FUNDAMENT - an open-source, fully-documented subwoofer platform (Build 001: Sealed 18″, BMS 18N862 + FA502)

[fivenp]

Hey there.... long-time lurker, but never really dared to write

I'd love to start an initiative on how we approach DIY Designs – mainly because of my own need for now

A little bit on the Background
Good subs are weirdly hard to get hold of in a lot of the world. E.g. Funk and Rythmik are the obvious ones and they're excellent - but shipping them across the oceans, plus duty, plus VAT, plus no local warranty, makes them a non-starter for a lot of people. Meanwhile there are brilliant DIY sealed builds scattered across forums that measure seriously well… and then get buried three pages deep and forgotten.

So the idea is simple: I'd love to take well known, working recipes, re-build them properly, in the open while having you as the community collaborate. Not another buried build — a documented, measurable, repairable, reproducible platform anyone can build locally, fix locally, and participate in improving going forward. (or potentially branch off and create new meaningful projects with other targets/budgets/equipment)

What I mean by open is really - having everything well documented, centralised and accessible / ready to be iterated on. Open enclosure plans, DSP presets, published measurements AND a consistent protocol, revision histories, accessible replacement parts, no black box - ever....

Down the line:
For folks who'd rather not cut cabinets and wire amps, We might team up as a community and offer built-to-order units - with reasonable prices. But the core stays open forever: plans, presets, measurements, HFD all public. BTO should be around convenience or skills, not a paywall around the knowledge!

Where I'm now

The first project I'm aiming for is the S18 - a sealed 18" "reference" level Subwoofer (afaik there are a lot of builds already out there in a similar direction)

• Driver: BMS 18N862 – should give clean and high output
• Amp/DSP: Hypex FA502, bridged.
• Enclosure: sealed, ~115–125 L net → Fc ≈ 47 Hz, Qtc ≈ 0.64.
• Three Hypex presets for starters, trading extension vs max SPL: P1 deeper/quieter · P2 balanced · P3 louder/shallower.

I personally need 2 of them to complement a full stereo set and be crossed at 120Hz (or potentially even higher?) - but I'm not 100% sure yet around beaming issues - I think measurements will show. But would love to hear your thoughts.

I've attached the first simulations of the presets in 125L Net

Simulations
https://simulator.00aud.io/app/share/18N862-125L-sealed-cmqpb8tnc0001jp04i3ffsmv2

It seems like we'd be running in pretty "okayish" territory regarding xmax + spl + amp power?
Goal was to have enough headroom for all type of home relevant applications.

Feel free to roast, give me reality-checks, criticism, suggestions - it's all very very welcome

I hope you all see some value and have some inputs. Let's build something nice together.
Michael

--------------
Files & Archives
The Open Sourced files will live in the GitHub repository over here - where I'll try to keep things as up to date and open as possible:
https://github.com/fundament-audio/s18

 

[Luvchampagne]

Hey there.... long-time lurker, but never really dared to write

I'd love to start an initiative on how we approach DIY Designs – mainly because of my own need for now

A little bit on the Background
Good subs are weirdly hard to get hold of in a lot of the world. E.g. Funk and Rythmik are the obvious ones and they're excellent - but shipping them across the oceans, plus duty, plus VAT, plus no local warranty, makes them a non-starter for a lot of people. Meanwhile there are brilliant DIY sealed builds scattered across forums that measure seriously well… and then get buried three pages deep and forgotten.

So the idea is simple: I'd love to take well known, working recipes, re-build them properly, in the open while having you as the community collaborate. Not another buried build — a documented, measurable, repairable, reproducible platform anyone can build locally, fix locally, and participate in improving going forward. (or potentially branch off and create new meaningful projects with other targets/budgets/equipment)

What I mean by open is really - having everything well documented, centralised and accessible / ready to be iterated on. Open enclosure plans, DSP presets, published measurements AND a consistent protocol, revision histories, accessible replacement parts, no black box - ever....

Down the line:
For folks who'd rather not cut cabinets and wire amps, We might team up as a community and offer built-to-order units - with reasonable prices. But the core stays open forever: plans, presets, measurements, HFD all public. BTO should be around convenience or skills, not a paywall around the knowledge!

Where I'm now

The first project I'm aiming for is the S18 - a sealed 18" "reference" level Subwoofer (afaik there are a lot of builds already out there in a similar direction)

• Driver: BMS 18N862 – should give clean and high output
• Amp/DSP: Hypex FA502, bridged.
• Enclosure: sealed, ~115–125 L net → Fc ≈ 47 Hz, Qtc ≈ 0.64.
• Three Hypex presets for starters, trading extension vs max SPL: P1 deeper/quieter · P2 balanced · P3 louder/shallower.

I personally need 2 of them to complement a full stereo set and be crossed at 120Hz (or potentially even higher?) - but I'm not 100% sure yet around beaming issues - I think measurements will show. But would love to hear your thoughts.

I've attached the first simulations of the presets in 125L Net

Simulations
https://simulator.00aud.io/app/share/18N862-125L-sealed-cmqpb8tnc0001jp04i3ffsmv2

It seems like we'd be running in pretty "okayish" territory regarding xmax + spl + amp power?
Goal was to have enough headroom for all type of home relevant applications.

Feel free to roast, give me reality-checks, criticism, suggestions - it's all very very welcome

I hope you all see some value and have some inputs. Let's build something nice together.
Michael

--------------
Files & Archives
The Open Sourced files will live in the GitHub repository over here - where I'll try to keep things as up to date and open as possible:
https://github.com/fundament-audio/s18

"
So the idea is simple: I'd love to take well known, working recipes, re-build them properly, in the open while having you as the community collaborate. Not another buried build — a documented, measurable, repairable, reproducible platform anyone can build locally, fix locally, and participate in improving going forward. (or potentially branch off and create new meaningful projects with other targets/budgets/equipment)

What I mean by open is really - having everything well documented, centralised and accessible / ready to be iterated on. Open enclosure plans, DSP presets, published measurements AND a consistent protocol, revision histories, accessible replacement parts, no black box - ever...."

I went through a very similar situation last year.

1. Your choice of architecture /amp/dsp
"repairable, reproducible platform anyone can build locally, fix locally,"
1. #1 failure point of a subwoofer is the plate amp. How do you repair a hypex plate amp? Most stereo repair shops won't touch it. I would highly recommend AGAINST this.
2. Why combo the amp and dsp together?
Get a minidsp 2x4hd or any other cheaper dsp solution and you're done with dsp. The more technical knowledge and the more a person wants to roll up their sleeves, the cheaper they can go. Some solutions are remarkably cheap but difficult to work with. Plug in the crossovers and filters and you're done
3. Agreed on box and driver need to be be set so then can the dsp settings be fixed. If a few alternatives to the driver are offered then that makes it more affordable depending on location and price range.
4. So many cheap amps available for subwoofers. Everything from a fosi/chip amp to an old class AB amp or a cheap class d amp. Again, why lock yourself in to a pricey Hypex plate amp?
5. Hypex amps are far from well documented. And far from easy to program.
Common theme - big mistake to design an open source sub build with an off the shelf plate amp like a hypex. Any other plate amp has the same problems. And when thst plate amp fails, good luck finding the exact same footprint to replace it.

I ended up going with GSG Audio Evolution Sealed 18 with a smaller Lavoce Wan154 driver x 3 subs. Minidsp flex balance as the subwoofer brain. Crown xls2502 x3 for the amps. I chose the smaller 15" driver to better complement the 35hz to 80hz target range. Not interested in sub 35hz reproduction. But that's just me.

 

[boXem]

@fivenp cool initiative you have here. Did you already simulate the impedance of the driver loaded by the enclosure to verify that the bridged amp would not be limited by it's current output?

 

[fivenp]

@fivenp cool initiative you have here. Did you already simulate the impedance of the driver loaded by the enclosure to verify that the bridged amp would not be limited by it's current output?

according to simulations the impedance should never drop below 6 ohm for the 8 ohm version of the BMS.

 

[boXem]

according to simulations the impedance should never drop below 6 ohm for the 8 ohm version of the BMS.

Then you are covered.

 

[kemmler3D]

I think this is a cool project! Tend to agree that Hypex is an appealing amp for a given build, but 1) somewhat overkill for subs, you don't need SINAD that high in this case and 2) Not very open / future-proof.

Implementing DSP with e.g. DSPi or CamillaDSP would make more sense for an open project. For the amp, I don't have an immediate suggestion but in terms of housing you could offer a 3D printable case of some kind, which could then be adapted if the amp module needs to be swapped out at some point in the future.

 

[101geo]

Good idea overall. I'm personally not a fan of plate amps, seen too many go pop! Having the enclosure designs available as a CNC file would be cool for the less able woodworkers. I would also like to see a collaboration for 15", 12" and 10" for those of us in Europe and Asia suitable for smaller homes.

 

[fivenp]

Thanks for all your thoughts so far. I do agree with some of the sentiments here
The beauty of being open by design is:

* We can easily separate DSP / Amps and branch off different Versions of this build - once the EQ settings are settled and confirmed "working"
* We can theoretically even use different Drivers and optimise for those

I did simply chose the Hypex and BMS combination because a) I have them here next to me and can build / validate / optimise based on them - b) they will def. deliver what on what you expect and c) the Hypex price is actually not too extreme for what you get (feature - but also power wise)

Thinking of ppl who are less versatile in DIY the approach with the Hypex all in one is just very convenient. But again - we can absolutely build different variants based on different DSPs and Amp combinations....

I would also like to see a collaboration for 15", 12" and 10" for those of us in Europe and Asia suitable for smaller homes.

1000% That's the goal!
I did have a 15, 12, dual 10s, and other builds on the back of my mind – maybe even something that goes more towards Genelecs W371A or Kii's BXT - a cardioid build that goes higher up to the mid frequencies.

I just wanted to get the initiative rolling

 

[Blankenstejner]

Hi,

Nice initiative. Thought I'd chime in as I have a pair of 18n862 subwoofers with FA502 bridged like you are suggesting, though in a smaller volume sealed enclosure (113 liters minus the hypex box etc.): https://www.soundimports.eu/en/denovo-audio-300-7079.html

I used them initially with a pair of KRF R3's with a crossover of 125Hz using a minidsp shd with dirac live and was never quite happy with them playing that high. Might very well be my ability to properly measure and set up dirac live (using a minidsp UMIK-2). With dirac and a fairly simple boost in the FA502 I had them playing down to 10Hz "flat" at my listening location (4.4mx4.4m room).

In my current build I plan to use them from around 60-65Hz down. Very good sounding subs with plenty of grunt, I just don't like them too high up (so far).

There's also a good test over at data-bass for a sealed BMS 18n862: https://data-bass.com/systems/5c48e01911126b0004ca12ec

 

[Salt]

Think about J_J's assumption that down to 40 Hz should be near to main speaker...

 

[fivenp]

Nice initiative. Thought I'd chime in as I have a pair of 18n862 subwoofers with FA502 bridged like you are suggesting, though in a smaller volume sealed enclosure (113 liters minus the hypex box etc.): https://www.soundimports.eu/en/denovo-audio-300-7079.html

Amazing to hear. I think the 10L should not really matter here

I used them initially with a pair of KRF R3's with a crossover of 125Hz using a minidsp shd with dirac live and was never quite happy with them playing that high. Might very well be my ability to properly measure and set up dirac live (using a minidsp UMIK-2). With dirac and a fairly simple boost in the FA502 I had them playing down to 10Hz "flat" at my listening location (4.4mx4.4m room).

That's very interesting. I had them in a similar chain (Dirac, SHD, etc) in an OB at some point playing up to ~150Hz and while I felt they were a little weird at the beginning – I've never heard better upper bass after....

Would be curious to know what boost you had on the FA502?

I remember Troels is using the BMS up to 200hz in his legendary 4-way - maybe there's something to learn from there

The-Loudspeaker

 

[Richard G]

Think about J_J's assumption that down to 40 Hz should be near to main speaker

Apologies, maybe I need stronger glasses. Who is J_J in this thread(?) …and an assumption re 40 Hz - was in this thread?

 

[Richard G]

Troels is using the BMS up to 200hz in his legendary 4-way - maybe there's something to learn from there

Without going and checking a graph of activity versus driver size, from memory that sounds entirely believable

IGWS the optimum crossover frequency depending on whatever driver you are crossing over to, size efficiency and amplifier driving it

 

[Salt]

Apologies, maybe I need stronger glasses. Who is J_J in this thread(?) …and an assumption re 40 Hz - was in this thread?

https://www.audiosciencereview.com/...ereo-bass-with-easyeffects.61716/post-2606935

 

[Richard G]

https://www.audiosciencereview.com/...ereo-bass-with-easyeffects.61716/post-2606935

There’s some validity in what he says. However, if you look at the distortion figures for just about any mid woofer, no matter how good it is, distortion begins to rise rapidly from typically around 80 or 100 Hz. And even though distortion might be not quite so bad in the range of upper base, taking the excursion away from a mid woofer and giving it to a subwoofer does remove the ability to reproduce without much distortion.

JJ is arguing from the point of view of directivity, and seemingly more in the context of speakers for theatres.

Whereas I am arguing from the point of view of minimising distortion in the upper upper bass

Choose your poison/ horses for courses

 

[fivenp]

Wouldn't we "sort of" solve both worries if we'd literally and physically create a modular 3 way out of a 2 way speaker?

In our specific case here I'd say if we stack e.g. a KH120 physically on top of one BMS 18" per side and cross them at ~ 120-150hz (or even a little higher) plus time align and level match them accordingly – we'd theoretically have a very coherent "main" system here?

Instead of treating the 18" as a classic subwoofer and stuff it into corners – we could treat it also as a bass extension module....

I think room interaction might become a different topic alltogether tho

 

[Luvchampagne]

Hi,

Nice initiative. Thought I'd chime in as I have a pair of 18n862 subwoofers with FA502 bridged like you are suggesting, though in a smaller volume sealed enclosure (113 liters minus the hypex box etc.): https://www.soundimports.eu/en/denovo-audio-300-7079.html

I used them initially with a pair of KRF R3's with a crossover of 125Hz using a minidsp shd with dirac live and was never quite happy with them playing that high. Might very well be my ability to properly measure and set up dirac live (using a minidsp UMIK-2). With dirac and a fairly simple boost in the FA502 I had them playing down to 10Hz "flat" at my listening location (4.4mx4.4m room).

In my current build I plan to use them from around 60-65Hz down. Very good sounding subs with plenty of grunt, I just don't like them too high up (so far).

There's also a good test over at data-bass for a sealed BMS 18n862: https://data-bass.com/systems/5c48e01911126b0004ca12ec

With dirac and a fairly simple boost in the FA502 I had them playing down to 10Hz "flat" at my listening location (4.4mx4.4m room).

Curious at what spl level you reached 10hz flat and at what distortion levels?

 

[Richard G]

Instead of treating the 18" as a classic subwoofer and stuff it into corners – we could treat it also as a bass extension module....

I think room interaction might become a different topic alltogether tho

Absolutely.

 

[fivenp]

Moving forward with the enclosure I'm wondering how crazy we should go with the bracing.
Initially I was expecting to have a central cross with 2 window braces (as in the attached dxf cutout sheet example) - given the force of the 18" I am not sure if we potentially need more tho

Thinking about budget – going with 22mm MDF and 44mm for the baffle

3x 550x550mm -> front/back
2x 536x550mm -> sides
4x 506x536mm -> for braces + top + bottom

We'd end up with ~ 2.6sqm of wood
This would give us some comfortable space on a 280x207cm sheet but will not work on more "default" sheet sizes where we'll need 3 sheets per pair

I'd love to also have another variant or two for e.g. 21mm Plywood Birch (it does cost ~ x3 of the 22mm MDF over here tho)

If anyone is good with DXF and annotations and all the things related or has hints on what to ideally use to shape the right files - please shoot! AI is helpful here – but my trust factor is pretty low

 

[mbd-00]

Hey! Thanks for sharing via the simulator. Made an account just to jump in on this thread. Would love to be a part of the wider open initiative.

Some ideas I've been thinking over/working on that might be relevant:
- Establishing open standards/formats for project sharing and drivers, and a cross-tooling compatibility layer (e.g design once, load into any tool). Existing formats aren't actively open to my knowledge, and there's a lack of consistency as to what they can include, or be extended to include - e.g things like traced datasheets, rolling in measured vs simulated data, EQs, ability to support TL/horns as well as more basic lumped impedance model simulation. Could also be things like amplifier specs, spec-sheet tracing, lot of potential scope
- Better open libraries of drivers and specsheets (actively working on this problem) and better comparison tools (e.g let's say this driver is not available in your region, what's the best local trade-off, and how do we understand those tradeoffs?)
- Price-aware simulation (e.g what driver behaves well against given parameters) - such as best sub for $X with low QTS, targeting max SPL at Y frequency in a vented box tuned to Z
- Published validation datasets against trusted tools