DIY project - cab design and build

[J Factor]

I just wanted to share a project I'm working on.

I have several cabs I use with my Mesa heads, but I've always wished for a more compact and lightweight 2x12. A lot of them are heavy and bulky. My current large vertical 2x12 sounds excellent, but it weighs about 90lb so I don't like to lug it around much.

A few weeks ago I started thinking at how to arrange a pair of 12" speakers in a very compact way to save cab size and therefore material. I came up with what I later learned is basically is a Marshall 1966 - a pair of speakers on a 45 degree diagonal, and they fit in a box that's 24" square on the outside. Not bad! I haven't ever come across one of those Marshall cabs in person but it seems like a good way to package a 2x12.

I'm also currently back in school full time as a mechanical engineering student (I should have finished nearly 20 years ago). This not only keeps me in the design and optimization mindset, but also I have access to the various wood and machine shops on campus, including a very modern maker space with a 3-axis CNC router - perfect for precisely cutting panels for a cab! I need only supply the materials; machine time is gratis.

After kicking around some more concepts, I decided to start drawing this thing up in CAD. I've gone through a few revisions and adjustments since then. I wanted to incorporate some reasonable approaches to keeping a light weight while maintaining the strength and sturdiness I've enjoyed in my Mesa cabs, so there are several design features here that cross Marshall and Mesa bits with some of my own. Here's the basics:

- Baltic Birch ply all around
- 12mm side panels and removable back panel
- 18mm thick baffle board, dado joined to each side panel (full perimeter dado)
- 18mm rear gusset panel, also full perimeter dado for strength. This serves as the attachment point for the rear panel
- Edges use a compound joint: half-round finger joints with an internal miter to the depth of the baffle/gusset dadoes
- Separate, removable grille board

The permanently installed dado-jointed baffle is a Mesa signature that I really like, as opposed to a baffle connected with cleats. It also provides very good strength and stability, resisting racking and distortion in the outer box frame. I figured I could add to it even more by doing the same sort of construction with the rear gusset, and the combination would allow me to use the thinner 12mm ply for the sides and back. In the pictures below you can see the rear gusset has large internal radii to avoid any stress concentrations. The two strong 18mm panels will also take the bulk of any vertical loads placed on the cab, so I won't hesitate to sit on it or stack other equipment on it.

The main box frame joints didn't need to be super fancy, but I figured if I have some precision machinery to do the detailed cutting, I may as well give the combination of miters and half-round finger joints a try!

Here are some pictures of the CAD models (with obvious cutouts for side handles and a rear jack plate)

full-dado baffle and gusset, section view

A small sample of the edge joining features - I will cut a few of these first to see how the geometry works in practice.

By my calculations, using some weights and measures of Baltic birch ply from various suppliers to get a general density, I expect the completed cab to weigh in about 40 ish lb fully loaded, give or take a couple pounds. This is using mid-weight speakers like Greenbacks, not even going for Neos. It might not be the lightest cab ever, but it's certainly enough of a weight savings for me to go ahead with the project.

Today I picked up the required Baltic Birch ply - one 5x5 each in 12mm and 18mm. I'm looking to get the panels cut some time in the next few weeks, with main assembly shortly thereafter.

Anyways this might not be the world's best cab, but I figured for the cost of materials I should at least give it a shot. I welcome any thoughts and feedback! I'll post updates here as I make progress, and I'll include some photos as I go.

Edit: latest updates

post 35

post 36

 

[gulliver]

Keep us updated … with pic.

 

[slayerbear17]

Looking good! How much roughly is the build going to be?

 

[Madrok]

Nice. I’m not sure you save much material, or weight, as opposed to a vertical orientation. Depends how wide the original was I guess. You should get a different response from a square baffle with maybe a larger surface area…more full-bodied?

 

[JimRad]

Looks good, I have just done almost the same thing but I decided to build a couple of 1x15" cabs for Celestion Fullbacks. It went really well and they became a lot lighter than the 2x12 cabs I was using.

I used 12 mm as well and 18 mm for the baffle and some stiffening of the back plate. Here is my build plans and pictures: Celestion Fullback cabinet plans

/Jim

 

[J Factor]

Looking good! How much roughly is the build going to be?

So far I've already purchased the Baltic Birch plywood and all the hardware (handles, screws, jacks and plate, wiring, tee nuts, rubber feet, etc). I haven't factored in the cost of finishes, but I have access to several options at the school shop that I can use so I might not spend up any further.

I already have a bunch of speakers I can use, so I'm not including any cost for speakers. For this cab I'll drop in my pair of Celestion V-Types.

I'm all in about $300 right now - certainly nowhere near as cheap as some store bought cabs, but they won't be this light, and it won't be as much fun!

Nice. I’m not sure you save much material, or weight, as opposed to a vertical orientation. Depends how wide the original was I guess. You should get a different response from a square baffle with maybe a larger surface area…more full-bodied?

I'm not sure what you mean by the original. My other (vertical) 2x12 is 30"h x 23"w x 15"d. It's built bombproof and is about 90lb loaded.

The Mesa 2x12 I had for a while was 30"x18"x15" or so, and was a bit over 60 lb.

This one should be ~40-45ish lb by my calculations. I'd say that's a savings, if my numbers are right! My 24x24x14 box volume should be right about what the Mesa was, but much lighter. I guess I'm just hoping not to lose too much low end compared to the bigger cab, but that's one of those things that I won't know until I get the thing put together.

 

[Chocol8]

Looks good! You could save a bit more weight with a thinner back panel. 12mm is way overkill, it just needs to block air.

Also, if the rear panel is supposed to be removable, put the jack plate on the permanent rear gusset.

 

[amphog]

The corner joints are always built square, then rounded over after assembly.

 

[Baelzebeard]

Square boxes are more prone to standing waves,(or so I've read). You might want to rectangularize it a little like 27x22 or so.
It will also let a full size head sit on top without hanging over.

 

[J Factor]

Looks good! You could save a bit more weight with a thinner back panel. 12mm is way overkill, it just needs to block air.

Also, if the rear panel is supposed to be removable, put the jack plate on the permanent rear gusset.

Thanks! I debated that, but there was also a cost economy to only having to buy two boards. I was also concerned with the kind of flex and drumming that happens in some cabs with thin backs - in fact, if there is any issue with back panel flex, I plan to add a few thin bracing ribs on a diagonal, much like you see under the top of an acoustic guitar.

This isn't meant to be a convertible cab, just that it wasn't to be permanently fixed - the rear panel will be attached with screws, and can be removed to install the rear-loaded speakers. I suppose it would be easy enough to cut the back into thirds and be able to remove a section for semi-open-back operation, but I'm aiming for a closed cab first and foremost.

The corner joints are always built square, then rounded over after assembly.

Ah, good point! Yes, the joints will be cut square and rounded with a table router after glue-up. What I meant by "half round" is that since the CNC router can not cut a perfectly sharp internal corner, those fingers will have a radius where they mate with the adjoining panel. I've designied in an equivalent radius (based on the tool insert bits we have available) on the mating external edges of the fingers so I won't have to do any corner sharpening with a file or table saw. This resource has a better graphic of what I'm talking about: https://www.woodsmith.com/article/episode-002-joinery-with-a-cnc/

It might be easier to show you what I mean without the wood texture:

Square boxes are more prone to standing waves,(or so I've read). You might want to rectangularize it a little like 27x22 or so.
It will also let a full size head sit on top without hanging over.

I have heard this as well, but there are also a lot of reports that the Marshall 1966 cabs sound great, and they are of very similar size and shape to this. Since I don't have one around to copy, I can't make it exact, but it seems like a reasonable starting point.

All my amp heads are 23" wide Mesas (Triple Crown and Mark V) so I was trying to stay as compact overall as possible. I do have a 4x12 and that other big 2x12 if I ever get into a full-width head again.

 

[J Factor]

Ok, update time!

I have been delayed a few weeks due to shop staff and tool availability as well as my own workload, but we're back on track. I have received all the components and raw materials, and made adjustments to fit the handles and jackplate after having them on hand for direct measurement.

I spent several hours working with the shop lead to bootstrap using RhinoCAM to set up the CNC toolpath. This is the final set of instructions that's sent to the machine so it knows how to move to cut the part correctly. There is a huge potential for either time waste or efficiency based on the paths and operations chosen. I worked the process time down from 20 min (for just one sample part) down to about 8. The complex half-round finger joint geometry takes the most time - the rest of the panels (baffle, gusset, rear cover) can be carved out in just a minute or two.

After several iterations and adjustments, I have a full toolpath for the small sample part shown in blue in a previous post. Once one more tool insert arrives (hopefully within the week) I'll cut out two of these sample parts, check their fit, and glue them up to prove stability of the final finger joint design. The samples use the same joint geometry, but are small to save time and material in case we need to make adjustments.

If all looks good with the samples, I'll set up the toolpath for the whole boards using the same operations, and hopefully be cutting parts within 2 weeks!

Here are some screenshots of the CAM process

We also did a quick test of an unknown vee mill bit on some scrap foam, just to get some eyes on how it cuts. The cut wasn't great so we're ordering a more appropriate one.

 

[J Factor]

Finally we got back to some progress! The shop had to order some tooling inserts, and we had a bit of a queue delay, but today I was able to cut out the test panels and fit them.

The CNC router worked perfectly once we got the setup right, and made quick work of the panels. They took just a little adjustment by hand with files to get the fit right afterwards. Unfortunately I forgot to get pictures of that process, but the pieces are now clamped in glue up.

We'll give them 24h and then perform a pull apart test. If they can withstand at least 20-30 lb pull on their own (even without the benefits of the full joint length, dado boards, and full box shape) I'll be satisfied and we'll go ahead with cutting the full parts.



You can watch the videos of the CNC running and me geeking out if you like:

 

[J Factor]

After being clamped for 2 hours, then 24 hours of glue set time, I ran the sample along a table router to give the edges a 3/8" round. Then I performed a stress test on this little 4" long joint section. There was a little bit of tearout from the router, but it's no problem as long as I use it in the climb-cut direction.

The sample was placed joint-up on the floor, forming an inverted V. I had a 130lb person stand on the joint, but not jump on it - no problem. The joint held firm. Next up was a 180lb person, same story. Finally I did, and I'm nearly 260lb... the joint still didn't budge. There are no signs of glue separation or ply issues at all. I'm extremely pleased and impressed with the strength.

The full size panels will have a much longer length to increase bending moments, but also will have more than 3x wider section of fingers and glue joint, plus will have the internal reinforcement from the dadoed baffle and rear gusset. I am more than confident in the strength of this construction, and it's full speed ahead to CNCing the full size project. I'm excited!

 

[Chocol8]

That looks extremely solid. If that cab doesn't hold up to regular uses, you will need to seriously rethink your life choices!

 

[gulliver]

That looks extremely solid. If that cab doesn't hold up to regular uses, you will need to seriously rethink your life choices!

Cab? OP made a boutique door-stop.

... as if I could do anything like that.

 

[J Factor]

Cab? OP made a boutique door-stop.

... as if I could do anything like that.

One that'll hold a hell of a heavy door!

I'll be working out the toolpathing on Friday I think, hopefully to start making lots of sawdust early next week. In the meantime, my cat likes to play around with the joint sample.

 

[gulliver]

One that'll hold a hell of a heavy door!

I'll be working out the toolpathing on Friday I think, hopefully to start making lots of sawdust early next week. In the meantime, my cat likes to play around with the joint sample.

I’m allergic to cats and dogs, so we have a mini rabbit. As indestructible as your joint is, a mini rabbit can turn it into a golf ball.

 

[J Factor]

Moving right along!

I got the toolpathing done over the weekend for the full 3 boards (all parts), got it verified this morning, and spent a good portion of the afternoon making sawdust. I have two of the 3 boards done, and will cut the third tomorrow or Thursday. I have two of the four sides, the baffle, rear gusset, and grille panel all ready for final edge cleanup, with just the two other sides and the back panel to go.

Machining time was supposed to be 2 hours for the 12mm board (those finger joints take some time to detail) and 15 min for the 18mm board. We got both knocked out in just under 2 hours, including setup time, so it worked out well.

We set the safety height on the CNC so it wouldn't machine into the support table (which is a spoil board, just in case there is a mistake). The machining process leaves a super thin, somewhat transparent layer of fibers that we remove by hand with a box knife or chisel. I didn't have time to do final edge cleanup with a file today, but I'll do that while I watch the CNC do the last board in the next day or so. As soon as everything is cleaned up, I'll test fit and then it's gluing time!

Here are a bunch of short videos should anyone be interested in watching the machining process. These are in the order of operations.

Horizontal roughing


Parallel finishing of finger joints - this process took about an hour


Facing of dado channels


Finishing the miter


2.5-axis profile cutting the outlines

 

[zenfreud]

Very cool speaker cab building project, thanks for sharing.
Surprised there aren’t more threads about designing and building light-weight cabs here on TGP. There’s quite a bit of that at the Talk Bass forum.

 

[J Factor]

Got the final parts cut out today! We had a little issue with the final clean up pass not cutting deep enough, so we had to trim out the parts with chisels. Not a big deal but there are a couple spots where the last ply layer is lifting, so I'll glue and press them for a day before continuing with assembly. That said, the joints line up great so far!


Parts from yesterday before edge cleanup

Now the rest from today. You can see some of those lifting spots here. Not too bad and it'll be ready to go by the weekend.

The dado channels are very tight, so I'm debating sanding the sides lightly or just chamfering the baffle edges and giving it a press fit.