2xDR250 CNC project

[kamal]

Hey folks!

I've been working on a pair of DR250s, and I'm excited to share my progress.

My overarching goal is to build a sound system that will work reasonably well for medium-sized parties at indoor/outdoor venues. It'll be used to play -a variety of electronic music- dope bass music. I already built a (non BFM) subwoofer that has me covered up to ~100hz, so now I'm filling in the spectrum.

I'm very interested in designing speakers (from scratch), but I'm a novice woodworker. I adapted the plans a bit to lean more heavily on design skills; notably, 3d modeling and CNC milling. Off the bat, I'd like to acknowledge that my deviations from the plans prolly leave me prone to all sorts of dumb problems, so I won't blame the plans if everything crashes and burns/sounds terrible.

Shout out to @eac (Don) for inspiration! I purchased one of his jack10lite flat packs from speakerhardware.

I live in a small shared apartment in San Francisco, so I don't own any large stationary tools. I'm a member of a maker space called theshop.build, which by and large has me covered (though their sawstop table saw has heavy demand/is often out of commission due to misuse). Notably they have a full size (4'x8' bed) shopbot.

Summary of the project so far:
I started by creating a model of the DR250 in fusion360, with parametrically adjustable wood thickness. This took a lot of hours. Since I'm cncing, I added braces/slots/pockets that guarantee everything's square before gluing (and preclude the need for assembly jigs).
Some pics of the model:
https://i.imgur.com/23GDpyV.png
https://i.imgur.com/5yQCSII.png

I inventoried all the parts, exported DXFs of all the flat surfaces, then arranged all the cuts in vcarve, adding dogbone fillets for all the slots.

I sourced quality baltic birch from Macbeath Hardwood in SF, and decided at the last minute to switch from 12mm(~1/2") to 9mm(~3/8") -- thereby committing to reference all angles etc from the fusion360 model instead of the plans. All the 9mm parts can be laid out comfortably on a full 4'x8' sheet.

All the shopbot work (1/8" compression end mill) went reasonably smoothly on the first try, though the slot cutouts needed a bit of finishing (couple seconds with a file).
Piles of cut pieces:
https://i.imgur.com/CGs5ObK.png
https://i.imgur.com/Dh6dTTy.png

Everything comes together pretty smoothly!
Here's an album with a few pics of everything slotting together: https://imgur.com/a/g6lgkRc (no glue yet!!)

I cut the rear reflectors yesterday, so the big remaining tasks are cutting the 1/8" pieces, gluing everything together, and figuring out the whole nacelle situation (I'm tempted to 3d print them, but this would be stupid for a lot of reasons).

Sadly, as a sign of the times, theshop.build's SF location has been shut down by the owner (evidently a developer who's prolly gonna raze the block to build luxury apartments). I've lost access to the shopbot until they relocate, but I'll probably be able to find a friend with a table saw (or I'll cram one into my bedroom ).

Modifying the plans seems to be a pretty touchy subject, so I'd like to reiterate that I'm a novice, I haven't validated *any* of my choices, my standards aren't super high, I don't believe my modifications will necessarily lead to an improvement compared to the result of faithful execution.

That said: with Bill's blessing I'd love to share any of the resources I generated with anybody who's interested.

Thanks for reading, and I'm curious to know what y'all think about this project!

[kamal]

Oh, related: I noticed a 4\ohm version of the deltalite 2510 drivers, but I was unable to find any mention of their use in BFM designs. I decided to use them to match the LF and HF impedances. This'll make it easier to use off the shelf crossovers (though I'm planning to bi-amp + dsp and delay for the compression drivers).

[cafedumonde]

Impressive! I am also embarking on a CNC build, though I'm using Sketchup and Vcarve: 4 Tuba 24 and 1 TAT. It would be very useful if you could summarize the mods that you made for your CNC design.

1. Was there a rule for using the slots and tabs? It looks like you added them for interior parts only.

2. Have you found that the BCD specification for the drivers was accurate for precutting the speaker holes? I have had challenges wit this, which I attribute to Sketchup's awkward handling of circles.

3. The recesses for the hurricane nuts is a great idea, but why didn't you go with the updated approach, which uses wood screws and cleats? This seems a little easier and eliminated the need for an extra panel for the baffle.

4. Did you generate g-code straight from Fusion? I will probably go this way in the future, though Vcarve is great for beginners.

Please share pics as you complete the build!

CDM

[kamal]

1. I certainly don't have a foolproof rule, but here are a couple notes:
- make sure to add dogbones in vcarve, and make sure to recompute the toolpath after adding them
- I made sure to set the width of all the pieces in the model to the actual width of the wood I purchased (measured with calipers) -- this was pretty straightforward in fusion 360 with a model parameter, but I'm not sure how easy it would be in sketchup

2. Yep! I didn't have any issues with this.

3. I suspect the answer is "I wasn't paying attention" That hurricane nuts are better than screws/cleats is certainly an unchecked assumption, but I hope the hurricane nuts will be at least as durable. And I figure that if I misaligned them or they get stripped I can just rotate the driver by 360/8/2 degrees and use fresh wood screws.

4. Nope. I created sketches from projections of the bodies, exported the sketches from fusion360 to .dxf files, then laid them out in vcarve. This was pretty smooth -- afaict the main downside is that it takes a bit of effort to propagate a change in the model through to the g-code (change model -> regenerate profile sketch -> export dxf -> replace part in vcarve). Doing everything in fusion360 seems appealing and I hope to give it a shot next time, but I haven't quite gotten the hang of the built-in layout/cam process.
I haven't used sketchup, but I highly recommend fusion360!

[Grant Bunter]

Oh, related: I noticed a 4\ohm version of the deltalite 2510 drivers, but I was unable to find any mention of their use in BFM designs. I decided to use them to match the LF and HF impedances. This'll make it easier to use off the shelf crossovers (though I'm planning to bi-amp + dsp and delay for the compression drivers).

Nice. I really like how you've approached some aspects of a CNC based build.

A few things. In your first post, you talk about designing and stuff.
It's common for many people coming here to think they can improve the design right off the bat, without asking any questions.
It's kinda an insult to the designer, in that somehow the design is "flawed", so needs improving, without ever having either built, or heard one of the cabs.

Here's where a few things that might have helped:

Ok, the 2510 8ohm is better because;
It has a more pronounced, higher, rising mid range response. This is a critical element in the cab design.
Also, if you expand to 2 cabs per side (paralled), the midrange woofers will only be pulling a 4 ohm load per channel with 8 ohm drivers.
Going to 4 ohm drivers means you need a 2 ohm per channel stable amp, and running it at 2 ohms has greater power draw.

As per the plans, the ASD's are to be wired in series. This increases their power handling capacity, and presents a nominal 16ohms load. You still require the resistor in the plans that prevents HF oscillation, even when biamping.

Amplifiers don't care what load is being presented to them, as long as that load is inside their performance specs.
You don't need to "match" mid woofer and HF, in fact, you can't, because impedance is determined by frequency.

Off the shelf crossovers? Where does it say do that?
NO NO NO
The filters in the plans are specifically for the design(s) in which they can be utilised. They improve response, are tailored to the cabs response, because it's the design that dictates response, not the drivers.
Why build a ferrari, then reduce it's performance?
Build the filters in the plans.

DSP/delay etc.
Many of us don't bother. You won't be able to notice the roughly 0.5msec delay required to address the MF/HF mismatch once applied.
What is much more important is addressing the mismatch between subs and mains with delay (which has to be pre crossover on your DSP or it will alter phase), depending on how they are set up each time, and ensuring that the subs and mains are in phase at the crossover point every time you set up...

[Bill Fitzmaurice]

As per the plans, the ASD's are to be wired in series. This increases their power handling capacity, and presents a nominal 16ohms load. You still require the resistor in the plans that prevents HF oscillation, even when biamping.

The series resistor is only used with piezos. A series capacitor is used when bi-amping compression drivers.

Oh, related: I noticed a 4\ohm version of the deltalite 2510 drivers, but I was unable to find any mention of their use in BFM designs. I decided to use them to match the LF and HF impedances. This'll make it easier to use off the shelf crossovers

Rule #1: Always follow the plans. Rule #2: Never break rule #1. If an off the shelf crossover worked it would be in the plans. If a 4 ohm woofer was a viable option it would be in the plans. If the HF drivers would work better wired parallel and if there was a reason for them to be impedance matched to the woofer it would be in the plans.

[kamal]

Thanks for the feedback! I hope I've been clear in asserting that I don't consider my choices to be improvements over the stock plans.

And thanks for explanations! I have a hard time following instructions if I don't have a thorough understanding of the justifications, and I definitely haven't thought hard enough about the electronics to understand the plans completely. Grant, you've convinced me to switch use the 8ohm woofers.

For biamp eq and limiting, do the behringer inuke/nx dsp amps work reasonably well?

[Bruce Weldy]

Another thing to consider about off the shelf crossovers is that most of them are a 12db slope whereas Bills designs are 18db, allowing a lower crossover point than the speaker specs recommend. This works. I considered bi-amping when I started building these boxes. I'm glad I went with the crossover in the plans. Easier setup and they sound great.

Stick to the plans, use the suggested drivers and build the crossover. Your results will be better than if you try to second guess.

I too, wanted to understand the "why". So, I asked a whole lot of questions until I was satisfied that the people on this forum knew more than I did about certain things. Then I was totally comfortable building by the plans.

After about 10 years and hundreds of shows on these boxes, I have absolutely no regrets that I did it the right way....and not my way.

[Grant Bunter]

That said: with Bill's blessing I'd love to share any of the resources I generated with anybody who's interested.

Mmmm, the code generated to make various cabs via CNC has been (generously) offered up numerous times.
The issue is then: how do you know that person bought plans?

Other than that, did you modify the dimensions to allow for 9mm stock?
And, given everything was cut from an 8 x 4 sheet, why didn't you use 12mm ply?
12mm ply is already much thinner than "convention" when it comes to building speaker cabs.
12mm though, using the plans, allows sufficient bracing for vibration free cabs. Whilst this is a much greater problem in subs, vibration does rob output, and, generally, those that have built cabs on thinner stock here in the forum, have generally increased the bracing.
That's why I'm asking...

[kamal]

Grant,

I'm trying to learn as much as possible from this project, so I'll be totally satisfied if everything turns out poorly in a way that I find informative.

My thoughts about bracing/wood thickness are as follows:
I added some bracing to the internal parts (around all the horn throats, and in the internal region surrounding the woofer) to make sure that everything would assemble square without clamps. Assuming this would add some stiffness, I then decided to reduce the stock thickness.

I considered just reducing the stock thickness for the parts with added bracing, but then decided to do it everywhere, for simplicity.

Wood thickness is a parameter in the model. When you change that parameter, all the parts are regenerated in a way that is at least mechanically consistent (no overlapping parts etc). I actually used 9.3mm, which was the measured thickness of the stock I purchased.

My biggest vibration concern due to the 9mm material comes from the sides, which have the largest <whatever the 2d equivalent of "unsupported span" is>.
Do you have some intuition about scenarios that give rise to vibration?

If one day I have the time, it would be fun to build more cabs from 1/4" and 1/2" stock, then try to measure the vibration with a contact microphone or something. Are there standard ways of measuring vibration?

[ACUA]

Pics please!!! And good luck!!

[Rick Lee]

Looks great! Buying/building a CNC machine is one of my future goals especially since I'm sure there are no maker spaces anywhere close. Concerning the thinner ply, I've built DR200s out of 1/4" and it worked well. I increased bracing as well but would say that extra bracing on the outer side panels would be the most important. No measurements, just observation. I would assume thinner ply would work as well for DR250s also.