Dipole Bass

Dipole Bass

I finally got around to replacing the sealed box Bass with a Phoenix type dipole bass. Wow! What a difference. I have been very happy with the "Ugly" subwoofer but I was using it up to about 50 Hz and extending the two Dynaudio 24W75 9 inchers down to 50 was great as long as I didn't try to play very loud. When really stressed, they protested loudly, sometimes bottoming with a great clacking sound. Now that I have a dipole bass that will easily work up to 100 Hz I changed the crossover frequency up to 100 Hz were it belongs and am much happier. My Dynaudios are much happier as well.

This is a good place to give credit were credit is due. The concept of this woofer system is entirely due to Siegfried Linkwitz. The W-baffle Box design is his, the crossover circuit design is his, and I give him complete credit for the great sound coming from these speakers. Thank you Siegfried. All of the information about how to design dipole woofers and the specifics of implementing them are on Siegfried's web site at Linkwitz Lab

I started out with four of the Lambda Acoustics SB-12 sealed box woofers. These worked really well but there were times that even these couldn't do all I wanted of them. So eventually I upgraded to the Adire Audio DPL-12. These have the low distosion and high x-max that I need! More on them farther down this page.

As most of you know, I don't like to build square boxes so I got some 2 inch radius quarter round MDF from AIT and made myself a box with rounded corners to match my main speakers. I built these a little wider than really necessary so I could have some flexibility with drivers. The inside dimensions of the cabinet are 13 inches high by 19 inches wide. Then I made a removable U shaped piece that actually holds the drivers. Removing a few dozen screws allows me to remove the piece holding the drivers, change out the drivers with ease and then slide the piece back in. This works better than I imagined and will make driver changes or upgrades a breeze.

This is the front of the speaker with the driver carrier installed.

This is a rear view of the speaker. The inside of the cabinet is just bare plywood. I applied a little paint for appearances but nothing fancy. I would normally build something like this from MDF. It is easy to work with, easy to finish and is fairly inexpensive. Plywood is much more problematic. It splinters, has huge voids between the plys and shows all the grain when painted. I only used it here because I had it. Normal particle board or MDF are much better choices.

This is the front of the speaker with the final paint job applied. You can see some light streaks on the inside of the cabinet were the paint has been scraped off by the driver carrier being test fit. The outside of these cabinets was treated to a layer of 4 ounce fiberglass cloth and several coats of polyester resin. This is the same finish as my main speakers and eliminates the problem with shrinking and expanding that plague cabinets made up of different materials. The final paint was misted on from a spray can and produces a nice "pebbled" finish. The complete story about this finish is back in the section about the main panels.

Here is the carrier with the drivers installed. I attached a brace at the open end of the U to give it a little more strength when installing the heavy woofers. This could be a piece of all-thread and then it could just be left in place.

Here is the carrier ready to install. The brace has been removed. These drivers are the older model 4-spoke Lambda sealed box woofers. They work pretty well but are limited to about 3/4 inch peak-to-peak excursion before they get real noisy.

The carrier is exactly 13 inches tall and the inside of the box is also 13 inches. It is quite easy to slide the carrier in and out. I actually built the carrier first and then used it as a jig to build the cabinet. That way it has to fit!

I wired the speakers up to a Radio Shack dual banana plug that is screwed to the floor of the cabinet. I drilled and countersunk a hole right through the plug and then attached it with a short sheet rock screw. Beats the hell out of $50 binding posts! I had a few cans of "Wrinkle Paint" hanging around so I finished the inside of the cabinets with it. It is more trouble than it is worth, but looks cool in the right application.

Here is the whole system in my living room. I want to put the dipole woofers next to the floor-wall intersection but I don't really have one. Someday I will have a room dedicated to Audio and it will be symmetrical! I am really pleased with the sound of my system now.

Change to Adire DPL12 Woofers and Integration of Stryke350 Amp into Cabinet

After listening to the Phoenix-Type Dipole woofers with the Lambda sealed box drivers a few things became apparent. First, I really liked the sound of these woofers. It was the most natural sounding Bass I had ever heard. Second, the Lambda woofers made too much aerodynamic noise when pushed to extremes. These are some great woofers and are designed for a sealed box. When used in this application, they are fantastic until they get up to about 3/4 inch P-P excursion. Then the pole vent starts huffing and chuffing like "The Little Engine That Could". When these are in a sealed box as they are intended, this noise is never heard.

I started looking for a new set of drivers that would have low distortion and low aerodynamic noise at the same time with great excursion. Along came the Adire Audio DPL12 ( http://www.adireaudio.com/diy_audio/drivers/adire/dpl12.htm ). This is a purpose built woofer just for this kind of application. Very large pole vent and a vented spider along with an Xmax of 14.3 mm and a Fs of 16 Hz. Just what I needed. I also looked at the Lambda dipole woofers. These have great potential also, but are way more expensive, have less Xmax and as of 2003 will no longer be manufactured or supported. Nevertheless, I bought 6 of them and will build something that optimizes there capabilities in the future. I just don't think they are the right driver to bury deep inside a Phoenix-type box. They are beautiful and deserve to live out in full view!

I replaced the Lambda woofers with the Adire DPL12 woofers and immediately heard a great improvement. I was using a Definitive Technologies plate amp with about 250 watts output and variable boost at 30 Hz and volume control. Using LAud to look at the in-room frequency response and adjusting the controls I got a pretty good sound from this combination. It was not perfect, but sounded great. I used the boost to provide the 6 db dipole correction slope. Neither the high pass or low pass frequencies are adjustable.

Next I ordered two 350 Watt plate amps from Stryke Audio. These have variable crossover frequency, phase and volume. I hooked these up to the DPL12 woofers and using LAud equalized as best I could. This was a little better that before but was still not what I wanted. I have worked on these Speakers for nearly six years now and I wasn't going to compromise at this point. I decided to bypass all of the filter circuits in the Stryke amps and build a circuit based on Siegfried's crossover. I already had an active crossover for the Dynaudio Mids and the power supply for that would easily handle a few more op-amps.

Using Siegfried's circuit design and measured data from LAud I entered the data into CALSOD and came up with a circuit that worked for these woofers. It was mounted in the same case as the active crossover for the mids. The two stacked boards on the right are the left and right sub crossovers.

The measured frequency response using a microphone right at the front plane of the speaker cabinet and the actual filter response shows that the filter is doing just what it is supposed to do. The 6 db dipole correction is in the filter response but does not show up in the driver response since it is close mic'ed

These are plate amps and measure about 12 inches by 12 inches. There was no place to mount them on a Phoenix type W-frame subwoofer. What to do? Cut the amps into two smaller pieces and put them inside the woofer enclosure. Hot Damn! This might actually work!

I built an umbilical using some Molex connectors and spare wire so I would be able to disassemble the whole contraption later. I cut off the lower portion of the control panel and remounted the IEC Power connector to make the parts a little smaller. I am not using any of the speaker terminals still on the board, but it was easier to leave them on.

I bypassed most of the filter circuits built into the amp with one simple wire. The only thing left is the volume control. The jumper takes off at the R25/C13/U3A junction, which is just after the volume control, but before the phase and boost circuits. The resulting frequency response is flat out to about 10K. Plenty good enough for subwoofer duty.

The driver carrier, described earlier, was the scene for the mounting of the main part of the amplifier. It is really crowded in there but with a little trial and error the transformer just cleared the spokes on the woofer basket. It is not visible here but it is right behind the woofer magnet.

Side view of the driver carrier with the amp installed. Here you can see some of the details of the Adire Audio DPL12 woofers. The vents around the spider have screens over them to keep out mice and other small animals. Twelve narrow spokes and 1.25 inch pole vent.

Sliding the carrier back into the cabinet most everything disappears except the heatsink. Pretty slick! Huh!

Speaker

1x12" PA Cabinets Using Recycled Parts

Cost

• 2lbs of Screws: free, already had them
• Carpet: $25
• "Cornbread": $5
• Metal boxes: $5
• Paint: $2
• Liquid Nails: free, left over from some construction
• Chicken wire: free, friend had a roll
• Wood: free, found behind a cabinet shop
• Speakers: free to me, friend's dad bought them for $8 from a thrift store
• Crossovers: free, capacitors that were lying around

Total: $37

Time: two weeks

Dimensions

(These assume wood of 3/4" thickness)

4pcs 23 1/8" x 14 1/2" (back/front)

4pcs 23 1/8" x 11" (sides)

4pcs 16" x 11" (top/bottom)

The tweeter is 4" x 10", with the center 5" from the top of the front piece.

The woofer has a 12" diameter and its center is 10" from the bottom of the front piece.

It's funny how these dimensions came out perfect. And I mean perfect. We simply drew them out on a large piece of MDF and luck saved us again and again. The chicken wire that went over the front, for example, was exactly as tall as the speakers. No matter how we chopped up the metal corners we used, there was not more than 2 inches of waste from each section I bought. The carpet from Home Depot could be evenly cut to cover the speakers with no waste. There is something godly, something universal in these dimensions. When a future Einstein publishes the Grand Unifying Theory it will no doubt come from careful study of these speaker cabinets. Until then, note that when the saw cuts, it will remove about a 1/16" of wood. We left an 1/8" gap between each section that would be cut out.

Remember to recess the front about half an i

nch so that the speaker surrounds do not interfere with the grill.

Crossover

In any serious discus

sion of crossover design you will have to have specs on the speakers you are using. For the 12" woofers, I could not find anything. For the '68 Magnovox tweeters, I wasn't going to find anything. Well, now those engineering classes have a chance to prove their use!

First you need to know the impedance of the speakers. The woofers were 8 ohm. I guessed that the tweeters would be between 8 and 16 ohms. I measured them with a VOM (which shouldn't work too well... but the tweeters are so low powered compared to the woofers that they are almost negligible anyway) and it read 12 ohms. I just needed an approximate value to calculate the size of capicitor needed.

The crossover should make the output from the cabinets at all used frequencies flat. Another consideration: the woofer can handle any frequency given to it. The tweeter, however, distorts at lower frequencies. So it must have a high-pass filter before it. The tweeter, being lower powered, also needs to have the signal attenuated with a resistor. The circuit should have an impedance of 8 ohms since it is being powered by a home stereo, car audio uses 4ohm amplifiers. A car amp could power the speakers, the output will just be quieter than with 4 ohm speakers. Powering 4 ohm speakers with an 8ohm amp is not a good idea since it was designed to drive a lesser load and turning the volume up can kill the amp. I actually tried this. The car amp had an automatic shut off. When I played chords at high volumes, it would turn off as I plucked the chord, then turn back on. After a while it overheated and shut off. After repeating this abuse several times, it refused to turn back on again.

I wanted to see what the frequency response of the speakers was. I tied a microphone to my desk lamp (attached to my computer) and put the speakers below it. Then I hooked them up to my stereo (which is attached to my computer). I downloaded test tones from a site I found on Google. I played each test tone and watched the levels in SoundForge coming from the microphone. This test is not very good because the microphone does not have a flat response and the preamp in the sound card probably colors the input even more. The microphone also fell off the lamp when I switched from the woofer to the tweeter, so I could not tell how the levels varied between them. Later I learned that one can put a VOM on the speaker's terminals and the measured voltage will be proportional to the volume.

Matlab graph of the results (available in text here)

The volume for both speakers, according to the computer, dropped off after 2khz. According to my ears, the tweeter worked just fine up to about 10khz. I guess the microphone or the sound card's preamp cannot handle higher frequencies. Also, the tweeter distorted at 200hz and below.

What does this graph tell us? Well, there is enough overlap that we can safely cut out all frequencies below 1khz for the tweeter. Since I wasn't sure about the actual impedance of the tweeter, I just chose 500hz as the cutoff frequency for the filter. Whether the impedance is 8,12, or 16 ohms the cutoff frequency will be safely above the 200hz that makes the tweeter distorts.

Fc=500hz R=12 --> Fc=1/RC -->

The tweeter is louder than the woofer for the same power. But since I couldn't measure how much louder... I decided to listen to it when it was built and add a resistor then.

Finding the Parts

Besides being short on money, I wanted to reuse as much material as possible. I went to home depot and it looked like the wood alone (3/4" plywood or MDF) would cost almost $100. Yet I know there is wood being thrown away everyday. So my friend Victor and I went on a bike tour of local dumpsters. Not surprisingly, we found quite a few construction sites. Most only had CDX plywood (containing knots and holes... not the kind of stuff you want to use for speakers). I did find enough of that and good 2x4s to build a small shed but... Then we took a car and hit up some cabinet shop's dumpsters. There was more than enough MDF. We took a 4ft x 6ft piece.

All those knots are no good for cabinets... but you might find carpeting and other stuff

I had to give in and go buy carpeting. It cost $25, the single largest expenditure in the project. It really helps the boxes look professional thouogh. While I was there I bought black spray paint ($1/can). I also needed to find something to use as handles and to put the jack socket into. I found electrical boxes for $0.69 each. Flat plates to cover electrical boxes were $0.30. Works for me... You do not want to use handles that stick out. They get caught on things and they are not comfortable for carrying heavy speaker cabinets any distance.

Electrical box

Cover we used as a backplate

You can fill the electrical box with water after soldering it to make sure it does not leak

Doh! The electrical boxes wouldn't go into the holes in the wood without chopping off the edges of the backplate.

Ben and I simply soldered the plates to the back side of the electrical boxes.

Corners were a problem. Door corners cost $3.15 for a pair and looked bad. With 8 corners per box... no way. I remember the metal edges that go over drywall. I had no idea what they are called. The clerk called it "cornbread" or something like that. I bought XXX ft.

Cutting the Wood

This step was the most time consuming and least fun. This is because all I had was a jigsaw and a rotary saw. Victor aided in this step of the construction. We killed his direct drive saw before I went home and got my worm-drive saw. It is hard to cut straight lines. We used a right angle and two clamps to hold a 2 x 4 to each piece of wood. Then, using the 2x4 as a guide, I would cut the wood. This took two days.

After it was all cut, we fit it together and... the lack of straight lines hurt. Some pieces were a little too long or short. We screwed the whole thing together and then it held but with some gaps. So we took the screws out and put Liquid Nails along all the edges. Then we screwed it back together again. A few cracks appeared around some screws but they were solid enough to jump on...

Liquid Nails is our savior

Ben, usually conservative before a camera, seems to be happy with how the cabinets are turning out

Painting

Why would you paint the boxes if they will be carpetted? Because if the carpetting rips you will have tan colored spots showing... Also the front will not be carpetted. You need one can per box. I did two coats.

All the rest of the construction

Now at Ben's garage, we found the center of gravity of the boxes so we knew where to put the handles. Just put the boxes on a broomstick and see where they balance. It was right about in the middle of the boxes. With deeper boxes and bigger speakers it would be towards the front of the box. We carried the boxes around and decided to put the handles 3/4 of the way up the side. A drill and a jigsaw were used to cut the holes out. The plates with the jack for the speaker cable went in at the same height on the back. For these, we drilled holes big enough to fit a 1/8" audio jack. Then we glued the jack with LockTite after tightening the nuts.

Adhesive will keep the jack from coming loose and also seal any gap.

Before putting the boxes in, we carpetted the box. We cut a piece that went from the bottom to the back to the top. The two other sides were cut out seperately. This was to keep the number of corners that had to be covered to a minimum. We used a staple gun to attach the carpet. The staples had to be hammered afterwards to hide them.

We wondered whether the speakers could blow a cable out through the jackhole. We calculated how much volume the woofers move and how much force would be exerted on the jack if that volume were to be pushed through the hole. The answer was a definite no. Also, as I mentioned above, if you make a hole just big enough for the electrical box to fit snugly, it won't go in with the back plate on. So we used a cutting wheel to cut the plates down to size. After the boxes are in, they should be sealed to prevent any leaks.

Silicone tub and tile sealer! Hey, don't sit on this stuff, it doesn't want to wash out. RIP to a new pair of shorts I was wearing.

After that, it was time to hook up the wiring. We had to put capacitors in parallel to get the right values. And we still weren't sure about the resistor, so we left it off. Yes, we'd get back to it right?

I keep burning myself with the soldering iron so Ben did the soldering himself

Now, an enclosed space will have a frequency at which it resonates. My living room, for example, would start to shake when I played a G3. The waves add constructively and soon the windows are rattling. The same will be true of the box. What is needed is some sort of dampening to keep the vibration from increasing. Batting is the solution. I used the insides of several old pillows.

Careful with the wiring while you put the batting in

After this step, the speakers went in. We painted the screws yellow earlier. You had better hope you made good fitting holes for the speakers otherwise your screws will have nothing to attach to or the speakers won't fit inside. We put silicone sealant on the bottom sides of the speakers to keep them from rattling and to prevent air leaks. Luckily Ben had black silicone that matched with the color of the cabinet.

Just don't get this stuff on the cones

Then it was time for the corners. We chopped the cornbread into sections of the right length. We put a screw in every other hole in the cornbread. The ends of the corners were sharp, so we cut them at 45 degree angles with tin snips on each side then filed and hammered them till they were round.

These corners are still problematic but better than nothing.

The rubber feet you see in the picture above are stoppers from discarded labware. We drilled holes halfway into them, then put a screw in the hole. They haven't held up very well but they do keep the cabinets from scratching hardwood floors.

Before we finished with the corners, we cut chicken wire and placed it over the front. We stapled it on then put the remaining corners in place.

We tested the speakers out. Very loud, but too much treble. The resistors! Gah! Now that the box is all sealed how do we get inside to add them? It was midnight by now, we both wanted to go to sleep, so the solution was to turn the treble on EQ down and leave that for another day. That day has not yet arrived...

I wanted to put some sort of badge on the speakers. The only pair we found were these Darwin fish.