Speaker building all by yourself

You might be one who is searching for World class speaker sound.
As it stands, perhaps you are not filthy rich and can not straight out buy 20k+ speakers.

Or you simply like to build stuff.

Or there isn't a speaker in the wide World just like what you have envisioned.

Welcome to the SBAF speaker building thread!

Share your vision and experience.
Design choices and thought process what made you do the decisions for driver selection and filters.
What materials, tools, and design soft was used etc..


Stuff that people have built.
In time it would cover nice selection of DIY speakers
with SBAF signal for the sound
(mods, feel free to add stuff to the list):

Builds:

• A Swope 2-way built by @ergopower and his son
  
• Troels' Discowery W18 built by @Jeb
  
• Alpair 6.2P & 7.3M built by @Jeb
  
• USM-2 built by @ultrabike
  
• Linkwitz LX521 built by @Ardacer
  
• SuperCube built be @OJneg
  
• A two way from 1'' compr to horn + 10'' woofer by @jpoyarzun
• Diy Overnight Sensations built by @Josh
• Diy Overnight Sensations built by @Jeb
  
• PAP Trio 15 built by @Btdk34
• AC 1.6 TMM OBs built by @sphinxvc
  
• Custom OBs built by @dBel84
• Serious' OB built by @Serious
• A BLH built by @JK47
• X-LS Encore built by @Puma Cat

You built a speaker and it's not int the list and you want it featured?
Let me know and I put it in the list. Or a mod can do it.

Contemplations:

• High Efficiency speakers for flea watt amps
• Advice for speakers for SET amps
  
• High Efficiency bookshelves myth
• On the shelf OBs discussion
• A speaker design discussion

Post pictures.
There is no solid definition of DIY, so clever mods, creative whack jobs and teardowns are also welcome at this time.
Showing off tools and workshop is fine.
Showing clever woodworking techniques is better.
Showing how to get best sound out of your particular crush of a midrange driver, even better.

 

A bit of a back story is in order of my road to speaker building.

At some point I realized the Genelec 8050s that I had been content with for almost 10 years were not really world class sound.

So, choices?

After some listening that I could do in my country there was easy decision to go for big 3-way speakers.

Problem was, what I would have been happy with cost more than my year worth of salary.

And even then, I thought, perhaps this would be only temporary satisfaction, after which I’d want even better or different one.

Decision to learn how to design and build speakers was easy. I was a complete newb. I had heard of T-S parameters and knew the concept of Helmholz resonator, all on the surface.

First practical step: I bought an old farm to have a place to build and to listen to whatever speakers I want.

Second practical step: I bought bunch of Makita tools, chisels, handsaws, clamps etc. For the money of the tools some decent speakers could have had already.

In the road to learn I exhausted everything that I could get my hands on during a 3-year period (and on-going).

Vision: I wanted a speaker that would be ideally full range. In learning process, I understood that a 3 way is not enough drivers to do this without compromising dynamics. So then, I settled with a -3dB point of 25 Hz and even that was pushing it with a 120 L enclosure that I set as max target volume. Obviously, it had to be ported. Sensitivity at least 90 dB/W/m. 95 dB/W/m would have been better.

I remembered the golden speaker compromise rule: high sensitivity, full range, sensible size – pick two.

I wanted to go with sort of a well paved road as my first high end speaker construction to make life a bit easier. So, no horns/waveguide (this changed), simple cone and dome drivers. Simple LR2 cross over (this too, changed).

Drivers:

I started my search from often recommended midrange driver perspective.

What a can of worms, or rather a rabbit hole. There is no simple method to chose between the drivers based on spec. T-S parameters are only of secondary importance here. Klippel analysis about as useful as the former.

Measurements of dispersion and harmonics are slightly more useful.

The sound of midrange will be impacted by materials of the cone and the cone construction more than that of the motor tech. Difficult to put this in numbers and graphs.

The practical guidelines I followed were: first; linear fr, lowest Mms while not showing much errors in impedance graph. Second; high Qms, short xmax, tiny coil, low harmonic distortion

In the end I have to admit I relied on impressions of people who had used similar midranges in similar concepts and were happy + added some of my backbone faith and decided to get a SBAcoustics Satori midrange (SATORI MR16P-8). Close contenders were Skanspeak Revelators. I even seriously considered high-end widebanders Like Seas Exotic and Fostex.

Bass was easy. First, I filtered out what I was looking for in terms of T-S parameters. Low Qes and big BL (strong motor --> fast firm bass), low Fs (duh), high Qms (more effortless), high Vas (more compliance --> more plankton). Also, Mms much more than 100g for 12’’ driver is probably better suited for sub duty, I prioritized mid-high bass more.

Next, I looked that it would have shorting rings and extended pole piece in the motor for lower inductance and cleaner mid-high bass. Funny enough, again a SBacoustics driver (SB34NRXL75-8) surfaced. There is a nice Revelator 13 inch that would have done roughly the same job, only it costed twice as much and was thus disregarded.

At this time in my journey I had no idea how good PA woofers can be. I thought the lowish Vas, high Fs and less advanced motor structure (ignorance in my part of not knowing where to look) were deal breakers. Well, next hi-fi speaker I make will use a PA woofer for sure.

For tweeter there was a lot to choose from. Great candidates from Seas, Scanspeak, SBacoustics and hell why not Accuton. Guess what? I got a SBacoustics driver, again (SATORI TW29RN).


The box:
For box material I was set on birch plywood and or solid wood elements. I simply dislike the MDF for its crappy durability to moisture and mechanical damage. I considered other materials, but since I already got woodworking tools, this was after the fact mental wandering.

The box design and appearance. Alright, now real trouble started. My engineering and physics background was of no help. I simply couldn’t understand how to make a speaker that women would accept as furniture piece besides the TV. This was a requirement, so it had to be solved.

I drew a clade full of isometric views of all the shapes I could come up with, at the same time having some faith that what I drew I could actually craft out of wood and plywood.

For a man cave which I already had I could have done a rectangular box and be done with it.
This was tempting.
Ahh, too easy. I needed some challenge. I had those fancy tools already. The conclusion is what is on pictures. Bent side panels, flat top-bottom, port on the bottom, drivers tilted on the front, yes mostly for looks.

I put some serious effort into the cross-bracing.
The cabinet side walls are 21mm thick and curved. Top is total 35 mm and bottom 28 mm. Front 45 mm. This cabinet turned out like a solid log, when picked up or knocked on the sides.
Haven't weighed the finished speaker yet, but picking it up it feels as heavy as my brother who is a big hulk of a man.

For cosmetics and for some added ruggedness some oak side strips were added along with top cladding.



For ease of access to the drivers the back panel comes off completely.

XOs and changes to the acoustic layout in another post. To be continued..

 

I'll play! Fair warning though, not very innovative.
Maybe the biggest factor was a project I could do with my son. He was a Biomedical Engineering student at Boston U, and for sophomore year, was stuck in an extremely small university apartment with a fellow engineering student. I helped him move in and wow, very little free space. He wanted something for music, and I suggested we build something, which he was happy to do.
So the requirements were really small mains for a shelf. We looked through a bunch of designs, and found the Swope by Paul Carmody. The rear surrounds were small and sealed, so should be OK with the placement available. I had an Energy 8" sub I wasn't using, and there was a place he could fit that on the floor underneath the shelf.
A friend had a really nice woodshop at the time, along with a bunch of accumulated hardwood. We found enough cherry boards for the job and laid out a cabinet design of our own based on the internal volume needed, the widths of boards that we had, and a pair of speaker grills I had laying around. To avoid seeing edges of solid boards, we made a design with the top and side panels with 45° beveled edges.
Pictures show some stages of the build. The final shot is without the screens, and you're seeing embedded magnets to which the screen attaches.
We then also built a box that combined a Lepai D-class amp and a MiniDSP. We used a MiniDSP to drive the sub so that the mains were high-passed so they and the amp weren't struggling to produce <100 Hz.
It all went together well, no errors that I recall. I played the Swopes in my system, and they were OK, not great, but not much affected by wall placement, which was one of the bigger goals. Overnight Sensations I think are a decent step up, but rear ported.
And my son learned something about speaker design, soldering and woodworking.

 

An understated quality of an excellent parent for sure.

I see that the tweeter and woofer are closer together than in the original design.
I haven't had time to play with this myself with small speakers to really find out whether or not it makes a difference in nearfield use.

 

Inspiring thread! Some day, some day... For now i’m gonna build something that matches my skill and available tools. New acoustic panels

 

Hah, yeah that wasn't done for sound quality reasons - they had to be that close to fit behind the speaker grills I had. I used a dremel and cut the arc off the flange right up against the housing.

 

Great work! I assumed at first sight that this was a kit pair that came that way.

The woodwork looks lovely too. Nice job

 

Can't forget the awesome build by @Jeb

https://www.superbestaudiofriends.o...nspeak-discovery-18-diy-troels-gravesen.4406/

 

Thanks for the reminder. It makes sense to link everyone's efforts in the first post.

 

I've been trying to build a pair of 2.5 ways for longer than I care to admit. They are sealed floorstanders based on Seas drivers: the U18RNX/P woven PP cones, and the 27TBFC/G tweeter. The PP woofers have a very well controlled breakup, a great motor, and most critically, were on sale at Solen. They are also wildly unpopular compared to the other 18cm drivers Seas sells, with most people opting for the other PP cones that Linkwitz uses; probably because these have much less bass output. Fortunately, I have two.

Here's the box that I definitely could not have built by myself.



There were many many many iterations. Right now I'm using a nanoDIGI as a Harsch crossover. This feeds I2S to a PCM1798, into a balanced discrete I/V stage. The power amp is a balanced transformer-coupled follower without global nfb, similar to Susan Parker's Zeus. There are some pretty neat optimization opportunities here! Since the tweeter will basically be out of the picture by 500Hz, for example, I re-stacked the DAC's OPT laminations to introduce a gap, trading bass performance for improved linearity. Likewise, you can use a much more aggressive lowpass filter on the DAC that feeds the woofers.



Why is there an unconnected tube socket? Don't worry about it.

Looking at the specs of the U18, my sense was that it would work quite well in a current drive system. I reconfigured the amps for differential current drive and got some pretty compelling results.



F3 went from 90Hz to ~50Hz, and it integrates much better with the sub. Note that this is nearfield, so is only valid to 500Hz or so. Interestingly the distortion hasn't changed much, even though this said to be one of the main benefits of current drive. My thinking is that since both amps are without global nfb, there is no back-emf induced distortion to deal with in either situation-- but the current amp could still stand some tweaking, and I should retake the measurements at a higher level.

I'm hoping to get the new amps done by the end of the month, and maybe swap the tweeters for some fabric dome, or something in a waveguide. And at some point, possibly upgrade to some FIR-capable DSP, and maybe even put everything in a box?

 

Nice work so far, can't wait to see how they end up looking finished.

Beautiful thread, thanks for including my build!
I'm planning a 4way extremely slim ob build that should have in theory the same performance as lx521. We'll see how that goes.
Not going to be cheap. In fact it will cost pretty much the same, but it should be as slim as physically possible with full force cancellation of sub drivers.

 

Also I remember reading somewhere that Marv built most of his speakers. That would be interesting to see if someone can find it if he posted it.

 

Ok, I'm totally cheating here (assembling, not building from scratch), but this vid has me interested:



Seeing some third party measurements confirm a few things about those woofers is even better. And, it seems Rick at Selah Audio has a kit. Sitting in my attic are a pair of old PE .75cu/ft curved cabinets sans baffles thanks to an earlier project. Hmmmmm.

I reached out to Rick about the Purezza kit to see if I could hack something together with existing cabs and some new baffles. This could get fun!

 

Fit in a car door, lol

I've seen this driver pop up around a few places and I can see why there's a lot of excitement for it. Troels recently released some measurements on it and this Joseph Crowe transmission line looks bonkers:



One thing to note, this is an 88dB driver so no low powered tube amps. This needs lots of clean watts. I think Bruno Putzey is involved with this driver so I wouldn't be surprised if this was developed as an ideal pairing for his NCore modules.

 

At last, a SBAF topic I can contribute some (hopefully) useful information!

A list of possibly useful links;
Freeware / speaker design software;

http://audio.claub.net/software/jbabgy/jbagby.html
Author, Jeff Bagby, was a super nice guy, and contributed many excellent designs, as well as this freeware. Unfortunately, after suffering kidney failure, and then a transplant, he contracted Covid-19. He was the first person in Indiana to succumb to the virus. A tremendous loss to all, particularly the DIY community.

https://kimmosaunisto.net/Software/Software.html Great, powerful design freeware. Author updates it constantly, and even provides a user manual, plus "how to" vids via YouTube.

Analysis Freeware, just need a laptop, calibrated mic (Umic or Dayton, not too pricey either way), and a USB/SPDIF converter (you then can hook directly to your DAC), or a USB cable long enough to move yourself and the laptop out of the sound field;

http://www.artalabs.hr/

https://www.roomeqwizard.com/

http://www.holmacoustics.com/holmimpulse.php

And, a few parts/kit suppliers, if needed;


https://www.parts-express.com/

https://www.madisoundspeakerstore.com/

https://meniscusaudio.com

https://www.soniccraft.com/

Other kit plans, and design info;

http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm

http://www.humblehomemadehifi.com/speakers.html
HHH above has designs in the CAD library, plus an interesting capacitor comparison.

 

I know a lot of us are happy with paper cones in a wood box, but I haven't seen too many builds of a wood cone in a paper box. (Maybe Harbeth? )



The woofer's an Airborne FR173B6 that's currently half off. The novel upcycled semi-open baffle is showing a lot of diffraction and baffle step effects, but it looks like it shouldn't be too difficult to work with these drivers.

 

I've definitely thought about using cardboard for cheap speakers. While it's worse than wood, it must be better than just going baffleless. Diyaudio has a big thread of people building foamcore speakers. Which I might actually do for a frugalhorn.

 

There's cardboard and there's cardboard. I've built subwoofers in Sonotubes and, properly designed, they work just as good as anything else. I know people have also used Sonotubes with cut-outs and flat front panels to good effect. The best speakers I've done didn't use parallel sides.

 

@yotacowboy you should take a look at some of the experimentation XRK is doing over at DIYA

That said, I have SELAH Veritas - Rick knows what he is doing

..dB

 

Nelson Pass subs are the most famous example