Build thread: Psalms tower speakers

[bimmerguy]

One of the stressful parts of this build is making the curved panels. I always find it a little nerve-racking working with glue you actually have to mix up. When you start out you're not quite sure if you're making too much (which would be wasteful of the expensive glue), or not enough (which would be worse since it takes a fair bit of time to mix it up, all the while you're already mixed glue is sitting around either in the pot or on the panel drying out). And truth be told, I was nervous about whether this was actually going to work.

If you've never worked with precatalyzed urea formaldehyde glue, it's weird stuff; it doesn't just dissolve into the water like you think it would. There is a lot of stirring involved and it takes a while for all the powder to hydrate. I sort of liken it to making pizza dough. Pizza dough doesn't have a nice smooth texture when you first mix it, but after you let it sit and hydrate for a bit and allow gluten to form, it becomes a lot smoother and more plastic. There's no gluten with the glue obviously, but the same sort of concept applies.

I tried to estimate the amount I needed based on the coverage listed on the container, plus 10%. I mixed the first batch up all at once - that is, enough glue to cover the 3 layers needed for one panel. That proved to be a little too big of a batch. I had difficulty getting a nice creamy texture and had trouble with lumps of unhydrated powder. They were a big pain as I had to scrape them off the sticky panel when I was rolling out the glue. The other issue with doing a big batch of glue was I wasn't quite sure how much I could use on the first two layers. I was afraid I was going to run out, so I probably used less than I should have until the final layer. I ended up with enough, but I don't like being to uncertain if I'm going to have enough. There were some anxious moments while I was waiting for the panel to cure in the Curvomatic wondering if the glue would fail and the curve would completely spring back.

So for the second panel I decided to make up three small batches in 3 separate cups - one for each layer. That way I would use an entire cup's worth of glue on each layer. And I made 20% extra instead of 10% (I have more glue than I need, and it doesn't have a long shelf life). Both the small batches and the 20% buffer were wise choices. Doing it in small batches, the mixtures ended up smooth with no lumps and I didn't need to guess about how much to put on each layer.

Here's my setup:



I weigh out enough powder for each layer in the panel in 3 separate cups. You add the water at two different times - 2/3 the first time, and 1/3 later, so I measure out the first dose of water into 3 other cups. I made a spreadsheet to calculate the weight of the powder, and first and second additions of water. I pour the water into the first cup and stir. And stir. And stir until it is smooth, and about the consistency of runny peanut butter. Then I let that sit while I do the same with the other two cups. By the time I'm done with those, the first batch is ready to stir again. Bubbles will have formed in the glue as the powder more fully hydrated and you want to stir those out. Then I measured out the second dose of water into each of the water cups. Then I add it to the first glue cup, and stir. And stir. And stir. Weird things happen when you added the second dose of water, but if you continue stirring, all will go well. Then I repeated the process with the other two cups. By this time they have all been sitting long enough that they are ready to go. With one more round of stirring.

It does not spread particularly nicely. It thickens fairly quickly, though it doesn't set quickly - so you do have time. But after about 6-7 minutes with a 6" glue roller, you can get a nice consistent film of glue across the sheet. Then you place the next layer of plywood on top and repeat...after you've stirred the next cup again, of course. Lather, rinse, repeat and you're ready for the clamps. Or Curvomatic. I like to wrap the panel is stretch plastic (like they use for wrapping pallets) to prevent the squeeze out from getting on anything. You can see the rolls of it in the next picture. I also put down a layer of baking parchment on the Curvomatic just in case the plastic wrap leaked.



The glue instructions are very clear that the temperature of the room and the parts must be above 70 degrees F for the glue to cure properly. My shop hovers right around 70 this time of year, so I took out a little extra insurance and use a couple heating pads with a few old pillows on top. I keep the heating pads and pillows around in my shop for exactly this sort of thing.



I'm never sure exactly how long I should leave the panel in the clamps, so I always err on the side of caution. I left each panel in there for 24 hours. Which means it take a minimum of 4 days to do 4 panels. But it was nice having an excuse for not working on it the rest of the day (though I certainly could have worked on other parts of the build).

When I took the first panel out of the Curvomatic the next day I was very pleased to see there was almost no spring back. Which was surprising - shocking even, given the incredible amount of force I had to apply to the panel to get all the layers to curve all the way. I was leaning on the ratchet straps with all my weight when tightening them up. Quarter inch Baltic birch plywood does not like to bend. But it turned out great. After they were all done, I removed the veneer tape and gave them a decent sanding.





I said _almost_ no spring back. There was a tiny bit. So I am going to remake my brace template to include the new radius. I also recalculated all my volumes - they changed by a fracture of a percent. It probably isn't strictly necessary to remake the brace template, but I want the panels to fit/glue to the braces perfectly. That was why I hadn't yet milled the brace blanks to their final size.

[JonMarsh]

This is some very fine work and very nice tooling you are using.

Do you use those aluminum bars and cable clamps for other projects, or were these bought dedicated to this build?

[Evil Twin]

One of the stressful parts of this build is making the curved panels. I always find it a little nerve-racking working with glue you actually have to mix up. When you start out you're not quite sure if you're making too much (which would be wasteful of the expensive glue), or not enough (which would be worse since it takes a fair bit of time to mix it up, all the while you're already mixed glue is sitting around either in the pot or on the panel drying out). And truth be told, I was nervous about whether this was actually going to work.

If you've never worked with precatalyzed urea formaldehyde glue, it's weird stuff; it doesn't just dissolve into the water like you think it would. There is a lot of stirring involved and it takes a while for all the powder to hydrate. I sort of liken it to making pizza dough. Pizza dough doesn't have a nice smooth texture when you first mix it, but after you let it sit and hydrate for a bit and allow gluten to form, it becomes a lot smoother and more plastic. There's no gluten with the glue obviously, but the same sort of concept applies.

I tried to estimate the amount I needed based on the coverage listed on the container, plus 10%. I mixed the first batch up all at once - that is, enough glue to cover the 3 layers needed for one panel. That proved to be a little too big of a batch. I had difficulty getting a nice creamy texture and had trouble with lumps of unhydrated powder. They were a big pain as I had to scrape them off the sticky panel when I was rolling out the glue. The other issue with doing a big batch of glue was I wasn't quite sure how much I could use on the first two layers. I was afraid I was going to run out, so I probably used less than I should have until the final layer. I ended up with enough, but I don't like being to uncertain if I'm going to have enough. There were some anxious moments while I was waiting for the panel to cure in the Curvomatic wondering if the glue would fail and the curve would completely spring back.

So for the second panel I decided to make up three small batches in 3 separate cups - one for each layer. That way I would use an entire cup's worth of glue on each layer. And I made 20% extra instead of 10% (I have more glue than I need, and it doesn't have a long shelf life). Both the small batches and the 20% buffer were wise choices. Doing it in small batches, the mixtures ended up smooth with no lumps and I didn't need to guess about how much to put on each layer.

Here's my setup:

[ATTACH=CONFIG]32330[/ATTACH]

I weigh out enough powder for each layer in the panel in 3 separate cups. You add the water at two different times - 2/3 the first time, and 1/3 later, so I measure out the first dose of water into 3 other cups. I made a spreadsheet to calculate the weight of the powder, and first and second additions of water. I pour the water into the first cup and stir. And stir. And stir until it is smooth, and about the consistency of runny peanut butter. Then I let that sit while I do the same with the other two cups. By the time I'm done with those, the first batch is ready to stir again. Bubbles will have formed in the glue as the powder more fully hydrated and you want to stir those out. Then I measured out the second dose of water into each of the water cups. Then I add it to the first glue cup, and stir. And stir. And stir. Weird things happen when you added the second dose of water, but if you continue stirring, all will go well. Then I repeated the process with the other two cups. By this time they have all been sitting long enough that they are ready to go. With one more round of stirring.

It does not spread particularly nicely. It thickens fairly quickly, though it doesn't set quickly - so you do have time. But after about 6-7 minutes with a 6" glue roller, you can get a nice consistent film of glue across the sheet. Then you place the next layer of plywood on top and repeat...after you've stirred the next cup again, of course. Lather, rinse, repeat and you're ready for the clamps. Or Curvomatic. I like to wrap the panel is stretch plastic (like they use for wrapping pallets) to prevent the squeeze out from getting on anything. You can see the rolls of it in the next picture. I also put down a layer of baking parchment on the Curvomatic just in case the plastic wrap leaked.

[ATTACH=CONFIG]32331[/ATTACH]

The glue instructions are very clear that the temperature of the room and the parts must be above 70 degrees F for the glue to cure properly. My shop hovers right around 70 this time of year, so I took out a little extra insurance and use a couple heating pads with a few old pillows on top. I keep the heating pads and pillows around in my shop for exactly this sort of thing.

[ATTACH=CONFIG]32332[/ATTACH]

I'm never sure exactly how long I should leave the panel in the clamps, so I always err on the side of caution. I left each panel in there for 24 hours. Which means it take a minimum of 4 days to do 4 panels. But it was nice having an excuse for not working on it the rest of the day (though I certainly could have worked on other parts of the build).

When I took the first panel out of the Curvomatic the next day I was very pleased to see there was almost no spring back. Which was surprising - shocking even, given the incredible amount of force I had to apply to the panel to get all the layers to curve all the way. I was leaning on the ratchet straps with all my weight when tightening them up. Quarter inch Baltic birch plywood does not like to bend. But it turned out great. After they were all done, I removed the veneer tape and gave them a decent sanding.

[ATTACH=CONFIG]32333[/ATTACH]

[ATTACH=CONFIG]32334[/ATTACH]

I said _almost_ no spring back. There was a tiny bit. So I am going to remake my brace template to include the new radius. I also recalculated all my volumes - they changed by a fracture of a percent. It probably isn't strictly necessary to remake the brace template, but I want the panels to fit/glue to the braces perfectly. That was why I hadn't yet milled the brace blanks to their final size.

Impressive... most impressive.

The Force is strong with this one...

[bimmerguy]

This is some very fine work and very nice tooling you are using.

Do you use those aluminum bars and cable clamps for other projects, or were these bought dedicated to this build?

Thanks, Jon!

The aluminum bars and straps are all part of the Curvomatic - which, yes, I did buy specifically for this build (I was sort of looking for an excuse to buy one), but which I plan to use on quite a few future projects -- not the least of which will be my center channel and subwoofer build. I will say, I have developed a sort of love-hate relationship with it. There is no other way I know of to simultaneously bend 3 sheets of 1/4 inch of Baltic birch plywood (short of a large dedicated hydraulic press). If using 1/4 inch MDF it would be a total breeze. So it allows you to do things that would be either be impossible or quite a bit more difficult and definitely more wasteful to do with traditional MDF forms. So in that sense I love it. It opens up some interesting design possibilities for other projects like cabinets, garden trellises, etc. You're not just limited to convex shapes, either. But it takes an age to set up. Assembling all those aluminum extrusions is quite tedious, and disassembling them is no better. And then I have to find somewhere/way to store everything. But all in all I think it was a good investment and will serve me well in the future.

[bimmerguy]

Impressive... most impressive.

The Force is strong with this one...

Thank you! I hope the speakers will turn out nice in the end. I'd really like to do Rick proud. I am not new to woodworking and machining, but I know so little about speaker design and acoustics. My undergraduate degree is in physics, but I didn't have much coursework in acoustics (I was focused more on optics/photonics). I am doing my best to learn, though. Guys like you and Jon and others here truly humble and inspire me with your great depth of knowledge and willingness to share.

[bimmerguy]

I haven't had a ton of time in the shop lately, but I have gotten a few things done. Now that the curved panels are done I was able to determine with a fair degree of precision what the actual radius of curvature was (though it was barely perceptible, there was a tiny bit of spring-back). I traced the inside curvature of a panel on a piece of paper (a couple pieces taped together actually to get a longer length). I used a 24 inch center-finding ruler to draw a line (cord) between two points on the curve 20 inches from each other, then at the mid point of the cord I drew a perpendicular line up to the curve. I could then use the length of the cord and the distance from it to the curve to calculate the radius. I set my homemade router compass to that dimension (the Jasper jig doesn't go big enough) and cut a new radius template. I used the template to check the consistency of the 4 panels and they were all -- amazingly -- exactly the same. With such an incredible amount of force needed to bend the Baltic birch plywood I thought there would have been more spring-back and more variability. I used that new template to make brace template 2.0, just like I did for brace template 1.0. It sounds perhaps like quite a bit of work, but it wasn't so bad.

I traced the brace template onto the blanks I had prepared earlier...



...and cut them to rough size on the bandsaw. Then I cut out the holes to rough size using the scroll saw. I could then double-stick tape the template to the rough-out blank, and template-route the piece to final dimension. I did a 1/4" radius roundover on all the holes.







After that was done, I did a quick test fit to see how they'd look assembled as the skeleton.

[bimmerguy]

I need to mill the back panel where the terminal connections will be. The plate that mounts the terminals needs to be perfectly centered in a recess in the panel, of course, and the rear panel needs to have its sides angled to match the curvature of the braces. Which presented me with a bit of a conundrum. Should I cut the angles first so I can find the its exact center more easily for the terminal panel? This would make it harder to clamp the piece in the vise for milling since the sharp angled edges are not able to be grabbed by the vise without being crushed. Or should I cut the terminal plate recess first when I can easily grab it in the vise, and then try to cut each angled side perfectly equidistant from the edge of the recess. I scratched my head for a long time about that. In the end I decided to cut the rear panel slightly undersized (narrower) so that even when the angled edges were cut there was still a slight flat on the sides. That way I could cut the angles and still grab it in the vise. And it had the added benefit of leaving a fixed reference edge that I could run against the fence when cutting the angle on the table saw. I could just turn it around 180 degrees and run the other flat edge against the fence, leaving the fence at the same location. This results in both the angled edges being perfectly centered on the piece (which could not have happened if the angles came to a sharp point).

These kinds of decisions really slow me down. I think when I make my center channel and subs in the future it will go much faster. So with my order of operations figured out for this step, I cut the angles on the rear panel...



...and I cut the angle for the piece that defines the rear of the MT section.

[bimmerguy]

The terminal plates will mount to the cabinet with very small screws. While I'm sure small wood screws would be perfectly adequate for the task, I'd like to make something a bit more...time consuming (?). So I decided to make some very small threaded inserts. I've made that sort of thing on quite a few projects in the past. I started out with some 3/16 inch brass rod (and some aluminum sheet I'll use later for the terminal plates).



I drilled and tapped a hole for a 4-40 thread.



Then I lightly knurled the OD to make a secure fit when pressed into the plywood. If necessary you can add a drop of CA glue when pressing them in.



After doing it 8 times, I had these:

[bimmerguy]

To make the terminal plate I started with the same 0.075 inch aluminum I used for the trim rings. I machined them to size and drilled the holes for the terminals and the mounting screws on the mill.



Then I milled out the recesses on the rear panel to accept the plate and the threaded inserts.



Next came pressing in the threaded inserts using the hydraulic press. I suppose I could have used a hammer, but this way prevented any damage to the soft brass and ensured the inserts went in perfectly straight. It takes a fair bit of force to push them in, surprisingly. I drilled the holes to be 3/16, the same as the original stock size, but with the knurling the diameter of the brass increased quite a bit. They are a very snug fit.





The plates sit just a little proud of the surface of the plywood. When the piece is veneered they will be flush.



I chose the Propeller posts for no other reason than I thought they looked cool. I might clean up the edge of the nylon isolation washers a bit on the lathe. They look a bit sloppy.

[Steve Manning]

One things for certain, you are not lacking in the tool department and the skill to use them.

[bimmerguy]

I have some questions for you experts.

One of the things causing my analysis paralysis is how to do the front baffle. As is typical, it will be made of two layers (3/4 inner, 1/2 outer) to form a final thickness of 1.25". I'll glue up the skeleton first - the braces, rear, and inner front baffle. Then glue on the curved side panels and trim those flush with the rear and inner front baffle. That leaves the outer front baffle. I plan on painting the front baffle gloss black. That's one of the reasons I veneered the curved sides when I did. The outer baffle will be black on its side (1/2 inch) dimension all the way around. It would probably have been easier (though taken longer) to hold off on veneering the sides until it was all assembled - that way the veneer could have covered the edge of the outer front baffle, too.

At any rate, now I have to decide how to merge the outer to the inner front baffle...and when to paint it, and when to route the holes for the drivers.

The transition from front baffle to side panel is what I'm concerned about. That's the area inside the green circle in this drawing.



Option 1: Alignment pins/paint the outer front baffle first. Install 4 alignment pins to allow the repeated and precise placing and removal of the outer baffle from the inner baffle. You could route all the driver holes and trim the perimeter to form a perfect edge with the side panels before you glue it, and then remove the piece for painting, and finally reinstall and glue it, already painted. I am assuming the pins would allow for enough precision that when placing the outer baffle back on it would align nearly perfectly. This has the benefit of allowing you to paint just the outer baffle by itself, which will result in a crisp painted edge when glued to the rest of the carcass.

Option 2: Glue everything first. After the side panels are trimmed flush with the inner front baffle, glue the outer front baffle in place, trim and shape the edges to fit with the sides, then route the driver holes. And finally, mask off all but the front outer baffle (and its 1/2" side edges) and paint it. This way is probably easier, but I am concerned there would be a thick/hard paint line where the outer baffle and side panel meet (that is, where the masking tape was). There will be several layers of paint (epoxy primer, sandable primer, color base coat) and the masked edge might be hard to deal with and ugly. And sanding the primer will be tough right next to the veneer.

Option 3: Is there another option?

In either case I will spray the clear coat after the entire speaker is assembled.

Are one of those two options (or a third) the way things are typically done? I'd like to do an inset aluminum strip to separate the inner and outer baffle, but I don't have any material long enough for that (I could get it, of course), but moreso I'd be worried about sanding through the veneer when trying to file the aluminum flush with the surface. And that still doesn't necessarily point me in the direction of option 1 or 2.

Any insight would be greatly appreciated.

[bimmerguy]

...and the skill to use them.

Thank you, Steve!

[tf1216]

How are your speakers coming along? We have another gorgeous weekend ahead of us, I think!

[bimmerguy]

Well it certainly has been a long time since I posted about the speaker build. I had a bit of a hiatus from the build due to...lots of things. My mom was hospitalized with pneumonia for a while, my beloved German Shepherd Karl died, my oldest human started kindergarten, and fall is a crazy busy time for yard work for me with the garden, massive amounts of leaves to pick up, and getting ready for the icy grip of winter. But I am back at it and making a good deal of progress. I try to catch you all up...

[bimmerguy]

Since I've chosen to make the front baffle in two separate pieces - an inner and outer - I needed a way to attach and then separate the pieces and still have perfect alignment. The typical way to do this is to use dowels that fit into fit into holes on the inside of the outer baffle and the outside of the inner baffle. So I started by making some precisely-dimensioned dowels that snugly fit into the holes but aren't so tight that you damage them getting them out.





I used my homemade dowel centers to transfer the hole locations from the inner to the outer baffle. This worked a treat, and I was then able to use the divots from the dowel centers to drill the mating holes on the outer baffle.



With the dowels inserted I can assemble and disassemble the two baffle pieces repeatedly with perfect alignment. I marked out the centers for each driver (and port) on the inner baffle and drilled a small pilot hole. Then I assembled the two baffle pieces and used the pilot hole to drill through the outer baffle too. Now I can route out the circles independently on both pieces and have confidence that everything will line up when I put it all together.


Then it was time again get out my routers and circle jigs and start cutting out about a million circles and recesses for the drivers/ports.

[bimmerguy]

I cut out all the necessary bits on the outer and inner baffles, including the recess for the trim rings I made earlier. I was surprised by how stringy the edges where. My router bit is fairly sharp, so I can't chalk it up entirely to a dull cutter; I guess the maple plywood is just really stringy. Anyway, the outer bafffle turned out pretty good after quite a bit of clean up with a chisel and sandpaper.




Next I cut out the notches for the little ears on the side of the tweeters where the wire connections are made. I used the mill for this part. I used a coaxial indicator to make sure I was perfectly centered on the tweeter hole each time since I had to mill the notches on each piece separately without any common reference other than the hole itself.





The finished notches look pretty good.

[bimmerguy]

Now is as good a time as any to relieve the back of the woofer and mid holes. The edges and dados do allow for a lot of room for a 45 degree chamfer bit to go to the depth I would like - if I made the chamfer deeper and wider it would interfere with the joint to the braces and it would cut through edge. So I started with a 45 degree, but then moved to 30, 22.5, and 15 degree chamfers to relieve as much material as I could without getting into the dados.


I wasn't too worried about getting too thin for the driver mounting screws. I only relieved the back of the inner baffle; there is still the 1/2 inch outer baffle on top of that. Here's what it looks like with the two baffle pieces together.




This was also a good opportunity to see how well the notches lined up for the tweeter. And now, finally, I can get to gluing things up. We'll get to that next.

[bimmerguy]

I decided to glue things up bit by bit so I could more easily ensure everything turned out square. The TM section of the speaker was the most involved since there are three pieces that connect together. I used a couple 123 blocks to help make sure everything was nice and square. I used quite a few clamps since I had to apply force in a few different directions.



Once that glue was dry I did the other internal braces. In addition to the 123 blocks again, I used the rear piece with its dados to make sure the spacing was right at the rear of braces. This also helped keep them square since at that location their just hanging out in space (for now).



For the bottom piece I used an angle plate along with the rear piece to get a nice square angle.



After that, the rear was easy to glue in place - just had to make sure the angle of the piece lined up with the curves of the braces.



Rinse and repeat with the other speaker and the skeletons are complete.

[bimmerguy]

Time to glue on one of the side panels. I thought quite a lot about how best to do this. I repeatedly tried clamps and ratchet straps, but I could never get good even force across the entire curve. Then in a monkey-touch-the-monolith moment of genius (not) I remembered the Curvomatic. All I'd have to do is make the new 'formers' that were the correct radius - this time concave instead of convex like before. To make the template I assembled the panel + a sheet of hardboard + the Curvomatic 'skin' to get the final curve radius and traced that onto a piece of hardboard. I cut it out fairly accurately on the band saw and sanded by hand to get exactly to the line. I didn't use a circle cutting jig since I was able to get a reasonably smooth curve this way, and match the needed radius very accurately. Then I just used the template and my router table to duplicate the shape onto some scrap particle board I had lying around.




I really scratched my head about what angle to orient the formers. When I put downward pressure on the assembly I don't want it sliding down hill, but I don't want it tilted so far the other way or the top panel will slide off when I set it on top before I can get the straps tightened. In the end I ended up with a compromise between both extremes and I think it worked out pretty well.



I clamped a couple pieces of scrap to hold the skeleton so it didn't slide down. After a few dry runs I was confident enough with the method to glue on the panel. I used the same urea formaldehyde glue as I did to glue the panels -- because it is super strong and doesn't allow for glue creep, but more so because it provides a longer working time and applying the glue carefully to all those edges takes a while.




The bottom panel is there just as a place holder of sorts right now. When I go to glue on the second side, the panel will already be attached, so I needed to account for it now.

[Steve Manning]

Man you got the mac daddy version .... nice!

[bimmerguy]

The next challenge was to trim the panels to size. They were intentionally made quite a bit oversized, so there was quite a bit of extra to remove. I started with my track saw.



I wasn't sure at first how I felt about using the track saw on an angle like that, but it when just fine. The track saw got me in the neighborhood, but to get really close I used a flat-bottom bit on the router. I've seen folks make a jig for the router that mounts to or replaces the router base. But I didn't want to take the time or material to do that, so I took the lazy way route. I just double-sticky taped some scrap MDT to the front baffle for the router to ride on, then the router bit could overhang the end and trim the extra panel overhang. This was a little precarious on the rears as there just wasn't much real estate on which to tape the MDF pieces. Here's what it looked like:







This got me pretty close. The rest was done with sanding.



The overhang on the top and bottom was trimmed off with the bandsaw. I was a little nervous about that part - more about possibly chipping out the veneer than anything else, but it went fine. But that brings me to the embarrassing part...

[bimmerguy]

Call it rushing, or not being careful, or not paying enough attention, but I committed the rookie mistake of sanding a little too...vigorously on the top and sanded through the top layer of the plywood veneer (fortunately not the ziricote veneer!).



Oi! This will end up being covered by the ziricote veneer, but I can't have any low spots - they'll be easy to see once the shiny finish is applied, and I don't want to risk the chance that the veneer might not stick completely in that area. So I decided to crack out the glazing putty. It's sort of like Bondo, but finer - it feathers out a lot nicer than Bondo, and sands a lot nicer too. I used this same glazing putty when I was painting my truck and I really liked it.



After some careful sanding with a long block, it is perfectly flat again. I checked it with a straight edge and feeler gauge.



That's the beauty of glazing putty - you can really taper that edge to nothing. But as soon as I feeling good about myself for fixing a stupid mistake, I realized that I might have just shot myself in the foot by putting down a surface to which I can't glue the veneer. Obviously paint sticks to glazing putty, but I didn't know if glue would. So I did a little experiment. I took a couple scraps of plywood and covered then with glazing putty, then sanded them perfectly flat.



I applied different glues on each piece (Titebond III and Better Bond Xpress). I left a little extra space on the samples in case I wanted to try out more glues (like clear epoxy). Then I applied the same glues to two different veneer sample pieces and let all the glue dry thoroughly.



I used a hot iron to then bond the veneer on each sample. Surprisingly, that worked great. The bond seemed very strong. But the proof is when you use a flush trim bit in the router to trim the veneer to size. That worked great, too! So apparently you can glue to glazing putty. Who knew? Well, I am sure some of you did...



It looks like there are a couple chips out at the corners, but that is due to the corners having a bit of the putty chipped off at those locations; so it wasn't perfect flat right there. Anyway, I am pleased that I could correct my mistake.

[bimmerguy]

The internal volume of the speakers as-built is slightly larger than my nominal ideal calculation. So I needed to take up some space. I happened to have some 1/4 inch Baltic birch cutoffs left over from making the panels that turned out to be exactly the right size. I decided to make some blocks from those cutoffs, to glue on the bottom of the interior, that fit shape-wise with the curved panels. This would have the double added benefit of providing a little more weight on the bottom end (they are a little top-heavy as-is with the extra structure for the sealed TM enclosure), and provide extra thickness into which I can screw the bases (although truth be told I am going to use threaded inserts so the extra thickness isn't really needed).

I started by making a new template. To define my outer edge I used a cutoff from trimming up the panel I'd already glued on. I couldn't just use my original template since there was just the slightest bit of springback in the panels. All that accuracy is probably unnecessary, but if you haven't figured it out yet, I'm a little anal retentive when it comes to building stuff. Anywho, I then used a compass to draw a line a bit inside my outer dimension.



Then I transferred my paper template to an obsolete template I had laying around from the braces, cut it out and sanded the edge smooth. I needed to laminate the 4 pieces of Baltic birch together to make one thick block. I did that with my vacuum press.



Then I cut them out and template-routed them to the final size.





When I went to glue them in, no matter what I did they would slide around a bit. There was such a large surface area with glue on it that it really wanted to skate around. So I finally abandoned clamps and just went with something heavy. Like a 4 inch diameter bronze rod. That thing must weigh at least 80 pounds. It's about as much as I am comfortable carrying around with one hand/arm.


With the new pieces glued in, my calculated volume is now within 2.5 cubic inches of my nominal ideal. Not too shabby.

[theSven]

Nice to see some updates posted here as the project progresses. Are you getting anxious to listen to them yet :-)

[bimmerguy]

I totally can't wait to listen to them! I know I have quite a bit left to go (and I won't be able to apply the finish - automotive polyurethane clearcoat), until the weather warms up again next spring/summer, but I'm already imagining how they will sound. I'll need to learn a lot more about how to measure properly so I have an objective sense of how they're really doing. I haven't done a lot of research into that yet, but I will.

I currently have a pair of Energy Veritas bookshelf speakers on stands as my main speakers (don't laugh), and an Energy sub. All of which are about 17 years old. They sound OK, but not great. I was in a professional wind ensemble for years, and I have a pretty good ear, and the Energies do fall quite shot, so these speakers will be a big upgrade. I have an old pair of Paradigms (first speakers I ever bought back when I was in high school) in my shop that actually sound better, but they're a little dusty. :-)

I'm a little concerned the room itself will be the biggest detractor from the new speakers' performance. It's a tough room, I think. It's got a ceiling that vaults to 14 feet, large windows on an adjacent wall, and it's open on the other adjacent side. Sort of hard to describe, but think of a box where the left wall and rear wall aren't there. It's a big space and weirdly shaped. But I don't want to get too ahead of myself. I haven't even thought about the design for the bases, or the grills. I'm not crazy about having grills, but I think it is a necessity since I have a 4 and 6 year old. I know they'll be careful around them, but I'd rather be safe than sorry. I'll probably just use the grills for 'storage' and remove the grills for listening.

[bimmerguy]

I needed a way to mount the flat crossover boards to the curved panel surface. I suppose I could have just screwed them in at the corners, but that wouldn't be in keeping with my needlessly complex approach. So I decided to make some mounting pieces that are flat on one side for the crossover boards and curved on the other to match with the panel. I started with a scrap piece of the Appleply. I was careful to mark the center and edges for laying out the arc. You can see how I used a parallel and angle plate to do that. With those three points marked I could ensure that, after the arc is cut and I run it through the table saw, the piece will be perfectly symmetrical about the center point - that is one side won't be thicker than the other. I might be going into too much detail about how I laid it out, but those are the types of things I think about constantly when making a project like this.



After that I could rough cut the arc on the bandsaw, and template route it to the final curve. I used the template I made for the braces since that had been made to perfectly fit the panel curve.



Here they are all done and test fit into the speaker. There are two different crossover boards. The smaller one is for the woofers, and the larger is for the mid and tweeter, hence the two sizes of mounting pieces.





Next I needed to drill holes in the crossover boards for the mounting screws. I could have just used the holes in the pegboard, but they were a little too big for my liking (I would have needed a much larger washer or bigger screw). So I drilled new holes that were sized appropriately for a No. 8 screw. I had to be careful not to nick any of the components while drilling, but it all went fine. I did the drilling on the mill because, I wanted the mounting holes spaced a precise distance that I could then replicate on the arced pieces; and well, because I like using the mill.



I'm not too crazy about the crossovers being mounted in pegboard; but Rick assembled those and they are done well with point-to-point wiring using nice components., Obviously the performance will be just fine, I'm just being picky. Few are as anal retentive (read: annoying) as me about the little details, even ones you can't see. Anyway, then it was on to the mounting points in the arced pieces.

[bimmerguy]

I didn't want to just use wood screws to mount the crossover boards (I will refer you back to the previous 'needlessly complex' statement). So I made threaded brass inserts into which I could install a No. 8 machine screw. The first thing was to drill the holes to accept the inserts.



Then I turned down some brass rod to the right diameter and knurled it. The knurling gives a good mechanical bite in the wood and gives the superglue recesses to hold on to.





Then I cut them to length, and drilled and tapped them.





And here they are all done.



The inserts were a tight enough fit that I needed to press them in. So I used the drill press like an arbor press (my arbor press was buried in another part of the shop and was all greasy). I put a flat 'plug' in the drill chuck so I had a flat surface to push with. And since the pieces are curved I used a cutoff from one of the braces to match the curve and ensure that I was applying force straight down. The thin CA glue sets up really quickly so you have to be fast about pressing it in after you apply the glue.



The bottom of the crossover boards have zips ties all over, so I had to mill out recesses for those so the boards would sit flat on them.



With all the recesses cut, the boards fit nice and flat.



And here they are all glued down in a speaker.

[bimmerguy]

Before I glue on the other panel of the speaker, I need to install the Sonic Barrier foam. I won't really be able to do it after the speakers are fully assembled. This is pretty straight forward for the assembled part of the speaker, but I needed to lay out the other panel so the foam wouldn't interfere with the glue locations. So I set the assembled part of the speaker on the panel, ensured that it was all aligned correctly, and traced the inner edge of all the spaces. Now I had a good diagram of where to stick the foam.



A quick note about how the panels were aligned - the 4 panels are essentially all a bookmatch of each other. So I was very careful to ensure the seams/edges in the veneer (there are 4 joined sheets per panel) were located at exactly the same place on both sides of the speakers. That is, if there was a seam 4 inches back from the front surface on one side, the seam needs to be 4 inches back form the surface on the other. This way the veneer will be symmetrical on both sides of the speaker. This also goes for its vertical orientation.

With all that layout done, I moved on to cutting the foam. It was quite a pain to cut. A utility knife blade was not quite long enough so I found a slightly longer blade. And the blade had to be razor sharp in order to prevent any tearing. Even so, it didn't cut easily. I think I spent 2 days cutting all the foam. I'm glad that's over.

[bimmerguy]

After cutting all the sonic barrier pieces, it was an easy task to just peal and stick 'em. But I didn't really trust the adhesive backing, so I added some pieces of VHB tape to the edges/corners where the pealing stress would be the greatest - just as a little anti-pealing insurance. It might not be necessary, and it might not be helpful, but I had the tape so I figured I use it.



I put some scrap pieces of foam under the crossovers. That's probably not needed either, but they were just going to go in the trash otherwise. Should I put some stuffing on the top of the crossovers? Or is that overkill/not needed?



The foam sticks up a little higher than the mounting surface of the crossover boards, but it smooshes down easily enough. I also cut some bonded polyester fill for the MT section. Not sure if this is the best material (Acousta-stuf and wool is quite a bit more expensive), but I figured it would be easy enough to swap out later if I don't like the sound. The bonded material is nice since it is a consistent density and it's easy to get a reproducible amount just by cutting it to size.



I made the internal wire runs nice and neat. Should I have been careful to keep all the lines separate from each other? I figured positive and negative run right next to each other in lots of external speaker wire, so it is probably ok inside too. I did not take great care to run the mid wires along a different path than the tweeter wires, for instance. Do you guys see any problem with that? I can still move the wire around if needed once the other panel is glued on.

So that brings us to gluing the second panel in place. I made new formers for the Curvomatic out of some scrap particleboard. That wasn't the best choice of material - it really wants to crumble under the pressure of the straps. But it worked...barely. If I had to do more than two speakers I would have needed something a little stronger like MDF or plywood. At least now I have a little less scrap particleboard lying around.



Next comes the worse part of the entire build for me. Trimming the panels to final size. The cabinets by this point are quite heavy and hard to move around through a router table; and they are now about 3/4 inch too tall to fit threw my big bandsaw. The way I trimmed the first panel wasn't the best, but at least then they were significantly lighter and easier to manipulate. I'll do some head scratching about how to do it easier. Maybe a router jig of some sort...

[bimmerguy]

Boy, 2022 was a tough year. The day after Christmas I got a call at 7:15AM that my mom had passed away. My family was just about to leave for a little after-Christmas trip when I got the call. What ensued was a whirlwind of activity and a week of 18 hours days planning a funeral and making other preparations. It's been a tough few weeks. But I really want to get back to some semblance of normality, and get back to making these speakers. I had to set them aside again with all the business of the Christmas season, and then the funeral, but I'm picking it up again.

[bimmerguy]

After the other side panels were glued on, it's time to trim them flush to the edge of the cabinet all the way around. I used my track say to remove the bulk of the material. I had a cutoff from one of the panels I'd glued on previously, so I was able to use that as a straight edge for the saw.



I could then use a flush-trim bit in the router to clean it up, followed by a little sanding. The long vertical edges required a bit more effort. I cut the bulk of the material off with the track saw. I thought it would be a little sketchy to clamp the track on a curved surface, but it worked ok. But I only dared to get within about a 1/4 inch with the saw. To get closer I once again needed the router. I made a trimming jig that I thought would be helpful.



Turns out it wasn't as helpful as I'd hoped. It worked ok, but in the end I preferred to just clamp a straight board on top. It functioned just like the jig, but being stationary it seemed like the router was more stable, especially on the narrow back side of the cabinet. And I didn't have the issue of sliding the router over the areas where I'd taped over the driver holes.


In the end it took a fair bit of careful work and a fair bit of sanding to get it perfect, but they turned out ok. The one downside (for now, anyway), is that they are HEAVY. They're not that big, but they are becoming quite a workout to move them around the shop.

[Browncoat]

So sorry to hear about your loss. It's never a good time for that kind of news, but the holiday break must be even harder.

Beautiful veneer work. I've got some zircote in my stash, and it's super chippy. Your work looks flawless.

[bimmerguy]

DISASTER!! I've worked with a lot of veneer in the past, but never with ziricote. Parts of it are very dry and not at all resinous. Other parts are extremely oily. That oiliness came back to bite me. The panels and their veneer are glued together with UltraCAT plastic resin glue. That is super strong stuff. But apparently it can have problems with ziricote. I should have known and done a little better prep by wiping the veneer down with acetone before gluing. Live and learn. Adhesion was generally good, but there was one spot where the veneer bubbled up a bit. I tried to cut a tiny slit and drip some thin CA glue in there. And that worked to a point, but it was difficult to get good glue coverage under the veneer without damaging it further. Because the veneer bubbled up, there was an extremely unfortunate sanding-through incident. I tried to fix it with black CA glue, but that just seemed to make it worse - I was chasing my own tail. I'd get one area, but then the adjacent area had problems. In the end it looked like this:



Not so bad? Have a closer look:



Oh boy. It's toward the back of the speaker where it won't be seen unless you specifically look for it, but I still can't leave it like that. There's only one solution as I see it. I have to cut out the offending section and replace the veneer. Delicate surgery, indeed. So as is the answer to many a woodworking challenge, I made a jig.

I needed a way to cut to a very precise depth along a curved surface. I started with making two arced pieces I could attach to the bottom of the router, essentially making a base whose curve matches the curve of the speaker cabinet. Here you can see me drilling the mounting holes on the mill.



That worked well enough. But now I needed a fence to ensure I cut a perfectly straight edge. Setting the curved piece on the speaker, I was able to get a pretty accurate guess as to the angle I needed to allow the fence to run flush with the rear of the speaker. I cut the slot to receive the fence piece on the mill.



With the fence piece installed it blocks two of the mounting holes. But it's plenty secure with the other two - I'm not putting a lot of force on it, and it is a one-time-use jig.



Here's how it looked just before use. I used two cutoffs from the curved panels as stops on the left and right of the cut.

[technodanvan]

I'm sorry to hear about your mother, the whole situation sounds very similar to when my dad passed away.

Your work, of course, looks as professional as ever!

[bimmerguy]

So sorry to hear about your loss. It's never a good time for that kind of news, but the holiday break must be even harder.

Beautiful veneer work. I've got some zircote in my stash, and it's super chippy. Your work looks flawless.

Thank you for the kind words.

Yes, ziricote is very chippy, very brittle. Even cutting it to rough size with the utility knife and veneer saw was a nerve-wracking endeavor. It's a lot different than working with, say, bubinga veneer - which is super hard, but not nearly as prone to cracking. It's not nearly as fragile.

[bimmerguy]

I'm sorry to hear about your mother, it sounds very similar to when my dad passed away.

Your work, of course, looks as professional as ever!

Thank you!

[bimmerguy]

The jig worked surprisingly well. The edges were nice and crisp, and the depth was perfect. It had all desirable qualities...except I didn't like it. Even if I selected a piece of replacement veneer and cut it flawlessly to fit, you'd still see a seam.



So after all that stress, I decided to cut away the entire sheet of veneer up to the bookmatch. I am certainly a glutton for punishment. So back to the drawing board with the jig. Well, almost. I needed to modify it to cut further up on the curve and add a new fence. So I milled off the old fence and cut the notches for a new one.





This time the fence ran along my track saw track. That way I could more easily tweak the cut line so it was perfectly aligned with the bookmatch seam.



It worked. But not perfectly. To use it I had to try to keep it perfectly parallel to the the track, which is easy when you're right up against it, but I needed to cut away all the material below it, too - which is a lot of passes with a 1/4 inch bit. Any deviation from parallel to the fence and you risk gouging a little into the surface. And that's a bit of a problem since the replacement veneer must sit flush with the existing veneer, and have a flat surface to bond to.



But after a little glazing putting and some sanding, it turned out to be a great surface on which to lay the new veneer.

[bimmerguy]

I sorted through the remaining veneer to find a piece that would work with the bookmatch. It's not flawless since the original bookmatch pieces were sequential and this one was several (maybe 4 or 5) layers down. But it's good enough that I don't think it will be too bad. I did the iron-on thing since there was really no good way to clamp the narrow piece in place. I wasn't sure how ironing would work on a curved surface. I spread a layer of glue (Xpress veneer glue - no need to buy the glue specific for ironing) on the veneer and the cabinet and let it dry thoroughly. Then I spread another layer of glue for good measure and let it dry, too, about a hour or so.



Overall I am pleased with how it turned out. Not an absolutely perfect bookmatch, but it is a good repair and the seam is good. There was a high potential for this repair turning out to be a complete catastrophic nightmare, even worse than the original issue. But in the end, I am pleased with it.

[Steve Manning]

Sorry for your loss. My mother also passed away last year, always a tough time.

[theSven]

Condolences for your loss.

I would classify you as a veneer surgeon with the work you did fixing that boo boo.

[bimmerguy]

With the veneer snafu fixed, I could move on to veneering the backs. This was a more straight-forward affair. Since I couldn't put the whole speaker in the vacuum bag, I did the iron-on thing again. I put a coat of glue on both the case and the veneer and let them dry for about an hour. The veneer wants to curl a fair bit when there is wet glue on one side, but with a piece this narrow it is manageable.



Then I could just iron it on.



I was torn a bit about how to best trim it to be flush with the edge. I couldn't use a router with a flush trim bit sides are curved. In the end I decided to just use a sharp chisel and essentially whittle away the edged until I got reasonably close. I then just sanded down to make the edge flush. I had to be quite careful that I didn't sand through the thin veneer, but it turned out all right.



The next challenge was cutting out the holes/ports where the speaker wire connections will be. I started by using a sharp knife to cut a rough hole. It was fairly easy to tell where the hole under the veneer was.



I then used my little router to free-hand cut out a close as I dared to the edge. I made sure to expose the threaded inserts I'd made to screw in the panel.



Then I could screw on the panel - just setting on top of the veneer, and scribe to the exact edge. I then removed the panel and used a sharp chisel to cut away the remaining overhang. There was no margin for error - if I cut over too far or if it chipped out there would be a noticeable gap between the panel and the veneer in which it is inset. Definitely not a design for manufacturability (a bezel to cover the gap would allow for more slop), but I like the look and this is a one-off so I can afford to do little things like that.



It's not pristinely perfect, but from more than 6 inches away it looks pretty good.

[bimmerguy]

Changing gears a little bit (for variety), I decided to tackle the fit of the aluminum trim rings I'd made earlier for the drivers. The rings for the tweeters were a bit too tight of a fit in the recess in the outer baffle, so I modified the OD jig I'd made for the router so I could mount it on the lathe. That was the only way I could think of to take off just a few thousandths evenly from around the entire circumference. I cut a groove slightly thinner than the thickness of the aluminum and the same ID. That way I could mount the ring on the wooden block/disk, and it would ensure the ring is perfectly concentric. Then it was an easy task to just turn off the needed amount.





the tweeter and mid rings are close enough together that they interfere with each other, so I needed to modify each of them so they'd fit in their places. I could have cut a arc out of the mid ring for the tweeter ring to nest in it. I'd then not have to alter the tweeter ring. But I decided it would look better (not sure if that's true) and a fair bit easier to have them meet at a flat line not an arc. I set the tweeter in place and then the mid ring resting on top of it and scribed a line where the two met.



I place the rings between two pieces of thick plywood and then filed down to the wood. The plywood provided a good safety stop for the file so I didn't over-file the rings and it also acted like a fence to make sure the line was nice and straight.



With them both done (for both speakers), they look pretty good.

[technodanvan]

I was just thinking about this project the other day. The rings look great!

Edit: I just noticed you didn't veneer the bottom of the speaker. Do you have any plans for finishing that?

[bimmerguy]

Edit: I just noticed you didn't veneer the bottom of the speaker. Do you have any plans for finishing that?

The top of the speakers will have veneer. More on that saga shortly! The bottoms will not. I am going to paint them black I think just to seal up the wood. I'd thought about veneering the bottoms (and I still might), but I don't have a ton of the ziricote veneer left and I need to have enough to do the center channel (that I haven't even designed yet) so it will match. And finding ziricote veneer that even comes close to looking similar will be like finding chicken teeth. So I am trying to conserve the veneer a bit. The subs that I build after the LCRs will likely be a different veneer entirely, both by necessity and by choice.

Anyway, the speakers will sit on either bases or outriggers (I haven't quite finalized the design for those yet) that will raise up the speakers about 1.5 - 2 inches. I think I'm leaning toward outriggers since I'd like a more minimal approach that doesn't distract visually from the speakers themselves. It seems like a large bulky base might be visually distracting. Not sure if I'll do spikes or not; the speakers will be sitting on a carpeted floor with a plywood subfloor (e.g. not on concrete) when they're installed.

[bimmerguy]

Well, on to the top veneer saga! Or shall I call it disaster #2? Getting all sorts of cocky about using the iron-on veneering method I decided to use that technique to glue the veneer on the tops of the speakers. So I cut some nice pieces of veneer and did the bookmatch thing like I've done a million times. Like this - x2:



Then I proceeded to apply the glue to the veneer and to the cabinet as I've done before. And that's where I ran into my first problem. You'll recall how on the back pieces the veneer curled a bit when the wet glue was applied. Well when the glue dried those pieces straightened out a bit - enough that I could easily iron them down. With the top pieces...not so much. Here's how they looked after a couple minutes with the glue on. I set a can inside the curl to keep it from doubling over on itself.



After the glue had dried they went full-on cinnamon sick on me. Check this out:


That one was the worst of the two. It cracked in half when I tried to straighten it out to iron it on (the heat does make it more flexible). So....back to the stack of veneer I went to make another bookmatched piece. Why I thought the new one wouldn't curly up I don't know. I put weights on the corners as it was drying and that did help enough so that I could at least lay it out on top of the speaker. So I went about ironing them on. The first one went on seemingly pretty well. But I had trouble getting the second one to stick down properly. Maybe that veneer glue isn't so great for ironing after all...

The problem wasn't that the veneer wouldn't stick at all, it's that it _nearly_ stuck. Bubbles everywhere. I worked for hours fighting those things. I've gotten pretty good at making a tiny (barely perceptible) slit in the veneer with a very sharp blade and using some thin CA glue to get through the slit and pressing it down to get the bubble to stick. That works to a point. But not for larger areas and not when there are a thousand of them. After all that time and frustration, it turned out to be a total disaster.



It looks even worse in person and it's not salvageable. I had no choice but to completely sand off the offending veneer and its glue and start over. Ugh. So much time and material wasted - mostly time. At first I tried using a router with my flush trimming jig, but that wasn't working well, so I switched over to course sandpaper. Sanding through the glue was tough, but I got it to a manageable state.



Well I wasn't going to make that mistake again. I don't have a good way of clamping the veneer to the top, but I needed to figure something out. I decided that all I could really do was to place a heavy weight on top. So to help distribute the weight evenly, I made a caul on which I would place heavy things. I made it out of the wrong material for every application, particleboard. I traced the brace template I'd made earlier, using a compass to give me a little extra area.



It's a nice fit on the top with just enough overhang.



I went back to the veneer pile again and made two more bookmatched pieces. I actually like these better than the first batch. They match either other more closely since they came from adjacent sheets in the flitch. Now I could glue them in the more traditional way. I used my granite surface plate, a large steel cube, and a large bronze cylinder for weight. It's all very heavy, but I weighed all the bits and calculated out the pressure and it wasn't very high - only about 2.5 PSI. Which sounds small, but I'm happy to report it worked a treat. The veneer is dead flat and strongly adhered to the tops.



When the glue was dry I trimmed the bulk of the extra with a big scissors and followed up with the router. Before trimming you can see how the veneer still really wanted to curl where it wasn't pressed down.



And after all that: