BMR Philharmonitor with Curved Cabinets

[DeruDog]

Hi All. I received the BMR Philharmonitor kits from Meniscus last week. I am planning to do some custom Baltic Birch Cabinets on my CNC with curved sides. Purely for aesthetic reasons and because I love designing things. My first renders are here, and I will keep the post rolling as I move along with the build. My first cabinet build with curved sides, first 3-way speakers, and first crossover build. With some things, if it's worth doing, its worth overdoing, I guess.

Front baffle has increased radius on the side round-overs. Otherwise the baffle is unchanged. The interior volume is virtually unchanged, with less than 1% lower volume.

 

[DeruDog]

The first critical step after the above design was checking and fitting the BMR cutout with the CNC. Hand router was one option, but I enjoy problem solving, so ... I made a CNC cutout design. First attempt (by eye and measurement) had some unattractive gaps; the last two photos are the successful cutout (by tracing a photo of the BMR). It's an even 1mm around the whole thing. Setting up the cut probably took about 90 minutes, including taking a photo and a little head-scratching about the best approach.

Test complete, and on to the final cuts in the coming days.

 

[headshake]

Any chamfering on the inside?

(pic credit http://www.troelsgravesen.dk/)

 

[DeruDog]

Any chamfering on the inside?

(pic credit http://www.troelsgravesen.dk/)

No call for it in the design so I had not planned on it.

I will look in to it and consider #4, since the face is an inch thick. That said, this would apply mostly to the BMR, since the Woofer is way deeper than the diagrams. The BMR is in a pretty small sealed compartment, so I don't think it behaves the same way as a normal speaker and hence might not behave the same way with a chamfered interior baffle.

 

[headshake]

No call for it in the design so I had not planned on it.

If you are curving the cab, won't that change the crossover?

If I add an edge radius in virtuixCAD I see changes around 1k-2k.

"The BMR is in a pretty small sealed compartment, so I don't think it behaves the same way as a normal speaker "
Oh, I see, the bmr on its own is all sealed up + the compartment to seal it inside the speaker box.

 

[DeruDog]

If you are curving the cab, won't that change the crossover?

If I add an edge radius in virtuixCAD I see changes around 1k-2k.

I am only considering an increase in the edge roundover. The original plans called for a roundover as well, of about .5in radius, which I increased to 1in radius. Dennis thinks it should work out alright, with the same crossover. If I have problems, I could build another cabinet or just allow my DSP to take care of it.

I will measure response with a calibrated microphone and REW when I am done. I am actually curious how much I might be messing up the tuning with the increased side bevel. Could be smoother, could be worse ...

I have a question, though, since you have virtiuxCAD: When increasing the roundover, if I am trying to keep the same tuning would I keep the flat part of the Baffle the same, or the radius-to-radius measurement the same, or the outside-to-outside of the cabinet the same? Logic dictates to me that I would keep the flat area the same and the additional radius would make the edge diffraction just roll off a bit more slowly, but I wonder what the modeling would tell us.

 

[bigjacko]

I don't know how to corretly measure the baffle size when it has round off, but I know there are software that can simulate the diffraction. There are complicated softwares to simple ones that use excel. For example this one: http://audio.claub.net/software/jbabgy/BDBS.html .

 

[R Swerdlow]

I am only considering an increase in the edge roundover. The original plans called for a roundover as well, of about .5in radius, which I increased to 1in radius. Dennis thinks it should work out alright, with the same crossover. If I have problems, I could build another cabinet or just allow my DSP to take care of it.

A number of years ago Dennis directly measured the same 2-way speakers, his MBOW1 design, in cabinets with squared off edges and with edges rounded over with a ¾" radius. (The MBOW1 speaker cabinets were 8" wide, as shown here.) The squared edge cabinet had a small diffraction peak at 4 kHz and a small dip at 7 kHz. The rounded edge cabinet made both nearly disappear.

Here is the square edged cabinet, on-axis:

And the rounded edge cabinet, on-axis:

A similar comparison made with the microphone 15° off-axis, with squared edged cabinet:

And with rounded edges, 15° off-axis:

I will measure response with a calibrated microphone and REW when I am done. I am actually curious how much I might be messing up the tuning with the increased side bevel. Could be smoother, could be worse ...

I have a question, though, since you have virtiuxCAD: When increasing the roundover, if I am trying to keep the same tuning would I keep the flat part of the Baffle the same, or the radius-to-radius measurement the same, or the outside-to-outside of the cabinet the same? Logic dictates to me that I would keep the flat area the same and the additional radius would make the edge diffraction just roll off a bit more slowly, but I wonder what the modeling would tell us.

I personally doubt if your cabinet design will require major crossover changes, but there's no better way to find out than to build a test cabinet and measure it. It might be an improvement.

 

[DeruDog]

I don't know how to corretly measure the baffle size when it has round off, but I know there are software that can simulate the diffraction. There are complicated softwares to simple ones that use excel. For example this one: http://audio.claub.net/software/jbabgy/BDBS.html .

So Cool! I am an excel nerd, so I love this free charting for speaker design.

I used the jbagby chart you linked and I got the below results for the ribbon. The first is my design vs the original design, my design in blue:

The next is if I narrow my planned box so the outside dimensions match the original:

Much closer to the original curve! Now I need to see how much I can get away with on roundover, since the woofer is close to the same width as the original baffle. Also, I did the same experiment with the BMR and came to the same conclusion on the curve, the narrower baffle is much closer. So thanks a million for that link, it is just what I was looking for when I originally started this design process. Back to the virtual drawing board, and the speaker is about to get a little deeper.

 

[DeruDog]

Sadly, there is only room for the 1/2 inch roundover due to the size of the woofer. I could do a pyramid-shaped speaker, or a wider roundover on the top half of the cabinet, but that strays from the original design, and would not be aesthetically as nice. So back to the 1/2 inch roundover it is. I am planning on a skinny version I made last night,

vs. a fatter version I designed earlier yesterday

 

[bigjacko]

Can you make the side even fatter and blend into the 1/2 inch roundover? It might be hard to make the side panels but seems a fun experiment.

 

[DeruDog]

I could do that but it will affect the crossover a little. It would be hard to know how much. That said, my AVR has DSP which would correct many problems. I would prefer not to rely on that, though.

 

[DeruDog]

Can you make the side even fatter and blend into the 1/2 inch roundover? It might be hard to make the side panels but seems a fun experiment.

See above

 

[DeruDog]

Can you make the side even fatter and blend into the 1/2 inch roundover? It might be hard to make the side panels but seems a fun experiment.

Also, the roundover is actually blended in to the side, one at 175 degrees and one at 165 degrees, if I remember right, so it should look quite smooth.

 

[DeruDog]

First test cuts yesterday. All proceeding to plan.

 

[Pygmy]

Following this, love to see what you're building.

I have a long term DIY project going based on the idea of the Philharmonic BMR's, I'm trying to design my own 3-ways using a BMR as mid, an Isodynamic Planar Tweeter, and a (sub) woofer driver.

I'll be doing an "active" build - the speaker drivers will be connected straight to outputs on the speaker, and I'll be building one "amp block" per speaker. Each "Amp block" will have a DSP, and 3 class D amplifiers to power each individual driver.
At least, that's the idea - it'll probably take a year or two to get it working..

 

[DeruDog]

I thought about the same thing. I think the Hypex is the easy button.

 

[Pygmy]

I thought about the same thing. I think the Hypex is the easy button.

If you mean their plate amps, sure.
But I want the amps externally, so I can use them later for different speakers, just having to change the DSP settings.

And I don't feel like making separate boxes with just a plate amp...

 

[DeruDog]

If you mean their plate amps, sure.
But I want the amps externally, so I can use them later for different speakers, just having to change the DSP settings.

And I don't feel like making separate boxes with just a plate amp...

Yes, I was talking about the plate amps, the Hypex 123, or a hybrid 122 with the woofer on one channel and the mid/tweeter on the other. I am uncertain about how they would test, so I think your idea of an external DSP solution would be ideal. Would you build that yourself, or just have 2-4 separate "boxes" for each speaker? Have you found the best DSP solution, yet? And Amps?

Again, I had thought about this before and may go there in the future before all is said and done.

 

[DeruDog]

Daily update. The 1/4 inch guide rods fit with a little coaxing from my mallet. The rest of the day was cutting and fitting a spoil board. This is a new cut for my CNC.