dyi guys are regularly view the dyi section.

Anyway, I did add some 5" wings to the 10" wide baffle I was using for my 7" Seas midrange and that made a world of difference. I'm now able to run it down to 200 hz before crosssing over to the two 10" Daytons.

After seeing your positive results with the H and U frames, I'm going to try unboxing the bottom end. The one thing thats bothering me is that it's so easy to tell where the transition from open baffle to vented box takes place. I think going open baffle on the bottome end will make the whole system more coherent. The Daytons Qtc may be to low though, but I give it a try. I probably won't get to this until July though. I'll try posting measurments.

Keep us updated on your progress.

 

[highfigh]

Erin- have you tested for differences when the backside of the speaker hole is beveled or rounded over? Also, are you planning to try curved wings, or just a flat baffle?

 

[ErinH]

Erin- have you tested for differences when the backside of the speaker hole is beveled or rounded over? Also, are you planning to try curved wings, or just a flat baffle?

Not yet. I plan to.
Troels has some info on his site, too. He didn’t obtain measurable results, but said he could still hear a difference. I wonder why. I don’t believe in magic fairy dust so I expect to see it. We can hear a difference of 0.5dB. That’s not much. That should easily show up on a measurement. Not saying he’s full of it… I just wonder what translates into the noticeable difference and why his results don’t show one.


I plan to try quite a few different things. Project is stopped right now as I put my IB15’s back in my after obtaining the JBL MS-8 and wrote a review for it (can be found on DIYMA here: http://www.diymobileaudio.com/forum/member-product-reviews/83066-jbls-ms-8-a-2.html#post1055836).

Hopefully I’ll get started back up within the month, but we’ll see.


Thanks guys.

 

[GO-NAD!]

I've been following your progress and it's quite interesting. I haven't commented before, because I pretty clueless about speaker design - any design. Woodwork is my thing and I have a mechanical background, so if somebody gives me a plan, or even simple directions, I can put it together.

I've been intrigued by the OB concept ever since discovering SL's website. I'd love an Orion setup, but they would require more space than I can provide - which is moot anyway, since head office would never approve...

Keep it up and good luck! By the way, do you always do woodwork in beach wear?