The Anvil 2-Way Active Accuton Bookshelf Build Log

[ohaple]

Since I am coming close to completing my Anvil bookshelf speakers and custom active crossover/amp module I thought it would be a good time to generate a build log. The last remaining step is reassembly and testing after the polyurethane dries.

This is a long post. I am sure, like with all DIY designs, there is room for improvement. This is the first speaker I have made, but I have the benefit of experience drawing from my father. The goal here was to build a set of bookshelf speakers focused on stereo listening at a desk, but capable of being used in a small living room later. I began by looking at transmission line designs and made a full design. The issue was, for the size, my measurements showed more bass extension was available from the same sized cabinet if we used a more traditional bass-reflex design (since a small bookshelf TL only allows for a small driver). So I scrapped the TL idea and went to work designing a bass reflex style speaker.

We chose the Accuton C158-6-85 for its good mid/bass characteristics, fairly flat response, known bass reflex design, and because we have been impressed with all of the Accuton ceramics we have used. The tweeter was a little more of a challenge since we have minimal baffle room and wanted to cross a little lower. Due to cost, the Accuton ceramic tweeters weren't an option. The Scanspeak Illuminator R3004/6020-10 was chosen since it would crossover well with the Accuton, fits the baffle, has good frequency response, and also matches the visual style of the Accuton.

I started by making a 3d design. It was important to make the design no greater than 15" in any direction so that I could cut the parts on the CNC machine and some on the laser cutter. Since I was planning to use a CNC machine, I wanted to make a design other than a rectangular box. The cabinet will be made using slices of birch plywood with walnut veneered sides and a maple veneered removable baffle. These will use an active external crossover so I did not leave room for the crossover internally.

I used SketchUp to help visualize the design in 3D and to confirm that the volume of the box is the 15L that Accuton advertises gives extended bass. Accuton advertises:

Vb: 15L, Port diameter: 40mm, Length: 100mm, Fres: 43Hz, F-3dB: 41Hz, Q: 0.50 (extended bass)

Using SketchUp's area tools I calculated that this design will have an internal volume of almost exactly 15L (prior to installing drivers, which will decrease the effective volume). The port chosen also exactly matches the recommended length and diameter.





I then designed the slices to be cut. The walls were made to be just over .5" thick all around. In order to waste less material I divided each slice in half, so they can puzzle together and be joined by dowels.



Everything started out exactly as planned. I would need ten slices per speaker, meaning 40 parts in total. I could get 8 parts from each sheet of material. We don't run very aggressive feeds and speeds so each sheet took about 35 minutes of CNC time.




We ran into trouble when we realized that some of the birch ply we purchased was defective. Though we bought cabinet-grade plywood, a portion of the sheet was not properly glued. A few of the parts delaminated completely and we needed to recut them. When we built a much larger set of speakers in the past this was not an issue, even using the same supplier.



After a full day of cutting we got all of the slices done. The puzzle pieces were a good tight fit so they were pounded together. You can see in the third photo why I named it the Anvil.





I inserted the dowels, applied glue, and clamped. The dowels were a tight fit so alignment was pretty easy. The dowel construction makes them very strong. Both inner sections glued up and insides sprayed with shellac to seal. I did not make the insides pretty since they will be covered in material and wont be visible anyways.




With the main "C" underway it was time to work on the baffle and sides. We veneered one sheet in maple for the baffles and two others in walnut for the sides. The sides were cut face-down to allow us to mill a pocket in for the baffle to sit and pocketed holes to allow for hidden dowel alignment.





The baffles were designed so that the drivers and port would mount flush to the surface.




The delamination began showing up over time on parts that previously looked like they were solid. These cracks would be sealed with a glue-sawdust mixture on the outside and gorilla glue on the inside. The boxes are very stable even with the filled cracks, in part due to the dowel construction.




After getting the cracks under control I sanded the boxes down smooth so they would be ready for finishing.



Now that the boxes are in their finished shape, we are ready to drill holes for the removable baffle inserts and binding posts. Rather than eyeball it, we cut drill jigs using a laser cutter.



Now time to cut the nameplate/binding post plate. The design was made using the same spacing as the drill jig so the holes would line up perfectly. The design was then laser cut and engraved from 1/8" Walnut



First time test-fitting the drivers and port. Everything fits like it should and looks good together.

[ohaple]

I then put a base coat of polyurethane on the entire box to help seal it. This won't be its finished look (it will have a thicker and glossier finish when done), but will be enough to seal it while it is hooked up to test how it sounds before dedicating more time to finishing.



After allowing the first sealing coats to dry, the inside was dampened. The sides have not yet been done, but will be. The material is glued in with contact cement.



Capacitors wired in-line to protect the tweeter.



Drivers temporarily mounted and wired.



Without the final finishing, but otherwise assembled. Its satisfying to finally see them assembled after several weeks of building.




Initial testing to make sure the speakers perform like we want prior to doing some of the finishing work. Since this is an active crossover design and the amplifier/crossover box still needs built, they are just loose boards for now. The speakers sound great. The port may need some tuning. It is currently setup to Accuton's recommended length/diameter for this volume. The crossover at 2200hz works great for these drivers.



Hooked up in the office. Overkill for a desk, but they sound great. The two-foot design allows them to be set half on speaker stands and half on the desk. They would be better suited to a deeper desk so they aren't quite as close to the listening position. The anvil shape makes them deeper than needs be for the given volume, which makes them harder to fit on this shallow desk. They would fit in well in a larger room as well. For our current living situation they will be at this desk, but may be in a living room in the future.




SPL of the left and right speakers measured individually from about 18" with only about 10 minutes of play time. Our hope is that we will get a little more bass response with some port tuning, room gain, and break-in of the drivers. There are no audible phase issues and they measure rather flat from about 150hz up. We will conduct room-style measurements once setup properly.





Now that we know that the speakers are working properly and the crossover module is set at a good point (2200hz), its time to get the crossover/amp boards mounted in a box. For this project I am using the Xkitz XAMP-M2 This is an initial prototype box. The inputs will come out of the front since the RCA inputs are mounted directly to the board. The power from the power supply and the outputs to the speakers will also come out the back. The prototype box was made from 1/8" ply, which is the same thickness as the acrylic I will use in the final.

The final box was cut from clear acrylic on the laser cutter. Labels for the inputs and outputs were engraved into the acrylic as well. Here it is all wired up. The 12 gauge we had on hand is a little overkill for such a small box but it worked out fine. The sides and bottoms were glued together but the top is just a tight friction fit. The switches are disconnects for the left and right speakers individually. The disconnects were intended to operate like anti-thump relays, but the thump still is there when the switches are flipped on after the amp powers up. The thump from the amps on their own isn't terrible but I would prefer to avoid it if possible. I revisit this issue later in the project.

The scored text was all filled with blue paint to match the boards and wire. It is subtle, but enough to allow you to read what you are doing.








Here is how the amp looks on the desk. The power supply is the grey box underneath. The Macbook is fed into the Emotiva DC-1 via USB. The Xbox is fed into the Emotiva DC-1 via optical. The DC-1 operates as a DAC and preamp, as well as a headphone amp for my headphones when speakers aren't an option. The Emotiva sends the RCA out to the crossover/amp. Unfortunately this ASUS monitor is causing noise in the Emotiva which is fairly loud. If the Asus is unplugged the system is silent. I may end up replacing the monitor if I cannot find a way to fix the problem. The other monitor and equipment causes no problems.



For perspective with where the speakers are in comparison to the amp and how they look approximately from the listening position. Now besides working on the anti-thump the amp is done and the speakers are ready to be finished. They have been sealed but need several more coats of polyurethane to give them a nice shine and I need to figure out some isolation pads for the speakers to sit on.



The interior walls still needed felted and the washers for under the baffle and tweeter. I used the vectors from the cabinets to laser cut the felt for the interior so it is the perfect shape, and similarly cut the felt washers.



The felt inside fit perfectly and was glued using rubber cement, which makes a nice clean interior.




Now back to wood finishing. More spray to get the cabinets glossy. After being sprayed and dried, they will be rubbed down with steel wool and then final coats will be hand wiped.



Now that the finishing process in under way its time to solve the amp thump issue. I wanted to avoid tinkering with my expensive drivers so I went to the thrift store and got this old speaker for $0.25. It actually doesn't sound that bad and will allow me to test out my new solution without risking my drivers. I cut off the speaker fabric to see what was underneath better since I don't intend to keep the speaker.



The idea is to use a 2-way switch with a center off position on the amp rather than the on/off DPDT i originally used. One of the on positions will be wired to a pair of dummy load 8ohm resistors, while the other is wired to the speakers.

Rough mockup of the circuit (values listed under components weren't changed to be accurate). When the power supply is turned on, the amp will generate the differential but will be in the off position. When the switch is moved forward, the dummy load resistors will be placed in the loop and they should absorb the "thump". Then with the thump dissipated, the switch will be moved to the rear position putting the speakers in the loop with no thump besides the minor noise made by the switch connecting. In the initial tests, the process works perfectly. At a desk this works fine, but this isn't a convenient solution for a home theater style use.




Next steps are to finish the polyurethane, integrate the dummy load resistors into the amp box, reassemble, and tune the port. I will update as I make the final steps. As a first speaker build I have been very happy with the results. I went straight into the deep end with ceramic drivers and custom amplification but I think its been a good experience overall. To get to this point it has been about 45 days, with most work done on weekends. In a subjective comparison, I was hoping to make something that sounded about as good (or better) as the B&W 705, which I really enjoyed listening to at our local audio store. While I don't own the 705s to do an A/B listening test, I believe I have been successful in my goal.

Probably more fun than the speakers themselves has been the opportunity to drive a couple towns over and spend each weekend with my father working on this project together. I definitely benefited from his experience and tools.

I am happy to answer questions or hear constructive criticism about the project. I am particularly interested to hear about port tuning suggestions and about the anti-thump design. I will update with final photos soon when complete.

[howard]

Nice looking speakers! Very good write up.

[malboro2]

You are crossing them at 2200khz. And so by active you mean that you have a separate amp for both the tweeter and the woofer, with the crossover before the power amps. What is the rms power of each amp for the tweeters and for the woofers? What is the slope of the crossover? What preamp are you using?

All the speakers I ever built have been active, but I like to be able to fool around with the cross, and not digitally, I use an analog adjustable L-R 24 db crossover, a preamp, and separate poweramps.

[ohaple]

You are crossing them at 2200khz. And so by active you mean that you have a separate amp for both the tweeter and the woofer, with the crossover before the power amps. What is the rms power of each amp for the tweeters and for the woofers? What is the slope of the crossover? What preamp are you using?

All the speakers I ever built have been active, but I like to be able to fool around with the cross, and not digitally, I use an analog adjustable L-R 24 db crossover, a preamp, and separate poweramps.

Thanks howard, it was painful to tear them down after getting to listen for a couple weeks so that I could finish them. Looking forward to getting them 100% this weekend.

malboro2 , I will do my best to answer. I am using an analog active crossover set to 2200hz. L-R 4th order 24dB/Octave. 12W RMS into 8 Ohms, 21W RMS into 4 Ohms. It is the Xkitz XAMP-M2, which is convenient since it handles the analog active crossover and bi-amping on a single board at a reasonable price.

Each driver is separately amplified. The crossover comes before the amplification (though on the same circuit board). I also like to tinker with the crossover points, which the XAMP-M2 allows easily. They provide a calculator to tell you what resistors to use to convert the crossover module to just about any frequency you would want without re-wiring the actual board. Resistors are pretty cheap so its not much investment to tinker with the crossover point. Unlike DSP it doesn't allow you to control delay and other functions. It does have separate gain adjustment so you can match the drivers.

I am using the emotiva dc1 for a preamp/dac.

[ohaple]

I finished up the pair a couple of weeks ago and have been thoroughly enjoying them.
We wired the switches on as diagramed. Each amplifier circuit has its own 8ohm dummy load resistor to dissipate the thump.
The resistors were mounted partially under the boards, but not touching.
To make sure the switch is clearly labeled we made a custom label. Off is center position, resistors are to the right. Left is to connect the speakers.
We ran a test on the dummy speaker and it works perfectly. The order to power on is: Amp off- Power supply on - Switch to dummy resistor - switch to speakers. To turn it off: Switch to off position - power supply off. Its less convenient than the timed relays used by most, but it works to solve the on thump and off thump that I was getting before. Again, the thump wasn't enough to damage drivers likely, but I would prefer to avoid it when possible.






Now that the amp is done, back to finishing the speakers. The feet were covered in adhesive cork to give a little bit of isolation and make them slide less. I am still investigating a better isolation solution.



These weigh 18lbs. Many high end bookshelf speakers weigh this much or more. Its a very manageable weight to move around when needed.



Finished photos. Happy with the end result. The end-grain plywood was significantly harder to get an even finish on than the veneer.






Measuring from the listening position.



With the port (green), accuton (yellow), and scanspeak (dark green) from 2". This shows the 50hz hump that is also visible in the next graph. Overall I am happy with how these drivers perform. 1/24th smoothed.



This SPL reading from the listening position shows the significant limitations of where the speakers are currently setup. The biggest concern is the 90hz dip and 50hz hump. 1/24th smoothed. I am hoping that I will be able to put them in a better position when I move. In the future I am planning to upgrade the power supply and give better isolation to the feet.



My next project is going to be a 5.1 home theater setup with emphasis on low budget and small size for the living room. Here is an initial rendering of the surrounds using a Dayton RS100-4 driver. I am calling these next group of speakers the Ions since they are loosely based on the visual style of a TIE fighter (twin ION engine).



As always, I am happy to answer questions about the build and take ideas for future projects.

[dst]

Very interesting design and spectacular cabinet and detail of construction. Well done!!! As to your choice of drivers, I would say the Accuton drivers are probably in a league well above and beyond a basic chip amp, IMHO. I am sure they will sound very good, but if and when you run them through a higher end, higher power amplifier I think they will pay even higher dividends.

[ohaple]

Very interesting design and spectacular cabinet and detail of construction. Well done!!! As to your choice of drivers, I would say the Accuton drivers are probably in a league well above and beyond a basic chip amp, IMHO. I am sure they will sound very good, but if and when you run them through a higher end, higher power amplifier I think they will pay even higher dividends.

Good point and I generally agree. Similar three-ways we have built (accuton woofer and mid, raal tweeter) we run on the xkitz active crossovers with three two-channel parasound amps. From what I can tell, my bottlenecks are (in order): the room, the amp/power supply, the isolation pads, the source. Due to money, four channels of high-end amplification isn't in the cards right now. Our thinking was that if we are going to spend many weeks and dozens of hours on a project, the drivers should be top-notch. The amp and power supply are modular in nature and can be upgraded any time, but the drivers can't.

I also have been surprised by the quality of the chip amp used. While I don't have a clear A/B test to confirm, I believe that the active crossover with chip amp sounds better than an average passive crossover with a better amp. Power is also not an issue at all since these drivers are of about average efficiency and you don't have any of the losses associated with a passive crossover. The gain on the amp is about 60% and the preamp never goes above -20db before it is too loud for my room. All that said, my plan is to handle the bottlenecks in order. When I move and get a more proper listening setup I will be able to eliminate some of the room issues I am having. Then when funds allow, I will see about upgrading the amp and swapping the crossover. I might tinker with the minidsp solution. Last time we used it we found that the xkitx crossover sounded better, but minidsp would give me more control over my limited listening environment.

My favorite thing about these speakers is the imaging. While the highs are not as transparent as the Raal ribbon in the 3-way we built, the imaging is actually better (likely from the near-field listening position) songs with clear depth and instrument separation are very fun to listen to.

[dst]

Your logic is quite sound and obviously planned out and future proofed. I have looked at the X Kits offerings on their web site for the past month or so, and they look to be most excellent for the price. If and when you do upgrade your amp, you can simply use one of their x-over only modules. And my intent is not to denigrate chip amps, as some sound quite good. Hey, Jeff Rowland even put one in the market, right! Thanks for sharing your design. Great work.

[ohaple]

Your logic is quite sound and obviously planned out and future proofed. I have looked at the X Kits offerings on their web site for the past month or so, and they look to be most excellent for the price. If and when you do upgrade your amp, you can simply use one of their x-over only modules. And my intent is not to denigrate chip amps, as some sound quite good. Hey, Jeff Rowland even put one in the market, right! Thanks for sharing your design. Great work.

Thank you! We have liked all of the xkitz products we have used. It's a small operation but the guy seems to do a good job with support and focusing on quality products first. Sent from my SM-G950U using Tapatalk